* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download FOMITES IN INFECTIOUS DISEASE TRANSMISSION: A
Document related concepts
FOMITES IN INFECTIOUS DISEASE TRANSMISSION: A MODELING, LABORATORY, AND FIELD STUDY ON MICROBIAL TRANSFER BETWEEN SKIN AND SURFACES. A DISSERTATION SUBMITTED TO THE DEPARTMENT OF CIVIL AND ENVIRONMENTAL ENGINEERING AND THE COMMITTEE ON GRADUATE STUDIES OF STANFORD UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY Timothy Ryan Julian December 2010 © 2011 by Timothy Ryan Julian. All Rights Reserved. Re-distributed by Stanford University under license with the author. This work is licensed under a Creative Commons AttributionNoncommercial 3.0 United States License. http://creativecommons.org/licenses/by-nc/3.0/us/ This dissertation is online at: http://purl.stanford.edu/cf347cn1097 ii I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy. Alexandria Boehm, Primary Adviser I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy. James Leckie I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy. Robert A Canales Approved for the Stanford University Committee on Graduate Studies. Patricia J. Gumport, Vice Provost Graduate Education This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file in University Archives. iii Abstract This dissertation examines the factors that influence fomite-mediated (e.g., indirect contact) transmission of viral gastrointestinal and respiratory illness. Specifically, the dissertation investigates virus transfer between surfaces and virus recovery from surfaces, models human-fomites interactions to estimate exposure and infection risk, and elucidates causal links between microbial contamination and illness in child care centers. Indirect contact transmission refers to person-to-person transmission of disease via an intermediate fomite (e.g., inanimate object acting as a carrier of infectious disease). The role of indirect contact in disease spread is poorly understood in part because the transmission route of viral pathogens is often difficult to determine. Transmission of respiratory and gastrointestinal viruses can occur through multiple routes (e.g., direct contact, indirect contact, airborne, and common vehicle), and the relative contribution of each route to total disease burden is unclear. The first study in this dissertation examines virus transfer between skin and surfaces, a necessary step in fomite-mediated transmission of viral disease. In the study, transfer of virus between fingerpads and fomites is explored in a laboratory setting. Bacteriophage (fr, MS2, and φX174) are used as proxies for pathogenic virus, and over 650 unique transfer events are collected from 20 different volunteers. The study concludes that approximately one quarter (23%) of recoverable virus is readily transferred from a contaminated surface (e.g., a fomite) to an uncontaminated surface (e.g., a finger) on contact. Using the large data set, the direction of transfer (from fingerpads-to-fomite or fomite-to-fingerpad) and virus species are demonstrated to both significantly influence the fraction of virus transferred by approximately 2-5%. To investigate the relative importance of factors contributing to fomite-mediated iv transmission, a child’s risk of illness from exposure to a contaminated fomite is modeled. Specifically, the model estimates a child’s exposure to rotavirus using a stochastic-mechanistic framework. Simulations of a child’s contacts with the fomite include intermittent fomite-mouth, hand-mouth, and hand-fomite contacts based on activities of a typical child under six years of age. In addition to frequency of contact data, parameters estimated for use in the model include virus concentration on surface; virus inactivation rates on hands and the fomite; virus transfer between hands, fomite, and the child’s mouth; and the surface area of objects and hands in contact. From the model, we conclude that a childs median ingested dose from interacting with a rotavirus-contaminated ball ranges from 2 to 1,000 virus over a period of one hour, with a median value of 42 virus. These results were heavily influenced by selected values of model parameters, most notably, the concentration of rotavirus on fomite, frequency of fomite-mouth contacts, frequency of hand-mouth contacts, and virus transferred from fomite to mouth. The model demonstrated that mouthing of fomite is the primary exposure route, with hand mouthing contributions accounting for less than one-fifth of the childs dose over the first 10 minutes of interaction. Based on the findings from the model that concentration of virus on a fomite influences a child’s risk of illness, we investigate methods to recover virus from fomites. In a literature review and subsequent meta analysis, we demonstrate that the outcome currently used to describe virus contamination, positivity rate, is biased by the authors’ selected sampling methods. We follow up, in the laboratory, with a comparison of the identified methods and demonstrate that polyester-tipped swabs prewetted in 1/4-strength Ringer’s solution or saline solution is the most efficient sampling method for virus recovery tested. The recommended method is compatible with plaque assay and quantitative reverse-transcription polymerase chain reaction, two techniques used to quantify virus. The link between hand / fomite contamination and infection risk was explored in a field study at two child care centers over four months. Both respiratory and gastrointestinal disease incidence were tracked daily, while hand and environmental surface v contamination were monitored weekly between February 2009 and June 2009. Microbial contamination was determined using quantitative densities of fecal indicator bacteria (e.g. Escherichia coli, enterococci, and fecal coliform) on hands and fomites as well as presence/absence of viral pathogens (e.g. enterovirus and norovirus). Health was monitored daily by childcare staff, who tracked absences, illness-related absences, and symptomatic respiratory and gastrointestinal illness. The resultant data suggests that increases in microbial contamination led to increases in symptomatic respiratory illness four to six days later, in agreement with typical incubation periods for respiratory illness. Similarly, respiratory illness led to increases in microbial contamination on hands during presentation of symptoms, and on fomites in the following three days. vi Acknowledgments Without the contributions of the people named below, as well as many people unnamed, the following dissertation would not have been possible. First, to Dr. Alexandria B. Boehm who served as my advisor throughout my time here at Stanford. It has been the utmost honor to have worked so closely with such a brilliant and enthusiastic scientist. Dr. Boehm’s immediate understanding of, and aid in resolving, the many obstacles I encountered along the way drove the dissertation ever onward. Without her innumerable contributions, the work herein would not have been possible. Second, to Dr. James O. Leckie for the many enjoyable meetings over the years. Our topics of discussion ranged from the intricacies of the experimental design to Stanford sports, from data analysis to the political system. I never left Dr. Leckie’s office without being excited by new research avenues and intrigued by his questions. The projects within would not have been possible without the interest and expertise of my other committee members. I thank Dr. Robert A. Canales for his advice and mentoring; he lit my interest in exposure assessment modeling and environmental statistics. I also thank Dr. Lynn M. Hildemann, whose contributions during the research proposal phase improved the quality of the work, and motivated the field portion. Finally, I thank Dr. Yvonne A. Maldonado, committee chair, for her contributions to the refinement of the dissertation through her expertise in pediatric infectious disease. Special thanks to Dr. Paloma Beamer, both mentor and friend. While a graduate student, Dr. Beamer proposed the application of the chemical exposure modeling framework to biological agents; her work was the impetus for this dissertation. vii Much of the brain power and laboratory work embedded in the dissertation was contributed by research colleagues and friends. Specifically, Willa AuYeung, Daniel Keymer, Karen Knee, Royal Kopperud, Blythe Layton, Joey McMurdie, Mia Morgan, Allison Pieja, Todd Russell, Alyson Santoro, Lauren Sassoubre, Nick de Sieyes, Francisco Tamayo, Emily Viau, Sarah Walters, George Wells, Simon Wong, Kevan Yamahara, and Valentina Zuin. Additionally, Amy J. Pickering contributed significantly, especially to the child care center study which would not have been possible without her seemingly inexhaustible contributions of time and effort. Thanks, also, to Joell Hamby, Brenda Sampson, and Sandra Wetzel for administrative support. The decision to attend graduate school at Stanford University was most influenced by my interactions with undergraduate advisors from Cornell University. Going forward, I hope that I reflect the enthusiasm of Dr. Louis D. Albright, Dr. Rebecca L. Schneider, and Dr. Michael B. Timmons in my approach to research and teaching. Thanks to the many friends I am lucky to have made both before and during my time at Stanford. Our time together, often spent camping, hiking, at dinner parties, at rock concerts, and sharing never ending pasta bowls, has passed too quickly. Thanks to Nathan, Naveen, and Sean for their lifelong friendships forged through heated debates (academic, political, and otherwise) over glasses of Scotch. To Sara, I cannot thank her enough for everything: from breakfast this morning to love, calm, and balance every day. From help with statistical modeling to watching music videos of The Darkness. She is my best friend. And to my family: Thomas, Eileen, Tommy, Missy, and Dani. For my entire life, they have provided an endless supply of encouragement, love, patience, and support. Without them, none of this would be possible. Nor would it have been as enjoyable. We should all be so blessed as to have such a wonderful family. The dissertation research was funded by the Shah Family Research Fellowship for Catastrophic Risk from Stanford University, the United States Environmental Protection Agency (USEPA) Science to Achieve Results Graduate Fellowship Program and the UPS Endowment Fund at Stanford University. EPA has not officially endorsed this dissertation and the views expressed herein may not reflect the views of the EPA. viii Contents Abstract iv Acknowledgments vii 1 Introduction 1 1.1 Fomites in Infectious Disease Burden . . . . . . . . . . . . . . . . . . 1 1.2 Transmission Routes of Infectious Disease . . . . . . . . . . . . . . . 3 1.2.1 Vectorborne Transmission . . . . . . . . . . . . . . . . . . . . 3 1.2.2 Airborne Transmission . . . . . . . . . . . . . . . . . . . . . . 4 1.2.3 Common Vehicle Transmission . . . . . . . . . . . . . . . . . . 4 1.2.4 Contact Transmission . . . . . . . . . . . . . . . . . . . . . . . 5 1.3 History of Fomite-Related Research . . . . . . . . . . . . . . . . . . . 6 1.4 Quantitative Microbial Risk Assessment . . . . . . . . . . . . . . . . 11 1.5 Dissertation Organization . . . . . . . . . . . . . . . . . . . . . . . . 13 1.6 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 1.7 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 1.8 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2 Virus Transfer 22 2.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 2.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 2.3 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 25 2.3.1 Volunteers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 2.3.2 Virus and Preparation of Inoculum . . . . . . . . . . . . . . . 25 ix 2.3.3 Plaque Assay . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.3.4 Virus Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.3.5 Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.4.1 Virus Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 2.6 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 2.7 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 2.8 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 2.4 3 Rotavirus Exposure Model 39 3.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 3.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 3.3 Model Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 3.4 Methods and Materials . . . . . . . . . . . . . . . . . . . . . . . . . . 45 3.4.1 Parameter Estimation . . . . . . . . . . . . . . . . . . . . . . 45 3.4.2 Model Approach . . . . . . . . . . . . . . . . . . . . . . . . . 49 3.4.3 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . 49 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 3.5.1 Parameter Estimation . . . . . . . . . . . . . . . . . . . . . . 50 3.5.2 Model Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 3.5.3 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . 55 3.6 Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 3.7 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 3.8 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 3.9 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 3.5 4 Virus Recovery from Surfaces 72 4.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 4.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 4.3 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 76 4.3.1 76 Review of Virus Surface Sampling Literature . . . . . . . . . . x 4.3.2 Laboratory–Based Surface Sampling Method Comparison . . . 78 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 4.4.1 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . 82 4.4.2 Laboratory–based Surface Sampling Method Comparison . . . 83 4.4.3 qRT–PCR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 4.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 4.6 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 4.7 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 4.8 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 4.4 5 Health and Surfaces in Child Care Centers 96 5.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 5.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 5.3 Methods and Materials . . . . . . . . . . . . . . . . . . . . . . . . . . 100 5.4 5.3.1 Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 5.3.2 Surveys / Demographic Data Collection . . . . . . . . . . . . 101 5.3.3 Sampling Scheme . . . . . . . . . . . . . . . . . . . . . . . . . 101 5.3.4 Health Data Collection . . . . . . . . . . . . . . . . . . . . . . 101 5.3.5 Hand Rinse Sampling . . . . . . . . . . . . . . . . . . . . . . . 102 5.3.6 Environmental Surface Sampling . . . . . . . . . . . . . . . . 103 5.3.7 Microbiological Methods . . . . . . . . . . . . . . . . . . . . . 103 5.3.8 Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 5.4.1 Surveys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 5.4.2 Health Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 5.4.3 Hand Rinse Samples . . . . . . . . . . . . . . . . . . . . . . . 107 5.4.4 Environmental Samples . . . . . . . . . . . . . . . . . . . . . . 108 5.4.5 Health Associations with Hand and Surface Contamination . . 108 5.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 5.6 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 5.7 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 xi 5.8 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 6 Conclusions 126 6.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 6.2 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 A Supplemental Material for Chapter 3 134 B Supplemental Material for Chapter 4 137 B.1 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 C Supplemental Material for Chapter 5 150 C.1 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 C.1.1 Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 C.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 C.2.1 Bivariate Correlations . . . . . . . . . . . . . . . . . . . . . . 152 C.2.2 Hand Contamination and Health Data. . . . . . . . . . . . . . 152 C.2.3 Hand Contamination and Environmental Contamination. . . . 153 C.2.4 Environmental Contamination and Health Data. . . . . . . . . 153 C.2.5 Health Associations with Hand and Surface Contamination . . 154 C.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 C.3.1 Use of Multiple Comparisons . . . . . . . . . . . . . . . . . . 154 C.4 Tables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 xii List of Tables 1.1 Gastrointestinal and Respiratory Viruses Transmitted Via Fomites . . 17 1.2 Evidence of Viruses in Fomite Transmission . . . . . . . . . . . . . . 18 2.1 Descriptive Statistics of Fraction Transferred for Each Subset. . . . . 35 2.2 Distribution Parameters for Fraction Transferred by Phage Type. . . 36 3.1 Input Parameters and Estimated Values for Exposure Model . . . . . 61 3.2 Sensitivity Analysis of Exposure Model . . . . . . . . . . . . . . . . . 62 4.1 Eluents Used to Remove Virus from Fomites . . . . . . . . . . . . . . 92 4.2 Implements Used to Remove Virus from Fomites . . . . . . . . . . . . 93 4.3 Comparison of Recovery of Infective Phage . . . . . . . . . . . . . . . 94 4.4 Surface Material, Implement, and Eluent Influence on Recovery . . . 95 5.1 Summary of Environmental Fomites Samples. . . . . . . . . . . . . . 115 5.2 Pathogen Detection PCR Parameters . . . . . . . . . . . . . . . . . . 116 5.3 Child Care Center Population Demographics . . . . . . . . . . . . . . 117 5.4 Child Care Center Population Health and Hygiene Knowledge . . . . 118 5.5 Frequency of Absenteeism and Symptomatic Illness in Child Care Centers119 5.6 Respiratory Illness as Function of Enterococci on Surfaces . . . . . . 120 B.1 Summary of Studies in Literature Review. . . . . . . . . . . . . . . . 147 B.2 Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 B.3 Positivity Rate Metrics By Virus . . . . . . . . . . . . . . . . . . . . 149 C.1 New Illness Episodes Model . . . . . . . . . . . . . . . . . . . . . . . 157 xiii C.2 Illness-Related Absences Model . . . . . . . . . . . . . . . . . . . . . 158 xiv List of Figures 1.1 Infectious Disease Transmission Routes . . . . . . . . . . . . . . . . . 20 1.2 Steps Required for Fomite-Mediated Transmission . . . . . . . . . . . 21 2.1 Histograms of Fraction Virus Transferred . . . . . . . . . . . . . . . . 38 3.1 Schematic Model of Virus Transfer . . . . . . . . . . . . . . . . . . . 64 3.2 Simulated Timing of Contacts . . . . . . . . . . . . . . . . . . . . . . 65 3.3 Simulated Concentration, Exposure, and Dose Profiles . . . . . . . . 66 3.4 Temporal Trends in Concentration and Exposure Profiles . . . . . . . 67 3.5 Dose and Risk Boxplots . . . . . . . . . . . . . . . . . . . . . . . . . 68 3.6 Dose and Infection Risk as Function of Virus Concentration . . . . . 69 3.7 Temporal Trends in Dose . . . . . . . . . . . . . . . . . . . . . . . . . 70 3.8 Temporal Sensitivity Analysis of Fraction Transferred and Contact Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 4.1 Fraction of Virus Recovered By Implement Eluent Combinations . . . 90 5.1 Time Series of Absences at Child Care Centers . . . . . . . . . . . . . 122 5.2 Time Series of Reported Symptoms at Child Care Centers . . . . . . 123 5.3 Time Series of Bacteria on Hands in Child Care Centers . . . . . . . 124 5.4 Time Series of Bacteria on Fomites in Child Care Centers . . . . . . . 125 xv Chapter 1 Introduction 1.1 Fomites in Infectious Disease Burden Informed and successful disease control choices are made on the basis of understanding infectious agent transmission routes (Hurst, 1996). Perhaps the most well-known example is also the first: Dr. John Snow’s removal of a water pump handle in London in 1854 halted a cholera epidemic. Similar contemporary interventions tailored to impede transmission include hygiene education (Aiello et al., 2008), improved water quality at the source and in the home, improved sanitation (Fewtrell et al., 2005), social distancing (Glass et al., 2006), and respiratory masks (Jefferson et al., 2009). The success of the interventions relies, in part, on the prior justification that the transmission route is a major contributor to the overall disease burden. Contemporary understanding is that, in particular for respiratory and gastrointestinal virus, transmission is complex, occurring via multiple, likely interrelated, routes (Goldmann, 2000; Nicas and Sun, 2006). Understanding transmission of respiratory illness (RI) and gastrointestinal illness (GI) disease spread, and how to prevent it, will aid reductions in burden. Annually, the average adult has about 2 to 4 acute upper respiratory illnesses. Children have approximately 6 to 8 (Heikkinen and Jrvinen, 2003). Per year, there are 400 million cases of lower respiratory infections, which compared to upper respiratory illnesses, are more likely to lead to hospitalization and death (Monto, 2002; Mathers et al., 1 CHAPTER 1. INTRODUCTION 2 2008). Gastrointestinal illness is responsible for 4.5 billion cases annually, leading to an estimated 1.7 million deaths in children under five every year (Mathers et al., 2008). Combined, acute respiratory infections and diarrheal disease account for 34% of the 10.4 million annual deaths among children under five (Mathers et al., 2008). Viruses, especially those presented in Table 1.1, are commonly responsible for respiratory and gastrointestinal illlness. Although medical treatments, like antibiotics, oral rehydration therapy, and zinc supplementation, are proving very effective (Hahn et al., 2001; Lazzerini and Ronfani, 2008; Roth et al., 2010), one estimate suggests that universal implementation of these methods can only reduce child mortality by an additional 20% (Jones et al., 2003). Combining medical treatments with prevention of disease spread can further reduce RI and GI morbidity. Reductions through prevention of disease spread require an understanding of transmission routes. However, the transmission of RI and GI is complex. Contributing to the complexity is an incomplete understanding of indirect contact transmission. Indirect contact transmission refers to person-to-person transmission of disease via an intermediate fomite (i.e., inanimate object acting as a carrier of infectious disease). Indirect contact, or fomite-mediated contact, is poorly understood due, in part, to the nature of the transmission route. There are a number of ways fomites can be contaminated with infectious disease, including contact with bodily fluids, body parts, or other fomites and settling from airborne particles by talking, sneezing, coughing, or vomiting (Hota, 2004; Boone and Gerba, 2007). Contamination of a fomite may provide no obvious or visible evidence of infectious disease presence. Additionally, the routes by which an infectious agent contaminates a fomite are equally able to infect a susceptible individual without the intermediate fomite. Therefore, it is often difficult to determine whether a transmission event occurred directly between an infected host and a susceptible host, or the event occurred indirectly via a fomite. Moreover, the factors that influence fomite-mediated transmission, as well as their relative importance, are poorly understood. Not only is contamination of a fomite a requisite step in indirect contact transmission, but viral persistence and transfer to a susceptible individual are also required. To initiate infection via fomites, a virus must be able to contaminate a fomite, persist on the fomite, come into contact CHAPTER 1. INTRODUCTION 3 with a susceptible host, and to initiate infection in the susceptible host. Common questions concerning RI and GI viruses include whether or not an etiological agent is capable of fomite-mediated transmission, and with what efficiency relative to other routes. Characteristics of viruses relevant to transmission via fomites are provided in Table 1.2. The dissertation presented here contributes to fundamental knowledge concerning the factors that influence fomite-mediated transmission for viral disease. To better understand fomite-mediated transmission, this chapter provides a background on infectious disease transmission routes, a history of research relevant to the contemporary understanding of fomites, and a description of quantitative risk assessment modeling. Modeling is a tool frequently used to assess factors that influence risk of illness in transmission of communicable infectious disease. 1.2 Transmission Routes of Infectious Disease Fomite-mediated transmission is a subset of contact transmission, one of the major routes of infectious disease transmission. There are, arguably, four major routes: vectorborne, airborne, common vehicle, and contact (See Figure 1.1) (James and David, 2001; Mangili and Gendreau, 2005). The major transmission routes are not mutuallyexclusive. Rather, an etiological agent may utilize multiple routes to transfer between infected and susceptible hosts. Similarly, the major routes are not necessarily distinct categories. As an example, indirect contact transmission during preparation may result in a foodborne (i.e., common vehicle) outbreak. 1.2.1 Vectorborne Transmission Vectorborne transmission is similar to fomite-mediated transmission only insofar as to replace the role of inanimate objects with a living vector. Although fomites are occasionally considered vectors (Lemon et al., 2008), this is not a strictly accurate definition (James and David, 2001; Mangili and Gendreau, 2005). That is, vectorborne transmission is the transfer of an infectious agent to a susceptible host via an CHAPTER 1. INTRODUCTION 4 arthropod or vermin intermediary. Vectorborne diseases are among the top twenty most common causes of death, worldwide, due in part to the ubiquity of malaria in low income countries (Mathers et al., 2008). The vector carriage of the infectious agent may simply be a case of mechanical transfer (like the role of a fomite), or the agent may undergo biological transformations during carriage. Examples of the former include dengue virus, west Nile virus, and yellow fever; of the latter include malaria and African trypanosomiasis. In research investigating fomites as causative agents in transmission of vectorborne diseases (specifically dengue virus and yellow fever), no individual exposed to fomites was infected (Ashburn and Caraig, 1907). 1.2.2 Airborne Transmission A second route of transmission is airborne, which typically refers to the aerosolization and movement, over long distances, of pathogens from an infected individual to a susceptible host. Airborne transmission often plays an important role in the contamination of fomites, as initially aerosolized particles may settle onto surfaces (Nicas and Sun, 2006). Similarly, resuspension from contaminated fomites may contribute to airborne transmission (Nicas and Sun, 2006). A common phenomenon in airborne viral transmission is the formation of viral droplet nuclei. Viral droplet nuclei are formed due to water evaporation from expiration by an infected host. At less than 5 µm in diameter, the viral droplet nuclei can remain suspended for long periods (Lowen et al., 2007). Recent evidence suggests that low humidity leads to increased formation of viral droplet nuclei, and therefore more efficient transmission (Lowen et al., 2007). The formation of viral droplet nuclei, however, is not a requirement for airborne transmission. Bacteria, for example, are capable of being transmitted via the airborne route. Tuberculosis is an example of an airborne bacteria (Mathers et al., 2008). 1.2.3 Common Vehicle Transmission Common vehicle transmission is often intertwined with fomite-mediated transmission, but refers more specifically to the potential infection of multiple individuals via a CHAPTER 1. INTRODUCTION 5 single carrier. Forms of common vehicle transmission include foodborne, waterborne, and iatrogenic transmission. For common vehicle transmission to occur, the vehicle needs to be contaminated prior to distribution to susceptible hosts. Food is often contaminated in the environment prior to harvest, during processing for distribution, or during preparation (e.g., inadequate hygiene or cooking) (Bresee et al., 2002). Water, both recreational and drinking, is often contaminated in the environment, as could occur due to poor sanitation, hygiene, and/or inadequate sewage or storm water control (Craun et al., 2006). Drinking water, even if it was previously treated, can be contaminated during delivery and/or storage. Examples of vehicles in iatrogenic transmission, or transmission during medical procedure, include nonsterile injection needles or catheters (Khan et al., 2000; Luijt et al., 2001), and/or infected blood or organs (Iwamoto et al., 2003). Fomites frequently contribute to infectious disease outbreaks that occur via common vehicle transmission. As an example, contamination of food during processing or preparation can occur due to contact with a contaminated surface, like a cutting board. Similarly, infection of a patient by using a nonsterile injection needle could also be considered fomite-mediated transmission. 1.2.4 Contact Transmission Fomite-mediated transmission is most often included in the fourth major mode of transmission: contact. Contact transmission occurs most often either through direct physical contact, which includes both casual (e.g., touching, kissing) and sexual contact, or indirect contact via a fomite. Two other forms of contact transmission are vertical transmission, which is defined as the transfer of disease from a mother to her fetus either in utero or during child birth, and zoonotic transmission, which is the transfer of disease between vertebrate animals and humans. Often, the distinction between contact and other transmission routes is blurred. Vectorborne transmission, as an example, is sometimes included as a subset of contact transmission, especially when vector carriage is simply mechanical (i.e., no biological transformations of the pathogen occur in the host) (Hurst, 1996). Similarly, the CHAPTER 1. INTRODUCTION 6 production of large respiratory droplets during expirations from talking, coughing, or sneezing blurs the distinction between contact and airborne transmission. Large droplets can be spread over short distances and intercepted by a susceptible host or fomite while settling. This form of transmission is sometimes considered a unique transmission route (Friedman and Petersen, 2004), and other times considered a form of contact transmission (Baron and Jennings, 1991; Hurst, 1996). Large droplet settling is one way respiratory pathogens contaminate fomites, contributing to the possibility of transmission via indirect contact (Nicas and Sun, 2006). Contamination of an inanimate object is only the first step in transmission via fomites (See Figure 1.2). Another requirement is that the etiological agent must remain viable on the fomite for a period sufficient for the fomite to come into contact with a susceptible individual. If it is able to persist long enough, the agent must then be able to transfer from the fomite to a point of entrance on a susceptible individual. For respiratory and gastrointestinal diseases, the point of entrance is most often a mucous membrane, such as through the mouth, nose, ears, or eyes. Once transferred, the agent must be able to initiate infection. In summary, the characteristics of the etiological agent, the fomite, the infected individual, and the susceptible host, as well as the interactions between the individuals and the fomite, influence efficacy of fomite-mediated transmission. 1.3 History of Fomite-Related Research Indirect contact via fomites was first identified by Italian physicist and scholar Girolamo Fracastoro, in 1546 (Ravenel, 1931; Clendening, 1960). Fracastero did so in his description of distinct transmission routes. His description included direct contact, indirect contact, and a predecessor to aerosolization (specifically, the “transmi[ssion of a disease] to a distance...merely by looking”) (Clendening, 1960). Fracastero posited that the etiology of a disease determined the transmission route, using specific examples of contemporary diseases (e.g., scabies via indirect contact, and smallpox via aerosolization). He also posited that contaminated fomites may remain so for “two to three years”, and described porous objects (“linen, cloth, and wood”) as more likely CHAPTER 1. INTRODUCTION 7 to act as fomites than nonporous ones (“iron, stone”) (Clendening, 1960). In so doing, Fracastero laid the groundwork for the concept that inanimate objects contribute to disease. The mechanism by which fomites acted remained elusive as Fracastoro’s work predated Agostini Bassi’s germ theory (1835) by almost 300 years. The knowledge that inanimate objects could spark disease outbreaks (in fact, the Latin definition of “fomes” is “tinder”) provided an opportunity for one of the most well-known examples of biological warfare in the New World. British colonialists, during their efforts to combat the Natives in the 18th century, provided the Natives with blankets and handkerchiefs from hospitals to “innoculate” them with smallpox (Fenn, 2000). Epidemics that tore through Native populations in 1763 and 1764 likely resulted from the provision of the blankets (Fenn, 2000). The knowledge of fomites also aided in the prevention of outbreaks. As an example during the height of a plague outbreak in 1835 in Alexandria, Egypt, the British Privy Council quarantined all ships carrying Egyptian cotton into England (Thompson, 1847; Plunket, 1879). The cotton was described as a “fomes”, referring to the etiological agent perceived to be the cause of both the Egyptian epidemic and an earlier epidemic in London that had occurred in 1665. To prevent the epidemic, the cotton bales were to be “rip[ped] open” to “purify... the cotton” through exposure “to sunlight and air” (Plunket, 1879). In the following decades, yellow fever outbreaks on ships led to the incorrect attribution of fomites as the causative transmission route. The prevailing evidence was the occurrence of outbreaks on ships weeks after they had set sail (Bell, 1901). As no sailor was symptomatic at launch, or in the days leading up to symptoms, sailors and scientists attributed the outbreaks to contact with objects such as “personal clothing and books” (Bell, 1901). The implication of fomites prevailed until randomized control trials conducted by Dr. Walter Reed (based on work first posited by Carlos Finlay) proved mosquitoes were the vector for yellow fever (Clendening, 1960). Specifically, Dr. Reed discovered that a period of approximately 12 days needed to pass for a mosquito that had consumed blood of an infected individual to be able to infect another (Bell, 1901). The uncertainty of the route of yellow fever transmission mirrors contemporary uncertainty in transmission routes. As an example, respiratory CHAPTER 1. INTRODUCTION 8 illness was perceived to be transmitted only through airborne transmission as little as forty years ago. Among the first work to rigorously test the hypothesis that a respiratory virus could be transmitted via contact came in the 1970s. J. Owen Hendley and Jack Gwaltney of the University of Virginia proposed that direct and indirect contact contributed to transmission of rhinovirus, perpetrator of over one third of the cases of the “common cold” (Hendley et al., 1973). Hendley and Gwaltney showed that rhinovirus was shed from an infected patient, capable of surviving outside the host on an environmental surface, and capable of infecting a susceptible individual who had contacted the contaminated surface (See Figure 1.2) (Hendley et al., 1973). Their work was among the first to delineate the steps necessary for a pathogenic agent to be transmitted via fomites. Laboratory work confirming the ability of rhinovirus to be transmitted via direct (hand-to-hand) and indirect (hand-to-surface) contact soon followed (Gwaltney et al., 1978; Gwaltney, 1982). The work by Hendley and Gwaltney was also among the first to demonstrate, conclusively, that fomites were a viable route for a respiratory pathogen. Prior to their work, symptoms of coughing and sneezing associated with rhinovirus were thought to contribute to its spread via airborne transmission (Hendley et al., 1973). Other studies, though, had tried and failed to demonstrate airborne transmission of rhinovirus (Hendley et al., 1973). Around the same time, interest in nosocomial infections was on the rise, particularly for a common respiratory pathogen (respiratory syncytial virus, or RSV). Caroline Hall and R. Gordon Douglas, Jr. of the University of Rochester, citing the early work of Hendley and Gwaltney, recognized that fomites may be contributing to the spread of RSV in hospitals (Hall et al., 1981). In a series of papers investigating RSV transmission via fomites, Hall et al. (1980); Hall and Douglas Jr (1981); and Hall (1983) demonstrated that RSV is capable of following the necessary steps to be transmitted via fomites. That is, RSV survives on surfaces, is readily transferred between surfaces and hands, and can infect susceptible hosts when a contaminated hand contacts their nose or eyes. Scaling up from the laboratory to the field, Hall et al. (1981) also examined the risk of handling infected infants at various levels of contact CHAPTER 1. INTRODUCTION 9 and demonstrated that large droplet contact and indirect contact with fomites were more efficient routes of RSV transmission than small particle aerosolization (Hall et al., 1981). The work by Hendley and Gwaltney along with the work of Hall and Douglas contributed to a renewed interest in fomites in infectious disease transmission as they demonstrated that fomites were an integral transmission route in diseases previously perceived to be primarily airborne. In the following decades (1980s-1990s), research on fomites increased four-fold, with studies examining their role in the transmission of respiratory (Dick et al., 1987; Brady et al., 1990), gastrointestinal (Butz et al., 1993; Wilde et al., 1992) and even bloodborne pathogens (Ferenczy et al., 1989). The work during this decade followed closely the work of Hendley, Gwaltney, Hall, and Douglas, in that it examined pathogen presence/absence on surfaces (Keswick et al., 1983; Piazza et al., 1987; Wilde et al., 1992), persistence (Keswick et al., 1983; Ansari et al., 1988; Abad et al., 1994), transfer (Jennings et al., 1988; Ansari et al., 1988), and the relative efficiency of the indirect route of transmission (Dick et al., 1987). The general acceptance of the role of fomites in infectious disease transmission was highlighted, perhaps, during this period with the First European Meeting of Environmental Hygiene in Dusseldorf in 1987. Research on fomites began to abate in the mid-1990s. The focus of most of the published articles on fomites during this time period continued to be their role in nosocomial infections (McCluskey et al., 1996; Bures et al., 2000; Neely and Sittig, 2002; Das et al., 2002). Additionally, new work was published investigating the role of fomites in animal diseases (Pirtle and Beran, 1996; Otake et al., 2002) as well as on tracking the role of fomites in GI and RI outbreaks (Cheesbrough et al., 2000; Rogers et al., 2000; Abad et al., 2001; Barker, 2001; Evans et al., 2002; Das et al., 2002). The latter proved influential in the resurgence of fomites research over the past several years (2004-2010). In particular, growing concern over two communicable diseases, norovirus (a gastrointestinal virus) and influenza (a respiratory virus), contributed to a resurgence in fomites research. Norovirus, first idenfied in 1972 (Kapikian et al., 1972), is the most CHAPTER 1. INTRODUCTION 10 common cause of gastroenteritis in the United States due, in part, to its extreme contagiousness (as few as 1 viral particles may be needed to cause illness (Teunis et al., 2008), and there are as many as 14 secondary cases for each primary case (Heijne et al., 2009)). Evidence suggests that direct and indirect transmission are important routes for norovirus transmission. For example, the U.S. Centers for Disease Control and Prevention report that 16% of norovirus cases are caused by person-to-person spread (Norovirus: Technical Fact Sheet, http://www.cdc.gov/ncidod/dvrd/revb /gastro/norovirus-factsheet.htm, accessed Sep 2010). Similarly, a series of studies of outbreaks traced the source to environmental contamination of norovirus (Cheesbrough et al., 2000; Evans et al., 2002). Over the last decade, researchers have sought to further investigate the relevance of fomites in norovirus outbreaks (Duizer et al., 2004; Clay et al., 2006; Jones et al., 2007; Girard et al., 2010). Similarly, outbreaks of influenza have increased interest in research on the potential role of fomites in transmission. Contemporary thought supports that aerosolization of small particles, including viral droplet nuceli formation and large droplet contact, are the primary transmission routes (Lowen et al., 2007; Tellier, 2009). Like rhinovirus and RSV, influenza is a respiratory virus. Nevertheless, the contribution of fomites continues to be debated (Brankston et al., 2007). Research to characterize the role of fomites, prompted by the concern over a future pandemic, has mirrored the early work on both rhinovirus and RSV. Specifically, published work has documented influenza survival on surfaces (Thomas et al., 2008; Sakaguchi et al., 2010), disinfection (Rudnik et al., 2009; Weber and Stilianakis, 2008), detection on surfaces (Boone and Gerba, 2005), and the relative efficacy of fomes-mediated transmission relative to other routes (Brankston et al., 2007; Weber and Stilianakis, 2008; Tellier, 2009). Much of this work has been done in the context of tailoring interventions to reduce infectious disease burden during a pandemic. In total, the research dedicated to fomites has concluded that indirect contact is an important route for transmission of respiratory and gastrointestinal illness. Nevertheless, better quantitative data is needed. Research over the last half century has delineated the steps required for an etiological agent to be efficiently transmitted via fomites. Laboratory-scale studies, typically focusing on specific pathogens, have CHAPTER 1. INTRODUCTION 11 demonstrated and quantified organisms’ abilities to transfer to and from fomites, persist on fomites, and to remain infectious. Studies scaling up to the field have demonstrated that contaminated fomites can initiate infection, and have also assessed the efficacy of fomite-mediated transmission relative to other routes. Nevertheless, a better characterization of the factors required for fomite-mediated transmission, and their relationships, is needed. In fact, in a review of the role of fomites in transmission of respiratory and enteric viruses, the authors (Boone and Gerba, 2007) noted the need for “better quantitative data”. Specifically, the Boone and Gerba identified the need for better data on viral inactivation rates, viral transfer between surfaces, and viral distribution and concentration on surfaces. The purpose of better data is to improve and develop “risk assessment models that associate viral infection with fomite contact” (Boone and Gerba, 2007). 1.4 Quantitative Microbial Risk Assessment Contributing to the Boone and Gerba (2007) assertion that data are needed for risk assessment models was the development, in the mid-1990s, of the framework for applying the reductionist approach of quantitative risk assessment (QRA) to infectious disease. QRA is the “technical assessment of the nature and magnitude of a risk caused by a hazard” (Jaykus et al., 1996) where the hazard can include “substances, processes, action, or events” (Covello and Merkhofer, 1993). QRA was first developed in the 1970s and later formalized with the benchmark publication “Risk Assessment in the Federal Government, Managing the Risk”, known colloquially as the Red Book, by the U.S. National Academy of Sciences in 1983 (NRC, 1983). Among the first applications of the QRA framework to assess infectious disease risk were assessments of waterborne transmission (Haas, 1983; Gerba and Haas, 1988; Regli et al., 1991; Rose et al., 1991), which led to the development of a codified framework for quantitative microbial risk assessment by the International Life Sciences Institute (ILSI) in 1996 (ILSI, 1996) and revisited in 1999 (ILSI, 1999). The framework for quantitative microbial risk assessment (QMRA) is adapted from the QRA paradigm. The ILSI framework for microbial risks consists of three CHAPTER 1. INTRODUCTION 12 phases: 1) problem formulation, 2) analysis, and 3) risk characterization (ILSI, 1996, 1999). Problem formulation “identifies the goals, breadth, and focus of the risk assessment, the regulatory and policy context of the assessment, and the major factors” (ILSI, 1999). Analysis develops exposure and dose-response assessments to work toward risk characterization, which is a quantitative characterization of the likelihood, type, and magnitude of human health effects (ILSI, 1999). Risk characterization also incorporates a transparent accounting of uncertainty or variability contributions to the final risk estimates (ILSI, 1999). A fourth phase (risk management) is occasionally included in the risk assessment paradigm and encompasses the risk mitigation and communication strategies (Haas et al., 1999; Covello and Merkhofer, 1993). A major contribution of the ILSI framework for microbial risks was its emphasis on the “dynamic and iterative process of the risk assessment process”, and that findings in a later stage (e.g., risk characterization) should be used to refine and improve findings from an earlier stage (e.g., analysis). The paradigm for QRA as outlined in the Red Book for human health effects was developed to account for risk from chemical exposures (NRC, 1983; Haas et al., 1999). To adapt QRA to microbial hazards, complexities unique to pathogens need to be considered (ILSI, 1996). The complexities include: 1) growth and/or inactivation of pathogens, 2) non-heterogeneous pathogen distributions in environmental matrices, 3) naturally or artificially acquired immunity, 4) asymptomatic infection, 5) secondary transmission (e.g., spread from an infected individual), 6) multiple endpoints (e.g., infection, illness, mortality), 6) potential for multiple exposure routes, and 7) uncertainty in environmental concentration measurements (e.g., accuracy of detection methods) (ILSI, 1996; Haas et al., 1999). In 1999, many of the first practitioners of QMRA (Haas et al., 1999) summarized and applied the QMRA paradigm to examples from many of the major transmission routes in the first and only textbook on the topic, Quantitative Microbial Risk Assessment. Despite the evidence that fomites play an important role in the transmission of disease, few studies have applied the framework of QMRA to model risk from fomites. In those that have, the estimated risk typically relies on simplistic exposure assessments that model human interaction with fomites based on estimates of the probability CHAPTER 1. INTRODUCTION 13 that a contact event occurs (e.g., 10% chance a fomite is contacted by hand), a constant frequency of the contact event (e.g., mouth is contacted by hand 0.08 times per minute), or a constrained sequence of events (e.g., fomes touches hand, hand then touches mouth) (Gibson et al., 1999; Chen et al., 2001; Gibson et al., 2002; Haas et al., 2005; Nicas and Sun, 2006; Nicas and Best, 2008). Examples of fomite-related quantitative microbial risk assessments include: 1) estimating risk rotavirus infection from clothes laundering (Gibson et al., 1999), 2) estimating risk from contaminated surfaces in health care settings (Nicas and Sun, 2006), 3) estimating risk of cross contamination during food preparation (Chen et al., 2001), and 4) estimating risk reductions acheived through hand hygiene (Gibson et al., 2002). Although the studies provide an important first step toward understanding the factors that influence fomite-mediated transmission, they function as simplified models and do not fully account for complexities of human-fomites interaction in field settings. 1.5 Dissertation Organization This dissertation consists of six chapters devoted to furthering knowledge of the factors that contribute to fomite-mediated infectious disease transmission. This introduction chapter (Chapter 1) provides background on the role of fomites in disease transmission. The four middle chapters (Chapters 2-5) present original research in the form of stand-alone manuscripts, each with its own introduction, methods, results, and discussion sections. The final chapter (Chapter 6), provides general conclusions and areas for future research. The references used throughout the dissertation are merged and appear at the end. Co-authors, along with their contributions to each chapter, are listed at the beginning in an introductory paragraph. I am first author on all publications that have been or will be generated from the work included in this dissertation as I was the primary person responsible for planning, conducting, and writing each project. If fomites play a significant role in viral disease transmission through hand contact, virus must transfer from contaminated fingers to fomites and transfer from fomites CHAPTER 1. INTRODUCTION 14 to fingers of a susceptible host. This sequence of events was explored through a laboratory experiment using three bacteriophage species as proxies for pathogenic virus. In Chapter 2 we demonstrate that approximately one quarter (23%) of recoverable virus is readily transferred from a contaminated surface (e.g., a fomite) to an uncontaminated surface (e.g., a finger) on contact. The chapter demonstrates, using a robust data set, that the direction of transfer (from fingerpads-to-fomite or fomite-tofingerpad) and bacteriophage species both influence the fraction of virus transferred by approximately 2-5%. The study also suggests that hand washing reduces the fraction of virus transferred on contact due, potentially, to altered skin properties. This mechanism may explain decreases in illness during handwashing interventions, along with the current explanation that handwashing reduces pathogenic virus and bacteria on the hands. In addition to implications concerning hand hygiene effectiveness, the developed data set contributes to work on quantitative microbial risk assessments examining fomites in disease transmission. In Chapter 3 we combine data sets from the previous chapters with a literature review to create a novel exposure and risk assessment model. The model, based on a stochastic-mechanistic framework using a simulation of a child’s interaction with a fomite, is among the first to incorporate detailed descriptions of sequential time series data modeling human-environment interaction in a microbial risk assessment. A combined sensitivity and uncertainty analysis identifies the factors that most significantly influence risk of infection. Although the analysis demonstrates that parameters describing human interaction are influential, uncertainty of, and variability in, virus concentration on fomites is shown to dominate risk of exposure, and therefore infection. To improve estimates of microbial contamination on surfaces, we compare methods used to recover virus from fomites in Chapter 4. The literature review and subsequent meta analysis demonstrate that the outcome currently used to describe virus contamination, positivity rate, is biased by the authors’ selected sampling methods. In the review, we identify the most promising virus recovery methods. We follow up, in the laboratory, with a comparison of the identified methods and demonstrate that polyester-tipped swabs prewetted in 1/4-strength Ringer’s solution or saline solution CHAPTER 1. INTRODUCTION 15 should be the standardized method for virus recovery. The recommended method is compatible with two common techniques used to quantify virus from the environment, plaque assay and quantitative reverse-transcription polymerase chain reaction. Quantification of virus from fomites is an important direction for future research, as few identified papers on virus surface contamination have quantified virus, or indicators of virus, contamination. Chapter 5 examines the relationship between microbial contamination on surfaces and adverse health outcomes in child care centers in Northern California. For four months in 2009, we quantified fecal indicator bacteria on hands and surfaces twice weekly at two child care centers. We simultaneously collected data on child absences and observable symptoms of gastrointestinal and respiratory infection. Using statistical modeling, we demonstrate that increased surface contamination both leads and lags observable respiratory symptoms. The study is among the first to infer, using longitudinal data, a causal link between indoor microbial contamination and health outcomes. The research presented in the dissertation addresses the role of fomites in infectious disease transmission. The dissertation also contributes to the development of ideas for future research directions. In Chapter 6, new hypotheses generated over the course of the dissertation are discussed. Also in the final chapter is a general conclusion on the role of fomites in disease transmission. 1.6 Acknowledgments The author acknowledges Sara J. Marks and the Stanford University School of Engineering Technical Communication Program for suggestions to improve the chapter, as well as to the website www.dezignus.com for hosting the royalty-free vector images of people used in Figure 1.1 and Figure 1.2. CHAPTER 1. INTRODUCTION 1.7 Tables 16 -2 0 before symp. 0 0 0 -6 0 0 0 2-5 1-3 2-4 3-7 1-2 13 +7 4-21 3-7 up to 21 8 20 14-21 2-4 4-7 Infectious Period (d) Onset Conclusion 2-14 0.5-4 7-14 0.5-2 1-3 Incubation (d) airborne, contact airborne, contact contact, airborne contact, airborne contact, airborne airborne, common vehicle, contact common vehicle, contact common vehicle, contact common vehicle, contact, airborne common vehicle, contact Routes Table 1.1: The epidemiological characteristics of common gastrointestinal and respiratory viruses transmitted via fomites. Infectious period onset refers to the number of days prior to presentation of symptoms whereby shedding occurs, with 0 representing same day as symptoms. The list of routes appear in order of most to least efficient, as currently understood. Data adapted and compiled from reviews by (Boone and Gerba, 2007; Donowitz, 1999) Gastrointestinal Adenovirus Astrovirus Enterovirus Norovirus Rotavirus Respiratory Coronavirus Influenza Parainfluenza Respiratory Syncytial Virus Rhinovirus Virus CHAPTER 1. INTRODUCTION 17 90-200 120-300 30 Parainfluenza RSV Rhinovirus icosahedral spherical icosahedral filamentous spherical spherical icosahedral icosahedral icosahedral icosahedral icosahedral Shape 6.4 - 6.8 n.a. n.a. 5-7 n.a. 8 5.5-6 4-7 n.a. 4.5 IEP +-sense ss RNA –sense ssRNA +-sense ssRNA ds RNA +-sense ss RNA +-sense ss RNA ds RNA +-sense ss RNA +-sense ss RNA +-sense ss RNA linear ds DNA Nucleic Acids door handles, phones table, bedding, remote control toys, counters, keyboards, toilets desks, computers, phones, tables, lightswitch - table, bedding, glasses, lamp phone, toilet, light switch phone, toilet bowl carpets, phones lightswitch, toilet phone, lightswitch, doorknob, toys, tables Detected on Fomites 0.21.25 2 0.0830.33 0.751.5 0.25 0.0028 n.a. 0.0010.002 0.0028 0.0028 >25 2.5 2-6 24-48 3-12 >1440 360>720 n.a. 1440 >720 n.a. 0.63-0.95 0.0280.042 0.5 1 0.006 -0.33 0.0028 0.002 -0.025 0.0028 0.011 n.a. 5-8 10 72 2-12 72>720 1601350 >1440 160 360 Persistence on Surfaces Porous Non-Porous Rate time(h) Rate time(h) 0.51% proven 0-1.5% n.a. n.a. 1-16% 7-13% proven n.a. n.a. Transfer (hand-fomite) Table 1.2: Common gastrointestinal and respiratory viruses transmitted via fomites, including chemicophysical attributes, detection on fomites, inactivation rates on surfaces, and evidence of transfer between hands and fomites. Nt “IEP” is the isoelectric point of the virus. Inactivation rate is in units (-log10 ( N )) as measured in days. “RSV” o is respiratory syncytial virus. Enterovirus persistence and transfer rates were estimated using porcine enterovirus. Norovirus persistence and transfer was estimated using feline calicivirus. “n.a.” is used where data are not available. Data adapted and compiled from Hall et al. (1980); Ansari et al. (1988, 1991); Abad et al. (1994); Long et al. (1997); Boone and Gerba (2007, 2010); Michen and Graule (2010) and Chapter 3 80-120 60-80 Rotavirus Influenza 27-38 Norovirus 80-220 17-28 Enterovirus Respiratory Coronavirus 28-35 90-100 Size (nm) Astrovirus Gastrointestinal Adenovirus Virus CHAPTER 1. INTRODUCTION 18 CHAPTER 1. INTRODUCTION 1.8 Figures 19 CHAPTER 1. INTRODUCTION 20 Infectious Disease Transmission Routes Vectorborne Airborne n atio pir ex vector susceptible host susceptible host susceptible host infected host infected host Pathogens: malaria, yellow fever, dengue African trypansimiasis Interventions: insecticides, environmental mitigation, bed nets, window screens, insect repellents Common Vehicle Pathogens: influenza, measles, rhinovirus, respiratory syncytial virus Interventions: respiration masks, social distancing, closing public locations, blocking expirations, mechanical filtration, ultraviolet radiation. Contact Foodborne Direct contaminated foodstuffs susceptible host infected host Waterborne susceptible host Indirect fomite drinking water bathing water recreational water Iatrogenic contaminated medical device, blood, or tissue infected host susceptible host Vertical susceptible host Pathogens: norovirus, enterovirus, rotavirus, poliovirus, rhinovirus, hepatitis A Interventions: water and food quality standards, hand and environmental hygiene, donor blood and organ screening, equipment sterilization infected host (mother) susceptible host (prenatal child) Pathogens: rotavirus, rhinovirus, norovirus enterovirus, hepatitis, human immunoviruss, Interventions: hand and environmental hygiene, pharmaceuticals, prophalxysis. Figure 1.1: Infectious disease transmission routes as grouped into four common categories with examples of common interventions used to reduce burden from example pathogens. Arrows represent movement or transfer of pathogen d to hand she mite persist on fo 2 tion 4 susceptible host e infec initiat trans fer fomit e to hand 3 mouth ite to m o f er nsf tra transfer hand to mouth Figure 1.2: For fomites to act as intermediaries in infectious disease, an etiological agent must be capable of following four distinct steps. The first step (“1”) is that an infectious agent most be shed from an infected host to the fomite. Two common pathways are direct shedding (e.g., large droplet setting from coughing, sneezing, or other expiration) or indirect shedding via hands (e.g., inadequate hygiene after using a restroom facility followed by handling a doorknob). The second step (“2”) is that an infectious agent must be able to persist on a fomite for a period sufficient for the fomite to come into contact with a susceptible host. The third step (“3”) is that an infectious agent must transfer from the fomite to a susceptible host, either by direct fomite-mouth contact (e.g., a child mouthing a toy) or by indirect fomite-hand contact (e.g., handling a colleagues cellphone) followed by hand-mucuos membrane contact (the average adult toches his lips or mouth 10-25 times per hour) . infected host fer trans o fomite t d n a h 1 shed to fo mit e Fomite-Mediated Transmission CHAPTER 1. INTRODUCTION 21 Chapter 2 Virus transfer between fingerpads and fomites The results presented in this chapter originally appeared as a research article in the December 2010 volume of the Journal of Applied Microbiology (Julian et al., 2010). James O. Leckie and Alexandria B. Boehm appear as co-authors, for their contributions to study design, data interpretation, and manuscript improvement. 22 CHAPTER 2. VIRUS TRANSFER 2.1 23 Abstract Aims Virus transfer between individuals and fomites is an important route of transmission for both gastrointestinal and respiratory illness. The present study examines how direction of transfer, virus species, time since last handwashing, gender, and titer affect viral transfer between fingerpads and glass. Methods and Results Six hundred fifty-six total transfer events, performed by twenty volunteers using MS2, φX174, and fr indicated 0.23 ± 0.22 (mean and standard deviation) of virus is readily transferred on contact. Virus transfer is significantly influenced by virus species and time since last handwashing. Transfer of fr bacteriophage is significantly higher than both MS2 and φX174. Virus transfer between surfaces is reduced for recently washed hands. Conclusions Viruses are readily transferred between skin and surfaces on contact. The fraction of virus transferred is dependent on multiple factors including virus species, recently washing hands, and direction of transfer likely due to surface physicochemical interactions. Significance and Impact of Study The study is the first to provide a large data set of virus transfer events describing the central tendency and distribution of fraction virus transferred between fingers and glass. The data set from the study, along with the quantified effect sizes of the factors explored, inform studies examining role of fomites in disease transmission. Keywords Virus transfer, surfaces, fomites, hand hygiene, environmental hygiene, quantitative microbial risk assessment, bacteriophage. CHAPTER 2. VIRUS TRANSFER 2.2 24 Introduction To better understand transmission routes for viral disease and develop more refined quantitative microbial risk assessment models (Atkinson and Wein, 2008; Nicas and Jones, 2009; Julian et al., 2009) additional information on the importance of fomites in the transmission of viruses is needed (Boone and Gerba, 2007; Brankston et al., 2007). Insight into the role of fomites in the transmission of infectious disease can be obtained by studying the transfer of viruses between skin and surfaces. Virus transfer between skin and surfaces can be described quantitatively by the fraction of virus on a contaminated (donor) surface that is transferred on contact to a recipient surface (Reed, 1975; Gwaltney, 1982; Ansari et al., 1988; Mbithi et al., 1992; Rusin et al., 2002). This fraction could be modulated by a number of factors including the donor / recipient surfaces and the virion surface. Previous studies have reported a wide range of transfer fractions (0.0001 to 0.67) for transfer of a single bacteriophage (e.g., PRD-1) or pathogenic virus (e.g., rotavirus, hepatitis A, human parainfluenza virus-3, rhinovirus) between skin and various surfaces (Reed, 1975; Ansari et al., 1988; Mbithi et al., 1992; Rusin et al., 2002; Bidawid et al., 2004). The range of transfer fractions is significantly influenced by the type of surface (porous or non-porous) contacted by the skin, with transfer between porous and food (e.g., cloth, lettuce, ham, beef, and carrots (Rusin et al., 2002; Bidawid et al., 2004)) surfaces generally lower than transfer to non-porous (e.g., stainless steel and plastic (Reed, 1975; Rusin et al., 2002; Bidawid et al., 2004)) surfaces. Only one published study has examined the transfer between skin and surface of more than one virus. In particular, Ansari et al. (1991) reported a difference in fraction transferred for rhinovirus and human parainfluenza virus-3 between fingers and metal disks. However, the small sample size of the study presumably precluded statistical analysis. The present study explores how viral species and factors including inoculum size, direction of transfer, and skin condition affects virus transfer. We quantify the transfer of three different viruses, MS2, fr, and φX174, between fingerpads and a glass surface. Additionally, we applied experimental treatments to isolate the effects of the following CHAPTER 2. VIRUS TRANSFER 25 on virus transfer: (1) inoculum size, (2) direction of transfer, and (3) skin condition defined by the gender and time since last hand washing. Inoculum size may influence fraction of virus transferred as the phenomenon was demonstrated in bacterial transfer by Montville and Schaffner (2003). Direction of transfer refers to the direction that virus is transferred, such as from skin-to-fomite versus from fomite-to-skin. Gender may influence virus transfer because men typically have a significantly lower skin pH (van de Vijver et al., 2003). Similarly, hand washing shifts the biological and chemical characteristics of the skin by decreasing organic and inorganic constituents (e.g., sebum, sweat, microflora), increasing pH, and decreasing hydrophobicity (Elkhyat et al., 2001; Kownatzki, 2003; Barel et al., 2009). To our knowledge, this is the first study to examine the effects of virus species, inoculum size, and skin condition on virus transfer between skin and a surface. 2.3 2.3.1 Materials and Methods Volunteers Permission of the Stanford University Research Compliance Office for Human Subjects Research was obtained prior to the study. Volunteers included 8 males and 12 females, with an age range of 20-32 years. To standardize unwashed state of volunteers’ hands, volunteers washed their hands for 15 seconds using soap and water at least 1.5 hours before the experiment, and avoided eating or going to the restroom within that time frame. No brand or type of soap was recommended or provided, and no effort was made to account for residual effects of soap products used before the experiment. 2.3.2 Virus and Preparation of Inoculum This study quantifies transfer of three different bacteriophage (MS2, fr, and φX174) obtained from the American Type Culture Collection (ATCC). MS2 (ATCC #15597B1), fr (ATCC #15767-B1), and φX174 (ATCC #13706-B1) strains were chosen because they have similar size (19-27 nm) and shape (icosahedral, no tail) to several CHAPTER 2. VIRUS TRANSFER 26 human viruses, such as norovirus (Abbaszadegan et al., 2007). MS2 and fr bacteriophage are both +-sense RNA viruses of the Leviviridae family, with similar surface characteristics but different isoelectric points (3.9 and 8.9, respectively) (Gerba, 1984; Liljas et al., 1994; Dowd et al., 1998; Herath et al., 1999). φX174 is a single stranded DNA virus of the Microviridae family with an isoelectric point of 6.6 (Gerba, 1984; Dowd et al., 1998). The inoculum used in the study was prepared by propagating the model viruses to a concentration of 108 -1010 plaque forming units (PFU)/ml in phage buffer (Reddy et al., 2006). The propagated virus was then enumerated and diluted to approximately 105 to 106 PFU/ml using tryptic soy broth (TSB, pH of 7.2 ± 0.2) to be used as virus stock. TSB is an organic-rich media intended to act as a model for the broad range of matrices in which respiratory and gastrointestinal viruses contaminate fomites (e.g. vomitus, urine, feces, mucus, and saliva). Use of homogeneous and well-characterized TSB was intended to reduce variability introduced by use of natural media such as fecal suspensions, mucus, or saliva. The virus stock was enumerated during every experiment to confirm titer. 2.3.3 Plaque Assay The double agar layer procedure was used to enumerate virus (USEPA, 2001). The hosts were Escherichia coli K12-3300 (ATCC #19853) for fr, E. coli HS(pFamp)R (ATCC #700891) for MS2, and E. coli CN-13 (ATCC #700609) for φX174. The double agar layer procedure was chosen to estimate the fraction of infective virus transferred on contact. 2.3.4 Virus Transfer To determine the amount of virus transferred on contact between a fingerpad and a nonporous glass surface, we used a protocol adapted from Ansari et al. (1991). Specifically, we inoculated either between 100 and 600 or between 1000 and 6000 PFU diluted in TSB on the donor surface in an aliquot of 5 µl to represent low and high titers, respectively. Borosilicate coverslips are uniform, smooth, and clean CHAPTER 2. VIRUS TRANSFER 27 surfaces providing a proxy for non-porous fomites with consistent characteristics. All surfaces, including the fingerpads, were subsequently allowed to visibly dry while supervised by the technician before contact between surfaces was made to mimic drying after natural contamination events. To verify that the inoculate remained on the fingerpads, the volunteer was supervised during the visible drying. All samples from which no virus could be recovered from either the donor or recipient surface following the transfer event were removed from analysis. The volunteer placed the donor and recipient surfaces in contact for 10 ± 1 s with an average constant pressure of 25 kPa (range of 16 kPa-38 kPa) controlled by counterbalancing a triple beam balance weighted to 500 g. 25 kPa is comparable to the pressure exerted by a child while gripping an object, the pressure exerted locally on the fingerpads for adults using handtools, and the pressure used in studies examining transfer of soil from surfaces to skin (Link et al., 1995; Hall, 1997; Ferguson et al., 2009). We used a cotton-tipped swab applicator, wet in 500 µl of phosphate buffer saline (PBS, 1 mM potassium phosphate monobasic, 155 mM sodium chloride, and 3 mM sodium phosphate dibasic, pH of 7.4 ± 0.05, from Invitrogen, Carlsbad, CA), to remove virus from the surfaces. The applicator was wiped firmly against the surface in a sweeping, rotating, motion for 10 s before being placed back into the remaining PBS and vortexed for 10 seconds. We used separate swabs to remove virus from the donor and recipient surfaces. Samples were aliquoted into 100 µl of 100 , 10−1 , and 10−2 dilutions in PBS; the dilutions were assayed using the double agar layer method (USEPA, 2001). The range of detection for this method is 10 PFU to 200000 PFU. If virus was unrecoverable from a surface, the lower detection limit of 10 PFU was used as an estimate for the virus recovered. The fraction transferred (f ) is defined as PFU recovered from the recipient surface (RR ), relative to PFU recovered from the sum of the donor (RD ) and the recipient surfaces, as previously described (Rusin et al., 2002): f= RR (RR + RD ) (2.1) Dessication, or the drying of the inoculum on the surface, results in a loss of virus titer (Ansari et al., 1988, 1991; Rusin et al., 2002). Because the surface is dried prior CHAPTER 2. VIRUS TRANSFER 28 to the transfer event, the seeded inoculum is higher than the sum of virus recovered from donor and recipient surfaces after the transfer event. We chose to calculate f using just recoverable virus from the donor and recipient surfaces (Equation 2.1) so that the virus inactivated by dessication is not included in the fraction of virus transferred estimated by Equation 2.1. We assume that the relatively short time of the contact event and subsequent hand and surface sampling does not contribute to loss in titer due to inactivation. The experimental design varied factors including low/high titer and direction of transfer, with blanks and replicates across the 10 fingerpads. Four randomly chosen fingerpads were assigned the following four titer/direction-oftransfer factor combinations: (1) low titer/glass-to-fingerpad, (2) low titer/fingerpadto-glass, (3) high titer/glass-to-fingerpad, and (4) high titer/fingerpad-to-glass. Four additional fingerpads were assigned the same factor combinations. As all factor levels of the fingerpads of the first set were identical to the factor levels of the fingerpads on the second set, the second set of contact events are defined as replicates for the contact events from the first set. In this manner, every contact event had a corresponding replicate contact event. The remaining two fingerpads (one on each hand) were selected to act as blanks. A blank is defined as a transfer event where fingerpad or glass was inoculated with TSB that did not contain any virus. After the initial 10 transfer events were completed, the volunteers washed their hands for 15 s usc antibacterial liquid hand soap (Colgate-Palmolive, New York, NY), ing Softsoap c scientific cleaning wipe (Kimberlyrinsed in tap water, and dried with a Kleenex Clark, Irving, TX) under the technician’s instruction. We then used the same factor assignments for each fingerpad to measure transfer for the ’washed’ hands. Twenty volunteers performed the experiment using MS2 bacteriophage, thirteen of the twenty volunteers repeated the experiment using φX174 bacteriophage, and ten of the thirteen repeated a third time using fr bacteriophage. Ten volunteers completed all 3 experiments. Temperature and relative humidity were recorded from a thermometer and hygrometer (Springfield Precision Instruments, Wood Ridge, NJ) kept at the sampling location. CHAPTER 2. VIRUS TRANSFER 2.3.5 29 Statistics All statistics were performed using the R statistical software package (R: A Language for Statistical Computing, version 2.9.0, R Foundation for Statistical Computing, Vienna, Austria). Where appropriate, descriptive statistics (mean, median, and standard deviation) are reported. Statistical significance was assessed using a significance level of α = 0.05. The significance of experimental factors (direction of transfer, gender, virus species, time since last handwash, and titer) on percent of virus transferred was assessed using n-way ANOVA on untransformed data. Tukey’s post-hoc test assessed significant differences between the transfer of each phage type. Distribution parameters for normal, lognormal, and Weibull distributions are reported for the data on fraction virus transferred between surfaces (f ) stratified by phage type. These distributions are used to describe microbial and/or chemical transfer (Chen et al., 2001; Beamer, 2007; Pẽrez Rodrı́guez et al., 2007). Five-fold cross-validation and Kolmogov-Smirnoff methods were used to determine distribution parameters and goodness-of-fit. 2.4 2.4.1 Results Virus Transfer f was quantified for 656 transfer events. Eleven transfer events (<2% of total transfers) of the original 688 were excluded because of a laboratory error (e.g. mislabeling and failure to add host) involving at least one of the two samples (donor or receipient surface). An additional twenty-one transfer events (<3% of total transfers) were excluded because virus could not be recovered from both donor and recipient surfaces after the transfer. All blanks were negative, implying fingerpads were not contaminated prior to study and no cross-contamination occurred during inoculation. Aggregating data for all three virus species, ranged from 0.001 to >0.999 with a median, mean, and standard deviation of 0.18, 0.23, and 0.22, respectively. Median, mean, and standard deviation of f were 0.32, 0.31, and 0.20, respectively, for fr; 0.18, 0.23, 0.21, respectively, for MS2; and 0.09, 0.19, 0.24 for φX174. CHAPTER 2. VIRUS TRANSFER 30 An n-way ANOVA investigated treatment effects on f . Gender (p = 0.42) and titer (p = 0.79) were not significant. Direction of transfer (p = 0.01) and time since last hand wash (p = 0.002) were significant, with glass-to-fingerpad and unwashed hands transferring a greater fraction than fingerpad-to-glass and washed hands, respectively. Additionally, virus species was significant (p < 0.001). f was larger for fr than for both MS2 (Tukey’s test p < 0.001) and φX174 (p < 0.001). f was not significantly different between MS2 and φX174 (p = 0.16). The mean, median, and standard deviation of f are presented in Table 2.1 grouped by significant factors (e.g., glassto-washed finger transfer of MS2 bacteriophage, unwashed finger-to-glass transfer of fr bacteriophage, etc). Parameters describing the distribution of f were determined for normal, lognormal, and Weibull distributions and are available, with estimates of goodness-of-fit, separated by virus species, in Table 2.2. Virus species impacts not only mean f , but also the best-fit distribution; MS2 and φX174 are right-skew while fr bacteriophage has a more left-skew distribution. As evidence, histograms of the data with corresponding best fit probability density functions are provided in Figure 2.1, separated by virus species and direction of transfer. Temperature and relative humidity ranged from 20-22◦ C and 45-60%, respectively, over the course of the study. No statistically significant correlation (using Spearman’s correlation coefficient) between temperature and f was found for fr (ρs = 0.06, p = 0.47), MS2 (ρs = 0.05, p = 0.65), or φX174 (ρs = −0.02, p = 0.75) or between relative humidity and f for fr (ρs = 0.08, p = 0.31), MS2 (ρs = −0.06, p = 0.57), or φX174 (ρs = 0.04, p = 0.59). 2.5 Discussion We demonstrate that viruses are readily transferred between skin and a model fomite surface. Aggregating 656 viral transfer events, the mean fraction of virus transferred, f , is 0.23 ± 0.22 (mean and standard deviation), consistent with previous studies on virus transfer (Ansari et al., 1991; Mbithi et al., 1992; Rusin et al., 2002) and may be applicable as transfer estimate for viruses of similar size and shape, such as CHAPTER 2. VIRUS TRANSFER 31 norovirus. The relatively large sample sizes of volunteers and contact events provide robust data to estimate distributions to describe f , an important parameter needed for quantifying microbial risk (Gibson et al., 1999; Nicas and Sun, 2006; Wein and Atkinson, 2009), especially in models that utilize activity data (Julian et al., 2009). f is influenced by the virus species, the direction the virus is transferred (i.e., fingerpadto-surface or surface-to-fingerpad), and the characteristics of an individual’s skin, in particular whether or not the hands have recently been washed. Although statistically significant, the factors we identified as influential may change the fraction of virus transferred by, at most, only 5-10%. This is small relative to the effect of changing the porosity of the fomite surface which has been shown to shift f by as much as 2 orders of magnitude (Scott and Bloomfield, 1990; Rusin et al., 2002). Although the contribution of fomites relative to other transmission routes in perpetuating disease burden remains uncertain, the present study suggests it is specific to the etiological agent and ameliorated through frequent hand washing. Virus species affects both the mean and distribution of f . Our work expands on the work of Ansari et al. (1991) who observed transfer differences between two human viruses using 18 total transfer events, by measuring over 600 transfer events with three different viruses. Our high number of observed transfers allowed rigorous statistical testing of treatments. Our results also demonstrate that f is influenced by the interaction of virus species and direction of transfer (Table 2.1). In other words, f depends on the direction of transfer, but precisely how well depends on viral species. This is consistent with observations described in the literature. Ansari et al. (1991) demonstrated human parainfluenza type 3 virus transfer is greater from fomite-to-fingers than fingers-to-fomite, while Mbithi et al. (1992), using hepatitis A virus, demonstrated the reverse: greater transfer from fingers-to-fomite than fomiteto-fingers. Washing fingerpads prior to a virus transfer event reduces f . The reduction in virus transfer due to washing is greater for fingerpad-to-glass transfer than glassto-fingerpad transfer. Changes in moisture level and pH on skin from handwashing (Gfatter et al., 1997), or other residual effects from the soap may contribute to this effect. To investigate the causal mechanism of reductions in f due to hand washing, CHAPTER 2. VIRUS TRANSFER 32 future studies could incorporate moisture and pH measurements of the volunteers’ fingerpads. The impact of hand washing with soap and water on reduction of gastrointestinal and respiratory illness is well documented (Aiello et al., 2008), and is generally attributed to the reduction of pathogenic bacteria and virus on the hands (Curtis et al., 2000; Pickering et al., 2010). The results suggest that reduced viral transfer during hand-surface contacts could also contribute to illness reduction. Further study of virus transmission may elucidate whether or not this finding extends to field conditions. The influence of virus species on f could be due to the physicochemical properties of the virus. The surfaces, suspension media, and contact mechanics were kept constant throughout the study, and the experiments were carried out in ambient laboratory conditions such that temperature and humidity varied over small, but realistic, ranges. Because the viruses were the same shape (icosahedral), we attribute the observed differences in f between virus species to the different sizes (19-27 nm) and chemical properties of the virus capsids. In this experiment, the bacteriophage studied (MS2, φX174, and fr) have different net surface charge, as evidenced by the different isoelectric points (3.9, 6.6, and 8.9, respectively) (Dowd et al., 1998; Herath et al., 1999) and different hydrophobicities. Specifically, φX174 was identified as the most hydrophilic and MS2 as the most hydrophobic in a study of 13 virus species by Shields and Farrah (2002); fr was not tested. Further research in this area is warranted. Neither gender, inoculum size, temperature, nor humidity significantly influenced f . Significant differences in skin characteristics due to gender, such as pH, have previously been documented but the differences are small (pH of male skin was 4.7, female skin was 5.0) (van de Vijver et al., 2003). This difference in pH was not large enough to affect viral transfer in the present study. Inoculum size also did not significantly influence f , in contrast to previous work with bacteria that showed inoculum size significantly influenced bacterial f over multiple orders of magnitude (Montville and Schaffner, 2003). Perhaps the range of titer we explored (one order of magnitude) was too low to observe an effect. Similarly, as neither temperature nor CHAPTER 2. VIRUS TRANSFER 33 relative humidity were explicitly investigated in this study, the range in temperature (20-22◦ C) and relative humidity (45-60%) may have been too small to observe an effect on f . There are several limitations to our study design. We minimized inter-trial variability by using glass surfaces, controlling for duration and pressure of contact, and using the same group of volunteers. In field conditions, such as when an individual contacts a virus-contaminated surface, variation may be greater as transfer events occur between a wide range of surfaces over a range of durations and contact pressures. The use of an infectivity assay (the double agar layer method) does not provide information on non-infective virus particles transferred on contact. Similarly, one plaque forming unit may be more than one infective viral particles (Galasso and Sharp, 1962). Accounting for the presence of non-infective virus particles or multiple infective virus particles in one plaque may alter the fraction of infective virus transferred. Future studies could incorporate molecular methods to better understand transfer influence of non-infective particles and multiple virus per plaque forming unit on transfer. 2.6 Acknowledgments This work was supported, in part, by the Shah Research Fellowship of Stanford University and by the United States Environmental Protection Agency (EPA) under the Science to Achieve Results (STAR) Graduate Fellowship Program. EPA has not officially endorsed this publication and the views expressed herein may not reflect the views of the EPA. The authors acknowledge the volunteers who participated in this study. Additionally, the authors thank the Boehm Lab, Robert Canales, Francisco Tamayo, and the anonymous reviewers who assisted with the work and/or provided suggestions for improving the manuscript. CHAPTER 2. VIRUS TRANSFER 2.7 Tables 34 CHAPTER 2. VIRUS TRANSFER Phage MS2 Direction Handwash n Finger-to-glass Unwashed 75 Washed 75 Glass-to-finger Unwashed 75 Washed 80 φX174 Finger-to-glass Unwashed 49 Washed 50 Glass-to-finger Unwashed 48 Washed 47 fr Finger-to-glass Unwashed 36 Washed 40 Glass-to-finger Unwashed 40 Washed 40 35 µ̂ median σ̂ 0.24 0.18 0.24 0.15 0.10 0.16 0.25 0.19 0.23 0.26 0.21 0.19 0.26 0.16 0.28 0.17 0.14 0.17 0.21 0.07 0.29 0.11 0.04 0.18 0.28 0.25 0.21 0.20 0.19 0.16 0.37 0.39 0.22 0.39 0.40 0.11 Table 2.1: The number of trials (n), mean (µ̂), median, and standard deviation (σ̂) of f for data subset by factors determined to be significant via n-way ANOVA (virus species, direction of transfer, and skin condition as determined by time since last handwash) CHAPTER 2. VIRUS TRANSFER 36 Normal Phage Type MS2 φX174 fr All Phage µ̂ σ̂ 0.23 0.22 0.19 0.24 0.31 0.20 0.23 0.22 p-value 0.09 0.03 0.45 <0.01 Lognormal µ̂ σ̂ -2.1 1.4 -2.6 1.5 -1.6 1.1 -2.1 1.4 Weibull p-value shape 0.18 0.43 0.14 <0.01 0.96 0.77 1.4 0.94 scale p-value 0.22 0.16 0.34 0.23 0.12 0.84 0.66 0.09 Table 2.2: The parameters (mean (µ̂), standard deviation (σ̂), shape, and scale) and goodness-of-fit for fitting normal, lognormal, and Weibull distributions to the fraction of virus transferred as determined by 5-fold cross validation. Parameters and goodness-of-fit are determined for each bacteriophage individually, and all bacteriophage aggregated CHAPTER 2. VIRUS TRANSFER 2.8 Figures 37 CHAPTER 2. VIRUS TRANSFER Glass-toFingerpad fr MS2 (a) (b) n= 99 n= 150 (e) (c) n= 76 (f) n= 95 0.0 0.2 0.4 0.6 0.8 1.0 ALL n= 155 0.0 0.2 0.4 0.6 0.8 1.0 Normal (d) n= 325 (g) n= 80 0.0 0.2 0.4 0.6 0.8 1.0 5 4 3 2 1 (h) 5 n= 330 Density Fingerpad -to-Glass φX174 38 4 3 2 1 0.0 0.2 0.4 0.6 0.8 1.0 Fraction Transferred Weibull Lognormal Figure 2.1: Histogram of f for (a) φX174 fingerpad-to-glass, (b) MS2 fingerpad-toglass, (c) fr fingerpad-to-glass, (d) all bacteriophage fingerpad-to-glass, (e) φX174 glass-to-fingerpad, (f) MS2 glass-to-fingerpad, (g) fr glass-to-fingerpad, and (h) all bacteriophage glass-to-fingerpad. The probability density function is overlaid on each histogram using the parameters reported in Table 2.2 Chapter 3 A Model of Exposure to Rotavirus from Nondietary Ingestion Iterated by Simulated Intermittent Contacts The results presented in this chapter originally appeared as a research article in the May 2009 issue of the journal Risk Analysis (Julian et al., 2009). Robert A. Canales contributed extensively to the modeling and statistical analysis presented and is a co-author on the publication. James O. Leckie and Alexandria B. Boehm also appear as co-authors, for their contributions to study design, data interpretation, and manuscript improvements. 39 CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 3.1 40 Abstract Existing microbial risk assessment models rarely incorporate detailed descriptions of human interaction with fomites. We develop a stochastic-mechanistic model of exposure to rotavirus from nondietary ingestion iterated by simulated intermittent fomes-mouth, hand-mouth, and hand-fomes contacts typical of a child under six years of age. This exposure is subsequently translated to risk using a simple static doseresponse relationship. Through laboratory experiments, we quantified the mean rate of inactivation for MS2 phage on glass (0.0052/s) and mean transfer between fingertips and glass (36%). Simulations using these parameters demonstrated that a childs median ingested dose from a rotavirus-contaminated ball ranges from 2 to 1,000 virus over a period of one hour, with a median value of 42 virus. These results were heavily influenced by selected values of model parameters, most notably, the concentration of rotavirus on fomes, frequency of fomes-mouth contacts, frequency of hand-mouth contacts, and virus transferred from fomes to mouth. The model demonstrated that mouthing of fomes is the primary exposure route, with hand mouthing contributions accounting for less than one-fifth of the childs dose over the first 10 minutes of interaction. CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 3.2 41 Introduction Viral agents transmitted primarily via the fecal-oral route, including enteric adenovirus, astrovirus, norovirus, and rotavirus, are responsible for 35% of hospitalizations for gastroenteritis, accounting for almost 5% of all child hospital visits in the United States (Malek et al., 2006). Enteric viruses have been detected on indoor surfaces and fomites in hospitals, day care centers, hotels, and houseboats (Keswick et al., 1983; Green et al., 1998; Cheesbrough et al., 2000; Jones et al., 2007). Evidence of the role of fomites in disease transmission includes the ability of the etiological agents to transfer between hands and fomes (Ansari et al., 1988) and between fomes and mouth (Rusin et al., 2002), and their ability to persist on fomes and hands (Hall et al., 1980; Casewell and Desai, 1983; Ekanem et al., 1983; Butz et al., 1993; Abad et al., 1994; Cheesbrough et al., 1997; Das et al., 2002; Clay et al., 2006). Despite evidence of the importance of fomites in the spread of disease, few quantitative microbial risk assessment models have examined their role in transmission of disease (Gibson et al., 1999; Nicas and Sun, 2006). The sporadic and sequential nature of multiple individual contacts between hands and fomites, hands and mouth, and fomites and mouth has generally not been considered in microbial exposure assessments. Instead, human interaction with fomites is modeled using estimates of the probability that a contact event occurs (e.g., 10% chance a fomes is contacted by hand), a constant frequency of the contact event (e.g., mouth is contacted by hand 0.08 times per minute), or a constrained sequence of events (e.g., fomes touches hand, hand then touches mouth) (Gibson et al., 1999; Chen et al., 2001; Gibson et al., 2002; Haas et al., 2005; Nicas and Sun, 2006; Nicas and Best, 2008). In the latter, frequently used in quantitative risk assessment as it pertains to food handling, researchers assume that contacts are inevitable, only one contact event of each type occurs, and the temporal sequence is static. In the present study, we further the understanding of the role of human temporal sequence is static. In the present study, we further the understanding of the role of human interaction with fomites on exposure to infectious agents by incorporating modeled sequential, intermittent contacts to encompass a wide array of activity levels. Previous work CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 42 in chemical risk modeling has suggested that using modeled sequential, intermittent contact events reduces uncertainty in estimates of exposure to fomites (Ferguson, 2003). Chemical risk models have been developed that account for an individuals sequential contacts with fomites using micro-level activity data. These data are obtained typically through the use of videotapes by transcribing the timing and sequence of an individuals interactions with environmental surfaces that can act as fomites (Ferguson et al., 2006). With these data in hand, along with concentrations of chemical residues on objects and knowledge of the ability of chemicals to transfer from objects to hands and mouth, a modeler is able to estimate chemical dose (Zartarian et al., 1995; Ferguson, 2003; Ferguson et al., 2006). The present study draws on this chemical risk model framework to conduct a microbial risk assessment of a childs interaction with a rotavirus contaminated ball in an indoor environment (e.g., a child care center). Using micro-level activity data allows us to examine the influence of sequential contact events on a childs exposure to rotavirus and subsequently estimate the risk of infection. After experimentally determining inactivation rates of virus on a surface and the transfer efficiencies of virus between a surface and human hands, we formulate a stochastic-mechanistic model of risk from nondietary ingestion of rotavirus resulting from fomes-mouth, hand-mouth, and hand-fomes contacts. The model is novel in that it uses variable and sequential microlevel activity data to quantify exposure to a contaminated fomes. One of the overarching goals of this study is to determine which model parameters need to be further studied so that more precise exposure assessments can be performed. 3.3 Model Description The stochastic-mechanistic model was developed with MATLAB (version 7.0; The Mathworks, Inc., Natick, MA, USA). The model estimates an individual’s viral dose over the specified time period by incorporating both direct contact between the mouth and a contaminated fomes and indirect contact between the mouth and the fomes via CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 43 intermittent hand-fomes/hand-mouth contacts. Discrete, mechanistic equations iterated by contact event relate an individual’s micro-level activity data (e.g., fomes-hand, fomes-mouth, and hand-mouth contact frequency) to virus-specific exposure factors (e.g., surface area contacted, hand-fomes viral transfer efficiency) and dose-response parameters to estimate exposure, dose, and risk of adverse health outcomes. A Monte Carlo sampling generates model parameter inputs from defined probability distributions, allowing the incorporation of parameter uncertainty. Parameter uncertainty is defined here to include both uncertainty and variability. The model output includes temporal concentration profiles for the fomes, left hand, and right hand and characterizes an individual’s cumulative and iterative risk from continued interaction with the fomes. Equations used in the model define an infectious virus (hereafter referred to as “virus”) as being in one of five states, as depicted in Figure 3.1: (1) located on the fomes, (2) located on the right hand, (3) located on the left hand, (4) irreversibly inactivated, and (5) absorbed in the facial membrane as dose. At the start of each model simulation, the fomes is contaminated with a uniform surface concentration of virus. Additionally, the individual’s hands and mouth are assumed free of virus. The movement of virus between the states occurs either through inactivation (states 1 → 4, 2 → 4, and 3 → 4) or through transfer of virus via contact (states 1 ↔ 2, 1 ↔ 3, 1 → 5, 2 → 5, 3 → 5). Viral inactivation is assumed to decay exponentially with time, causing virus to move from states 1, 2, and 3 to state 4: Cx (tx ) = Cx0 e(−kx tx ) (3.1) where Cx (tx ) with units virus/cm2 is the concentration of virus on surface x (e.g., fomes or hand) at time t, Cx0 is the initial concentration of virus on the surface (virus/cm2 ), kx is the inactivation rate of the virus on the surface (s−1 ), and tx is the elapsed time(s). The transfer of virus between surfaces upon contact is modeled by assuming that CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 44 transfer is driven by a concentration gradient: CX = CX0 − AXY T EXY (CX0 − CY 0 ) (3.2) where CX is the concentration of virus on surface X after contact with surface Y (virus/cm2 ), CX0 and CY 0 are the virus concentrations on surfaces X and Y , respectively, prior to contact (virus/cm2 ), T EXY is the percentage of virus transferred from surface X to surface Y (%), and AXY is the ratio of surface area of the contact event between surfaces X and Y to the total surface area of surface X (cm2 /cm2 ). The transfer is assumed to occur instantaneously and uniformly, and the duration of contact is assumed to not affect transfer. The latter is based on the work of Cohen Hubal et al. (2008), who found that duration does not increase the amount of both lipophilic uvitex and nonlipophilic riboflavin tracer residues transferred between surfaces on contact (Cohen Hubal et al., 2008). It is assumed that, after transfer, virus is distributed evenly over the entire surface. Dose (D) is the number of virus that transfer from a surface to the mouth and depends on contact area between the surface and the mouth, as follows: D = Sx T Exf Cx (3.3) where Sx is the contact area between object x and the mouth (cm2 ), T Exf is the percentage of virus transferred from the object to the mouth, and Cx is the concentration of virus on object x (virus/cm2 ). The mouth is assumed to be an absorbing state, as previously described (Nicas and Sun, 2006) so virus in contact with the mouth is instantly absorbed into the body. A dose-response curve for rotavirus is used to determine the likelihood of adverse health outcome (Haas et al., 1999; Teunis et al., 1999). Because our model results in multiple ingested doses from subsequent fomes-mouth and hand-mouth contacts, we assume that the likelihood of an adverse health outcome is determined from the additive effect of multiple subsequent exposures, as follows (Haas et al., 1999): RT OT = f I X Di i=1 + J X j=1 Dj + K X k=1 Dk (3.4) CHAPTER 3. ROTAVIRUS EXPOSURE MODEL f (D) = 1 − 1 + 45 −α D 1 (2 α − 1) N50 (3.5) Here, RT OT is the likelihood of adverse health outcome(%), f is the dose-response function (Equation 3.5), which is unique for rotavirus (Haas et al., 1999). I, J, and K, are the total number of fomes-mouth, right hand-mouth, and left hand-mouth contacts resulting in a dose event, respectively. Di , Dj , and Dk , are the doses (in units of virus) resulting from the ith fomes-mouth, jth right hand-mouth, and kth left hand-mouth contacts, respectively. 3.5 is a beta-Poison function with shape (α) and scale (N50 ) parameters equal to 0.265 and 5.597 plaque-forming units (PFU), respectively (Haas et al., 1999). The model accounts for both direct and indirect transmission routes. Direct transmission describes mouth-fomes contacts that transfer virus from the fomes to the mouth (Equation 3.3). Indirect transmission describes hand transfer as intermediary between the fomes and the mouth, with hand-fomes contacts transferring virus to the hand (Equation 3.2), and subsequent hand-mouth contacts resulting in dose (Equation 3.3). Viral transfers are modeled as discrete contact events occurring at intervals tF M , tRM , tLM , tRF , and tLF (s) describing subsequent fomes-mouth, right hand-mouth, left hand-mouth, right hand-fomes, and left hand-fomes contacts, respectively. Viral inactivation continuously occurs on surfaces and hands, albeit at different rates (kf and kh , respectively). 3.4 3.4.1 Methods and Materials Parameter Estimation The model parameters used to estimate a child’s dose due to interaction with a contaminated fomes include the initial concentration of virus on surface (Ci ), inactivation rates of virus on surfaces (kf , kh ), percentage of virus transferred between contacted surfaces (T Eom , T Eoh , T Ehm ), length of time between contact events (tF M , tRM , tLM , tRF , tLF ), surface area of fomes and hands (Af , Ah ), and surface area of contacts (Sf , Sm , Sh ). For each parameter, we provide estimates, with justification, of values in CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 46 the form of approximated distributions and associated parameters. The distributions and range of associated parameter values are summarized in Table 3.1. Initial Concentration of Virus on Surface Though multiple studies have demonstrated the presence of rotavirus genomic RNA on indoor environmental surfaces using assays that require at least 100 rotavirus per sample to detect, no work has quantitatively determined the concentration of rotavirus on surfaces (Keswick et al., 1983; Wilde et al., 1992; Butz et al., 1993; Soule et al., 1999) To reflect the uncertainty in the initial concentration of virus on a fomes and the potential variability of the severity of contamination events, we used a uniform distribution with minimum and maximum parameters of 0.001 and 10 virus/cm2 . Inactivation Rates on Surfaces Two inactivation rate parameters are required: rate of viral inactivation on dry environmental surfaces, kf , and rate of viral inactivation on hands, kh . Experimental studies using MS2 phage as a surrogate for pathogenic virus were performed to estimate inactivation rates on environmental surfaces, kf . Glass slides (1 × 2.5 cm2 ) were inoculated with 107 PFU MS2 phage suspended in tryptic soy broth (TSB). Borosilicate glass was chosen to represent a nonporous material and was prepared by washing in soap and water, wiping with 70% ethanol, rinsing in distilled water, and air-drying. TSB was used as the suspension media to include potential effects of particle shielding, though previous studies have demonstrated no significant difference in the persistence of virus on fomites due to suspension media (Abad et al., 1994). After inoculation, the surface samples were kept in 6-well plates at 20◦ C with 65% humidity in the dark to provide a conservative estimate for viral inactivation on typical indoor environments. The surfaces were swabbed with cotton-tipped applicators wetted in 500µl of phosphate buffered saline (PBS) at increasing intervals for a period of 50 days. The wetted cotton-tipped applicators were vortexed in microcentrifuge tubes for 15 seconds in 500µl PBS, and 100µl of this was assayed using the double agar layer method (USEPA, 2001). After swabbing, the surfaces were rinsed in 5 mL of PBS CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 47 for 15 minutes, and 100µl of the rinse was assayed to determine the concentration of recoverable infective phage that were not removed by the cotton-tipped applicator. The log of the total recovered (C) PFU over the log of initial count (C0 ) was plotted as a function of time (t). Linear regression was used to determine the inactivation rate (kf ) such that the concentration of viable, infective phage removable from the surface of a fomes follows exponential decay (Equation 3.1). We assume that the viral inactivation rate on hands, kh , follows an exponential decay similar to previously reported inactivation rates for viruses on surfaces (Boone and Gerba, 2007). However, kh has been shown to be greater than inactivation rates on surfaces, possibly due to temperature and moisture or chemicals on the skin (Ansari et al., 1988). Ansari et al. (1988) demonstrated a 93% reduction in rotavirus titer on the surface of the skin over more than four hours, and this was used to estimate kh . Percent Viral Transfer Between Surfaces Three parameters describing viral transfer between surfaces are used: transfer between fomes and mouth (T EF M ), fomes and hand (T EF H ), and hand and mouth (T EHM ). To estimate percent transfer during fomes and hand contacts, MS2 phage was used as a surrogate virus in laboratory studies. Borosilicate glass and fingertips were used as proxy surfaces. Four fingertips from each of 10 volunteers were inoculated with low (∼2×103 PFU) or high (∼2×104 PFU) titers of MS2 phage suspended in TSB using a micropipettor. After the inoculation was allowed to dry, the fingertips were placed against a glass surface for 10 seconds with an average constant pressure of 25 kPa (range 16-38 kPa). The process was repeated with four glass surfaces inoculated to represent surface-to-hand transfer. For both directions of transfer, a fifth surface (either finger or glass) was inoculated with PBS, representing a blank control. Cotton-tipped swab applicators wetted in 500 µL of PBS were used to remove phage from both the glass surface and the fingertip. The samples were stored at 4◦ C and, within 48 hours, were vortexed for 20 seconds and enumerated using the double agar layer technique (USEPA, 2001). This resulted in a total of 80 samples, not including blanks, in four categories: low titer hand-to-surface, high titer hand-to-surface, low titer surface-to-hand, and high titer surface-to-hand. The transfer of phage between CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 48 surfaces was quantified as the count of recoverable infective phage from the recipient surface as a percentage of the total recoverable infective phage from both the recipient and the donor surfaces (Rusin et al., 2002). Micro-Level Activity Data Five parameters are used in the model to describe a child’s discrete contact events during his/her interactions with a toy: the time intervals between subsequent fomesmouth, right hand-mouth, left hand-mouth, right hand-fomes, and left hand-fomes contacts (tF M , tRM , tLM , tRF , and tLF , respectively). A recent study determined that the Weibull distributions best describe the frequency of contact event data (Xue et al., 2007). As such, the time intervals in this study are described using the Weibull distributions (Law and Kelton, 1997). To determine the durations between subsequent hand-toy (tRF and tLF ) and mouth-hand contacts (tLM and tRM ), we use data of children’s interactions with their environment, as previously collected and described (AuYeung et al., 2006; Ferguson et al., 2006). The data were collected by videotaping one- to six-year-old children in both indoor and outdoor environments for 2-hour time c software (SamaSama periods. The videotapes were translated, using VideoTraq Consulting, Sunnyvale, CA, USA), into second-by-second accounts of the contact events between a child’s right hand, left hand, and mouth and into 36 object categories for 20 children. The data set was subjected to quality control, as previously described (Ferguson et al., 2006). The resulting micro-level activity data provide detailed descriptions of a child’s object contacts necessary for modeling the complexities of intermittent contaminant loading and removal (Beamer, 2007). We assumed that the time intervals between a child’s hand contacts with the object “Hard Toy” (defined as any hard, nonporous toy) could serve as a valid proxy for repetitive contacts with a toy ball. In total, 1,340 right-hand and 1,433 left-hand contact events of 15 of the 20 children were used to develop the Weibull distributions for tRF and tLF . The time intervals for the remaining five children were used to cross-validate the Weibull distributions using the Kolmogorov-Smirnoff test for goodness of fit. The same method was used to determine the time intervals between hand-mouth (tRM and tLM ) and fomes-mouth (tF M ) contacts. The number of data points used CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 49 to create the Weibull distribution were 129, 127, and 43 for right hand-mouth, left hand-mouth, and fomes-mouth contacts, respectively. 3.4.2 Model Approach The model is a discrete-time model, iterated by contact event. First, a sequence of events describing fomes-mouth, right hand-mouth, left hand-mouth, right handfomes, and left hand-fomes contacts is simulated using a Monte Carlo sampling from the interval distributions described by the respective parameters. An example of a simulated sequence is shown as a time series in Figure 3.2, with solid vertical lines representing contacts between the specified hand and the fomes, unfilled circles representing contacts between the hand and mouth, and filled circles representing contacts between the fomes and mouth. Once a sequence of contacts is generated, the initial and final concentrations of virus on each surface (left hand, right hand, and fomes) are determined for each contact event using sampled virus-specific exposure factors, dose-response parameters, and the specific equations in Appendix A, formulated from Equations 3.1, 3.2, and 3.3. From this information, temporal exposure, dose, and risk profiles are generated and metrics of interest are recorded. Examples of exposure and dose profiles as a function of time are presented in Figure 3.3. The illustration of the left hand in Figure 3.3 was omitted for simplicity. Vertical dashed lines represent the timing of hand-fomes contacts, and vertical solid lines represent the timing of hand-mouth contacts. 3.4.3 Sensitivity Analysis The sensitivity analysis method used in this study was previously described by Xue et al. (2006) Briefly, the model is run using single-point parameter values to investigate the sensitivity of the model to variations in a given parameter. The model is run twice, with the value of one specified parameter set first to the 25th percentile (p25) and then to the 75th percentile (p75) of its probability distribution while all other parameters remain set to their median values, and the model output is calculated. The median, p25, and p75 values are used as normative values to describe distributions CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 50 because the model relies on multiple, different probability distribution functions. The effects of parameter variation on output are investigated by calculating a ratio of the cumulative dose resulting from the use of the p75 to the cumulative dose resulting from the p25. The ratio of the results (p75:p25) quantifies the sensitivity of the model to the parameter over the middle 50% of the probability distribution. If p75:p25 is equal to 1, then the model output is unaffected by variations of the parameter. A ratio greater than 1 demonstrates that increases in parameter value, either from the p25 to the median or from the median to the p75, increase the cumulative dose by a factor equal to the ratio. A ratio less than 1 demonstrates a decrease in the cumulative dose by a factor equal to the inverse of the ratio. To track the changes in model sensitivity to a given parameter as the length of time of child-fomes interaction increases, we investigate the temporal change in the p75:p25 ratio for each parameter. The parameters are ranked by the order of influence on the cumulative dose by comparing the absolute values of the log of the p75:p25 ratio. 3.5 3.5.1 Results Parameter Estimation Inactivation The inactivation rate, kf , with 95% confidence interval was determined to be 0.0052 ± 0.0014/h for the representative nonporous surfaces at ambient conditions (20◦ C and 55-65% relative humidity). This value is within an order of magnitude of previously reported viral inactivation rates for rotavirus p13, astrovirus (serotype 4), and hepatitis A (Boone and Gerba, 2007). From Ansari et al. (1988), the inactivation rate of rotavirus on hands, kh , with 95% bootstrapped confidence interval was estimated to be 0.27 ± 0.03/h. This value is within an order of magnitude of other studies investigating microbial inactivation on the skin for other organisms (Musa et al., 1990; Traore et al., 2002) and is used in this study as an estimate for viral inactivation on the skin. CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 51 Transfer Efficiency The transfer from glass to hand was determined to be 36%, with a standard deviation (SD) of 26%, and the transfer from hand to glass was determined to be 27%, with a SD of 23%. The Kruskal-Wallis test showed no statistical difference (p > 0.05) in these populations, demonstrating insufficient evidence that the direction of transfer influences percent transferred. Thus, our model assumes that the percentage of viral transfer is direction-independent (Nicas and Sun, 2006). Additionally, there is insufficient evidence that virus transferred between surfaces is dependent on the initial viral titer, as demonstrated by a Kruskal-Wallis test for significance (p > 0.05). Pooling the data, T EF H used in this model is represented by a normal distribution, with a mean of 32% and a SD of 25%. Although the mean is similar to reported values for viral transfer (Ansari et al., 1988; Mbithi et al., 1992; Rusin et al., 2002), the spread of this distribution is greater than previously reported for viral transfer, but is similar to the values for lipophilic and nonlipophilic compounds (Cohen Hubal et al., 2008). The transfer of virus between hand and mouth (T EHM ) was estimated using a study by Rusin et al. (2002) examining PRD-1 phage transfer from fingertips to lips (Rusin et al., 2002). Laboratory experiments investigating the transfer from 20 volunteers resulted in a mean transfer of 41% of recoverable phage onto the lips, with no SD reported. We assume that the distribution spread (SD) for each percent transfer parameter is similar to our experimentally determined distribution for T EF H , and therefore we assume a SD of 25% for T EHM as well. To our knowledge, no work has yet investigated the amount of virus that transfers directly between a fomes and mouth (T EF M ). We assume that this transfer (T EF M ) has a similar distribution as the transfer of virus between hand and mouth: a normal distribution with a mean of 41% and a SD of 25%. Micro-Level Activity Data Using data from the videotapes, the time between subsequent right and left handfomes contact events, tRF and tLF , is modeled as aWeibull distribution with scale and shape parameters 33 seconds and 0.62 for the right hand and 32 seconds and 0.63 for CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 52 the left hand, respectively. The Kolmogorov-Smirnoff goodness-of-fit test indicates that the Weibull distribution is sufficient for describing 17 of 20 childrens right handfomes contacts and 16 of 20 childrens left hand-fomes contacts (p > 0.001). The times between subsequent right and left hand-mouth contact events, tRM and tLM , are modeled as a Weibull distribution with scale and shape parameters of 270 s and 0.47 for the right hand (p > 0.05 for 20 of 20 children) and 420 s and 0.61 for the left hand (p > 0.05 for 20 of 20 children), respectively. The time between subsequent fomes-mouth contact events, tF M , is modeled as a Weibull distribution with scale and shape parameters 140 s and 0.41 (p > 0.05 for 20 of 20 children), respectively. Surface Area Parameters describing the total surface area of the hand (AH ) and the fomes (AF ) are required. Additionally, surface areas of contact between the fomes and mouth (SF ), the fomes and the hand (SH ), and the hand and mouth (SM ) are needed. We estimated total surface area of a childs hand (AH ) as uniformly distributed with a range of 270–390 cm2 , based on calculations using data available from the ChildSpecific Exposure Factors Handbook (interim report) (Tulve et al., 2002; USEPA, 2006). The toy ball is given a diameter of 9–11 cm, resulting in a uniform distribution for surface area of the fomes (AF ), with a range of 250–380 cm2 . The literature values were used for both surface area of contact between a fomes and a hand (SH ) and surface area of contact between a hand and mouth (SM ) for children playing with toys outdoors. The surface area between a fomes and a hand on contact were observed to fall within the range of 8–27% of total hand surface area (AuYeung, 2007). For this case study, a uniform distribution using the 5th and 95th percentiles (13–24%) of that range as endpoints is used, and we assume that surface area of childs contact with outdoor toys is a sufficient proxy for surface area of childs contact with indoor toys (AuYeung, 2007). Similarly, the 5th and 95th percentiles of SM were estimated as a range of 6–33% of total hand surface area (AuYeung, 2007). We assume a uniform distribution using these values as endpoints,and that hand-mouth contacts outdoors act as a sufficient proxy for similar childs contacts indoors. To our knowledge, no data are available on the percentage of surface area of a CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 53 round ball contacted during fomes-mouth contacts. Therefore, we make a conservative estimate that the surface area is similar to the surface area of childs contacts between hand and mouth (6–33%). This estimate is conservative because, though we assumed similar surface areas for the round ball fomes and the hand, less surface area will likely be contacted during a mouthing event for a large round ball than for small, irregularly shaped fingers and hands. 3.5.2 Model Results Temporal Exposure, Dose, and Risk Estimates Prototypical profiles of right- and left-hand exposures and virus concentration on the fomes are provided in Figure 3.4. Virus concentration decreases on the fomes as inactivation and continued hand-fomes and mouth-fomes contacts remove virus. Conversely, virus concentration initially increases on the hands, as the presumed original state of the hands is virus free. As described by Equation 3.2, the difference in the concentration of virus between the fomes and hands drives the transfer of virus between the two surfaces, forcing an eventual pseudoequilibrium between the surfaces. Once this is reached, hand- and fomes-mouth contacts, combined with inactivation, remove the virus from the fomes and hands, causing the concentrations to decrease gradually. Figure 3.5 displays the estimated cumulative dose and corresponding risk as a function of time. The median cumulative dose increases approximately linearly with the length of time the child interacts with the contaminated fomes, ranging from a median dose of 13 virus (corresponding risk of infection (RI) of 60%) during 10 minutes of interaction to 42 virus (RI 70%) during one hour of interaction. As rotavirus has a low median infectious dose, or dose at which half of individuals exposed will experience adverse health effects, of 5.6 PFU (Haas et al., 1999), the majority of risk occurs within the first 10 minutes. This would not be true for other pathogenic agents transmitted via the fecal-oral route such as Shigella and enteropathogenic Escherichia coli, which have higher median infectious doses of 103 and 107 , respectively (Haas et al., 1999), The 5th and 95th percentiles of dose, using the specified probability CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 54 distributions for each parameter, are 0 (RI = 0%) and 430 (RI = 84%) rotavirus with 10 minutes of child-toy interaction to 2 (RI = 35%) and 1,000 (RI = 87%) rotavirus with one hour. The virus concentration on the surface of the fomes is reflective of the degree of severity of the contamination event, so we explored the dose and risk of illness as a function of initial concentration on the fomes (Figure 3.6). As demonstrated in the results of 1,000 model simulations of 10 minutes of child-fomes interaction with best-fit functions (Figures 3.6a and 3.6B), there is large variability in the resulting dose and risk of illness for a given initial virus concentration. Despite this variability, the median dose linearly increases with the initial concentration on the fomes (Figures 3.6a and 3.6c). The dose and corresponding risk as a function of initial concentration on the fomes for simulations between 10 and 60 minutes of child’s interaction with the fomes fits the beta-Poisson function (Figures 3.6b and 3.6d), with a shape parameter (α) similar to that used for the beta-Poisson doseresponse model (Equation 3.5). The beta-Poisson scale parameter (equivalent to N50 ) describes the concentration on the fomes for which there is a 50% risk of illness (Haas et al., 1999). The scale parameter decreases with increasing child-fomes interaction time (from 0.3 PFU/cm2 at 10 minutes to 0.04 PFU/cm2 at 60 minutes), suggesting the risk of illness from a given virus contamination increases the longer a child plays with the toy, consistent with the previous observation given in Figure 3.5. The relative importance of direct (fomes → mouth) and indirect (fomes → hand → mouth) transmission of viral pathogens was explored (Figure 3.7). As a child initially interacts with the fomes, the fomes-mouth contacts contribute more than 80% of dose, demonstrating that a child’s direct mouthing of a toy is the most likely route of viral transmission. As a child continues playing with a toy, the virus on the fomes is transferred to the hands, and indirect transmission via mouthing hands contributes more to the child’s total dose. Therefore, the proportion of total dose from mouthing a toy decreases, and the proportion from mouthing hands increases. Because the majority of the ingested dose occurs within the first 10 minutes, fomesmouth contacts contribute to the majority of a child’s risk of adverse health effects. This may not be true for other pathogenic agents with higher median infectious doses CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 55 than rotavirus. 3.5.3 Sensitivity Analysis Model-estimated dose relies on the stated assumptions concerning input parameter values and distributions. Thus, we emphasize the relative importance of input parameters over absolute model output values (Zartarian et al., 2000) by examining the parameter’s influence on model output through a sensitivity analysis. The sensitivity analysis (Table 3.2) demonstrated that the parameters that most influence a cumulative dose after 10 minutes of interaction are, in order: (1) initial virus concentration on the surface of the fomes (Ci ), (2) frequency of fomes-mouth contacts (tF M ), (3) frequency of right hand-mouth contacts (tRM ), (4) transfer of virus between fomes and mouth (T EF M ), (5) frequency of left hand-mouth contacts (tLM ), (6) percentage of the fomes that contacts the mouth on fomes-mouth contacts (SF ), (7) surface area of fomes (AF ), and (8) percent transfer of virus between hand and mouth (T EHM ). The commonly used Spearman correlation sensitivity analysis (Gibbons, 1985; Siegel, 1988) supports these findings (data are not shown). The sensitivity analysis method allows investigation of the changing influence of each parameter over time (Table 3.2). As the child continues to interact with the toy, the ratio of the dose resulting from the p75 to the p25 (p75:p25) value of a parameter changes to reflect the parameters changing influence on resulting dose. For example, the importance of fomes-mouth viral transfer decreases as child-fomes interaction increases (Figure 3.8A). This further supports the finding that direct contact between the fomes and mouth is the primary exposure route within the first 10 minutes, but that the indirect contacts between the hand and fomes and the hand and mouth become increasingly important as the child continues interacting with the toy. Similarly, the influence of the frequency of fomes-mouth contacts on resulting dose decreases temporally (Figure 3.8B). The model also demonstrates that handfomes viral transfer, examined over the duration of the child-fomes interaction from 10 to 60 minutes, has little influence over resulting dose. When the percentage of virus transferred on hand-fomes contacts is varied between 18% and 54%, the range CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 56 between the p25 and the p75, the change in dose is low. Presumably, this is because the expected value of handfomes contact frequency (0.05/s) is an order of magnitude larger than the median frequency of hand-mouth contacts (0.006/s). That is, there are almost 10 hand-fomes contacts made for each hand-mouth contact. Because multiple hand-fomes contacts occur, the virus concentration on the hands and fomes reach equilibrium, regardless of the percentage of virus transferred on each individual contact. 3.6 Implications We use micro-level activity patterns in a mechanistic-stochastic model of dose to more fully understand the role of fomites in pathogen transmission. We simulated a single individuals interactions with a contaminated fomes, demonstrating the ability to model pathogen transmission on a contact-by-contact basis. Previous work incorporating human-environment interactions has demonstrated the importance of sequential contacts in understanding microbial exposure and risk from specific activities (Gibson et al., 1999; Chen et al., 2001; Gibson et al., 2002; Haas et al., 2005; Nicas and Sun, 2006; Atkinson and Wein, 2008). We further this work by demonstrating the importance of modeling a wider range of activity level by incorporating stochastic simulations of activity into microbial exposure assessment. Together with previous studies investigating bacterial and viral transmission via contacts, this study provides a foundation for incorporating human-environment interaction in dynamic infectious disease models used to describe population-based pathogen transmission for fecal-oral diseases (Elveback et al., 1971; Nasell, 2002; Stone et al., 2007). An overarching goal of our work is to identify model parameters that require further study to improve future risk assessments. The model parameter most strongly linked to estimated dose is, unsurprisingly, the concentration of virus on the fomes.Our use of a uniform distribution for the initial concentration bounded by parameters differing by over four orders of magnitude was motivated by not only a lack of quantitative data describing distributions of viral contamination on indoor surfaces but also a desire to examine model output over the full range of plausible values. As CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 57 Laborde et al. (1994) found, the concentrations of indicator bacteria on surfaces in child care centers regularly differed by over four orders of magnitude, emphasizing the need to understand the exposure and risk resulting from interactions with surfaces over a range of contamination levels (Laborde et al., 1994). Clearly, as the initial concentration increases, one would expect the dose and risk of illness to increase as well (Figure 3.6). Nevertheless, there is a need for further evaluation of the quantity and spatial distributions of viral contamination on indoor surfaces, particularly for surfaces such as toys that are most likely to act as fomites. This finding also reflects the importance of reducing the presence of virus on surfaces to reduce or eliminate fomes-mediated disease transmission. The average time between the fomes and mouth contacts was a significant contributor to model output and should be investigated further. The importance of this parameter decreased temporally as the child continued interacting with the toy. This is explained by the increased contribution of right hand-mouth and left hand-mouth contacts to dose (Figure 3.7). Implementing an intervention to reduce fomes-mouth contacts would result in a reduction in an individuals dose. A decrease of 50% of the rate of fomes-mouth contacts, with all other parameters unchanged in the stochastic model, reduced the median dose by 31%, with a corresponding reduction of risk for rotavirus of 4%, after 10 minutes of child-fomes interaction. However, modifications in a childs behavior may be difficult or impossible to implement. The amount of viral transfer between surfaces may be an example of a parameter that could be readily modified, as different toy surface properties or environmental conditions may influence the amount of virus transferred between surfaces. As demonstrated, the transfer of virus between fomes and mouth is the fourth most influential parameter in determining dose and risk. Reduction in the transfer between fomes and mouth by 50%, without any changes in the values or distributions of the other variables in the stochastic model, reduces the median dose of the simulations by 22%, with a corresponding reduction of risk of 2%, after 10 minutes of interaction. This highlights the importance of better understanding the effect of environmental conditions on viral transfer between surfaces to reduce the incidence of disease, as environmental conditions are easily modified via the use of humidifiers/dehumidifiers CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 58 and thermostats. Additionally, the use of toys with surface properties that inhibit or reduce viral transfer during periods when viral outbreaks are most likely to occur may reduce or slow the transmission of disease (Nicas and Sun, 2006). Finally, the model demonstrates that viral inactivation plays little or no role in determining an individuals dose over the time scale of our model. Reduction of viral inactivation on both fomes and hands by 100% (i.e., assuming no viral inactivation) resulted in a change in the median dose of one viral particle, with corresponding difference in risk of only 0.5% after 10 minutes of interaction. After 60 minutes of interaction, assuming no viral inactivation, the median dose changed by only four viral particles, with corresponding difference in risk of 0.5%. This result applies only to ambient environmental conditions and the time scale of this model, which focuses on child’s interaction with the fomes at a temperature of 20◦ C and relative humidity of 65% over a period of at most 60 minutes. Changes in environmental factors significantly alter inactivation rates of virus on surfaces, with rapid increases in inactivation rates at a higher relative humidity and temperature (Ansari et al., 1991). Viral persistence on surfaces likely influences transmissibility via fomites on a time scale more closely aligned with the time scale of inactivation rates, for example, days or weeks (Boone and Gerba, 2007), and at more extreme environmental conditions, for example, relative humidity >85% and temperature >30◦ C (Ansari et al., 1991). The results of the model rely on assumptions and simplifications. For example, we assumed uniform distribution of viral agents on both fomes and hands, before and after contacts. This assumption, common in chemical exposure modeling and in investigation of aerosol dispersion of pathogenic agents (Nazaroff et al., 1998; Zartarian et al., 2000; Liao et al., 2008), clearly impacts the resulting simulated risk of adverse health effects. Future work should examine uneven pathogen distribution on surfaces. Additionally, the model does not explore the effects of environmental conditions, such as temperature and humidity, on the transmission and inactivation of virus. As temperature and humidity have recently been implicated in the seasonal patterns of influenza transmission (Lowen et al., 2007), future work could focus on discerning the effect of temperature and humidity on inactivation and transfer to elucidate their role in the transmission of fecal-orally transmitted viruses. CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 59 Examination of an individual’s interactions with his/her environment to assess exposure to infectious disease lays the groundwork for incorporating activity and virusspecific exposure factors into broader, secondary transmission models. Although the model incorporates individual’s contact events and virus-specific exposure factors such as transfer of virus between surfaces and viral inactivation rates in disease transmission, the model is limited to assessing a static examination of a single individual’s risk. The dose-response model used to determine an individual’s risk was based on data obtained from healthy adults; we did not account for the uncertainty associated in applying this model to children to determine the risk of illness (Ward et al., 1986). Similarly, the model does not account for immunity or interaction between multiple children, which would enable the dynamic modeling of secondary transmission via person to person or person to fomes to person (Abbey, 1952; Elveback et al., 1971; Nasell, 2002; Lawniczak et al., 2006; Giraldo and Palacio, 2008). Despite the limitations of the model, we elucidate the roles of an individual’s contact events, viral inactivation, and viral transfer on an individual’s risk of adverse health effects. This mechanistic-stochastic model of microbial dose incorporates contact-by-contact human-environment interactions and can therefore serve as a basis for future highresolution microbial risk assessment. 3.7 Acknowledgments The authors acknowledge the volunteers, members of the Boehm and Leckie research groups, and anonymous reviewers who assisted with the work and/or provided suggestions for improving the article. This publication was supported by the Stanford University Shah Research Fellowship and the STAR Research Assistance Agreement No. F07D30757 awarded by the U.S. Environmental Protection Agency (EPA). It has not been formally reviewed by the EPA. The views expressed in this article are solely those of the authors, and the EPA does not endorse any products or commercial services mentioned in this article. CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 3.8 Tables 60 kf kh AF AH tF M tRF tLF tRM tLM SF SH SM T EF M T EF H T EHM Inactivation Rate Fomes Hand Area of Object Fomes Hand Contact Frequency Fomes and Mouth Right Hand and Fomes Left Hand and Fomes Right Hand and Mouth Left Hand and Mouth Percent of Object Contacted Fomes (Fomes-Mouth Contact) Hand (Fomes-Hand Contact) Hand (Hand-Mouth Contact) Percent Transferred Between Fomes and Mouth Between Fomes and Hand Between Hand and Mouth % % % % % % s s s s s cm2 cm2 1/s 1/s virus/cm2 Units (140, 0.41) (33, 0.62) (32, 0.63) (420, 0.59) (270, 0.47) Normal (0.41, 0.25) Normal (0.36, 0.26) Normal (0.41, 0.25) Uniform (0.06,0.33) Uniform (0.13,0.24) Uniform (0.06,0.33) Weibull Weibull Weibull Weibull Weibull Uniform (250,380) Uniform (270,390) Normal (1.4×10−6 , 2.0×10−7 ) Normal (7.5×10−5 , 4.3×10−6 ) Uniform(0.001,10) Distribution (Parameters) Study Study Study Study Study Assumption This Study Rusin et al. (2002), Assumption Assumption AuYeung (2007) AuYeung (2007) This This This This This Assumption USEPA (2006) This Study Ansari et al. (1988) Assumption Source/Reference Table 3.1: Parameters Used in Model, with Corresponding Distributions and Median Values for Determining Exposure and Dose Distributions Ci Symbol Virus Concentration Initial Concentration Variable Description CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 61 0.001 1.27×10−6 7.21×10−5 283 300 6.50×10−4 5.10×10−3 5.50×10−3 4.80×10−4 4.60×10−4 0.13 0.16 0.13 0.24 0.18 0.24 Virus concentration Initial concentration Inactivation rate Fomes Hand Area of object Fomes Hand Contact frequency Fomes and mouth Right hand and fomes Left hand and fomes Right hand and mouth Left hand and mouth Percent of object contacted Fomes (fomes-mouth contact) Hand (fomes-hand contact) Hand (hand-mouth contact) Percent transferred Between fomes and mouth Between fomes and hand Between hand and mouth 0.41 0.36 0.41 0.20 0.19 0.20 1.60×10−3 1.25×10−2 1.32×10−2 1.20×10−3 1.10×10−3 315 330 1.40×10−6 7.50×10−5 0.1 p50 0.58 0.54 0.58 0.26 0.21 0.26 3.20×10−3 2.50×10−2 2.60×10−2 2.40×10−3 2.20×10−3 348 360 1.53×10−6 7.79×10−5 10 p75 1.9 0.9 1.3 1.6 1.0 1.2 3.0 1.0 1.0 1.2 1.2 1.2 1.0 1.0 1.0 10000 10 min 1.6 0.9 1.3 1.5 1.0 1.3 2.3 1.0 1.0 1.4 1.3 1.3 1.0 1.0 1.0 10000 20 min 1.5 0.9 1.4 1.4 1.0 1.3 2.1 1.0 1.0 1.4 1.3 1.3 1.0 1.0 1.0 10000 30 min 1.5 1.0 1.4 1.4 1.0 1.3 2.0 1.0 1.0 1.4 1.4 1.3 1.0 1.0 1.0 10000 40 min p75:p25 1.4 1.0 1.5 1.3 1.0 1.4 1.9 1.0 1.0 1.4 1.4 1.3 1.0 1.0 1.0 10000 50 min 1.4 1.0 1.4 1.3 1.0 1.4 1.8 1.0 1.0 1.4 1.4 1.3 1.0 1.0 1.0 10000 60 min 8 3 4 5 6 2 7 1 Rank Table 3.2: Sensitivity Analysis Results for Cumulative Dose (Number of Virus) During Increasing Length of Time of Child-Fomes Interaction p25 Variable Description Input Values CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 62 CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 3.9 Figures 63 CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 64 Figure 1 4: Inactivated 2: Right Hand 5: Facial Membrane/ Dose 1: Fomes 3: Left Hand Legend - Viral Inactivation - Transfer by Contact Figure 3.1: The relationships between the five potential reservoirs for virus represented by this model. At time 0, the fomes is the only contaminated object, and the right and left hands are free of virus. The arrows represent the possible pathways between the states. CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 65 Figure 2 Right Hand Left Hand Fomes 0 Legend 100 200 300 - Hand-Fomes Contact 400 500 600 Time (s) - Hand-Mouth Contact 700 800 900 1000 - Fomes-Mouth Contact Figure 3.2: Example of timing for randomly generated sequence of contact events between left hand, right hand, fomes, and mouth. Vertical solid lines represent handfomes contacts, unfilled circles represent hand-mouth contacts, and filled circles represent fomes-mouth contacts. CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 66 Figure 3 Right Hand Conc. (Ch) Fomes Conc. (Cf ) 200 150 100 100 50 20 40 60 80 150 100 100 50 40 60 80 Dose (D) 20 Time 0 Legend 0 0 - Hand-Fomes Contact - Hand-Mouth Contact - Dose Figure 3.3: Example of trends of concentration, exposure, and dose over time, simulated from randomly generated sequence of contact events, describing interaction between right hand, fomes, and mouth. Vertical dashed lines represent right handfomes contacts, and vertical solid lines represent right hand-mouth contacts. Each hand-fomes contact results in virus transfer between fomes and hand. Each handmouth contact results in virus transfer from hand to mouth. Different inactivation rates (kh and kf ) continuously decrease viral concentration on hand and fomes. CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 67 Figure 4 10 Fomes Right Hand Left Hand 2 Virus Concentration (virus/cm ) 10 8 8 6 6 4 4 2 2 0 0 0 10 10 20 20 30 30 40 40 50 50 0 6060 Length of Time of Child-Fomes Interaction (min) Figure 3.4: Example of modeled concentration and exposure profiles for fomes, right hand, and left hand, demonstrating the temporal change in concentrations. CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 68 Cumulative Ingested Dose (virus) Figure 5 1200 1000 800 600 400 200 0 10 20 30 40 50 60 10 20 30 40 50 60 Risk of Illness (%) 100 80 60 40 20 0 Time of Child-Fomes Interaction (min) Figure 3.5: Modeled distributions of (top panel) dose and (bottom panel) risk of infection from 10,000 simulations of child-fomes interaction after specified interaction time. Boxes depict the median, 25th percentile, and 75th percentile. Whiskers represent the 5th and 95th percentiles. Figure 6 CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 1200 (a) 1000 1.0 0.9 800 (b) 0.8 0.7 Simulation Results Best Fit Line 69 Simulation Results Best Fit Line 0.6 0.5 600 0.4 400 0.3 0.2 0 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 (c) 1200 1000 800 600 60 min 50 min 40 min 30 min 20 min 10 min 0.1 0.0 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 1.0 0.9 0.8 0.7 0.6 0.5 0.4 400 0.0 0.5 0.0 0.5 1.0 (d) 60 min 50 min 40 min 30 min 20 min 10 min 0.3 0.2 200 0 -3.0 Risk of Illness Ingested Dose (virus) 200 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 0.1 0.0 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 1.0 Initial Virus Concentration on Fomes (Log(virus/cm2 )) Figure 3.6: Dose and risk of illness as a function of initial concentration of virus on fomes. Best-fit lines for (a) and (c) are calculated using the linear relationship of resulting dose from the initial concentration on fomes from 10,000 model simulations. Best-fit lines for (b) and (d) are calculated using the beta-Poisson distribution resulting from the risk of illness as a function of initial concentration on fomes from 10,000 model simulations. (a) Dose as a function of initial concentration resulting from 1,000 model simulations plotted with the best-fit line for 10 minutes of child-fomes interaction demonstrates the variability of modeling results and distribution around the best-fit line. (b) Risk of illness as a function of initial concentration resulting from 1,000 model simulations plotted with the best-fit line for 10 minutes of child-fomes interaction. (c) Dose increases as a function of child-fomes interaction as well as initial concentration on fomes, as shown in 10-minute increments for 10–60 minutes of child-fomes interaction. (d) Risk of illness exhibits increases similar to those of dose as a function of initial concentration on fomes, as shown in 10-minute increments for 10–60 minutes of child-fomes interaction. CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 70 Figure 7 Ingested Dose (virus) 21% 50 20% 30 20 15% 8% 11% 67% 60% 21% 18% 81% s-M Fome 10 0 56% 23% 58% 23% 19% 40 22% 54% 24% outh acts Cont Right Hand - M outh Contacts Left Hand - Mouth Contacts 10 20 30 40 50 60 Time (min) Figure 3.7: Dose contributions over time from left hand-, right hand-, and fomesmouth contacts increase as child-fomes interaction increases, as demonstrated by the line graph. The pie charts demonstrate the percentage of dose contribution from left hand-, right hand-, and fomes-mouth contacts at the specified time, with the contribution from fomes-mouth contacts decreasing as a percentage of the whole as child-fomes interaction increases. CHAPTER 3. ROTAVIRUS EXPOSURE MODEL 71 Ratio of Cumulative Dose (p75:p25) Figure 8 (a) 2.0 2.0 1.5 1.5 1.0 1.0 0.5 0.5 (b) 5 4 3 TE Fomes-Mouth TE Hand-Mouth TE Hand-Fomes 5 t Fomes-Mouth 4 t Hand-Mouth t Hand-Fomes 3 2 2 1 1 0 0 10 20 30 40 50 60 Length of Time of Child-Fomes Interaction (min) Figure 3.8: Sensitivity analysis examining the effect of parameter variation on model output as it changes over time. The ratio of the cumulative dose calculated using the 75th percentile value of the probability distribution to the cumulative dose calculated using the 25th percentile value of the specified parameter. The ratio of the cumulative dose refers to the factor by which it changes when the parameter value is increased for: (a) viral transfer between surfaces, where T EF omes−M outh is the percentage of virus transferred from fomes to mouth per contact, T EHand−M outh is the percentage of virus transferred from hand to mouth per contact, and T EHand−F omes is the percentage of virus transferred between hand and fomes per contact, and (b) the frequency of contact events, where tF omes−M outh is the frequency of fomes-mouth contacts, tHand−M outh is the frequency of right hand-mouth contacts, and tF omes−Hand is the frequency of fomes-right-hand contacts. Chapter 4 Surface Sampling Methods for Virus Recovery From Fomites The results presented in this chapter will be submitted to a peer reviewed journal in Winter 2011. Francisco J. Tamayo contributed to the experimental design and data collection, and will be co–author on the resulting publication. James O. Leckie and Alexandria B. Boehm will also appear as co–authors, for their contributions to study design, data interpretation, and manuscript improvements. 72 CHAPTER 4. VIRUS RECOVERY FROM SURFACES 4.1 73 Abstract The role of fomites in infectious disease transmission relative to other exposure routes is difficult to discern due, in part, to the lack of information on the level and distribution of virus contamination on surfaces. Comparison of outcomes of studies intending to fill this gap is difficult because multiple different sampling methods are employed and authors rarely report their method’s lower limit of detection. In the present study, we demonstrate that the sampling method significantly influences virus recovery from surfaces, and therefore influences study outcomes. We compare sampling methods chosen from a literature review to identify the most efficient method for recovering virus from surfaces in a laboratory trial using MS2 bacteriophage as a model virus. Recovery of virus is determined using both plaque assay and quantitative polymerase chain reaction. From this, we conclude that polyester-tipped swabs prewetted in either 1/4–strength Ringer’s solution or saline solution most effectively remove virus from nonporous fomites. Our results also demonstrate that the recommended sampling method is an appropriate method for quantifying virus on surfaces. CHAPTER 4. VIRUS RECOVERY FROM SURFACES 4.2 74 Introduction Preclusion of infection is the most effective method to combat the respiratory and gastrointestinal diseases that cause over 6 million annual deaths, worldwide (Boone and Gerba, 2007; Mathers et al., 2008). Successful interventions to reduce disease burden include hand and environmental hygiene (Siegel et al., 2007; Bell, 2006), but the impact of these interventions is difficult to quantify because the importance of contact with contaminated surfaces, or fomites, relative to other transmission routes is uncertain (Mubareka et al., 2009; Brankston et al., 2007). Evidence of the importance of fomites comes from both laboratory and field studies. Laboratory studies have demonstrated that handling either artificially–or naturally–contaminated fomites by susceptible hosts indoors results in subsequent infection (Gwaltney, 1982; Hall et al., 1980). Additionally, virus can be transferred between hands and fomites on contact, and survive on fomites for hours or days (Bean et al., 1982; Rusin et al., 2002; Abad et al., 1994). In a field study, environmental hygiene as an intervention significantly reduced illness–related absenteeism in classrooms (Bright et al., 2009). Additionally, fomites, such as carpets (Osterholm et al., 1979; Evans et al., 2002), towels, and medication cart items (Morens and Rash, 1995) have been implicated as the primary cause of multiple outbreaks. Despite this evidence, questions remain regarding relative efficacy of fomite–mediated transmission relative to other exposure routes (Jennings and Dick, 1987; Atkinson and Wein, 2008) and likelihood of virus transfer from fomites to hosts (Pappas et al., 2009). Surface contamination is most often described by the positivity rate, defined as the fraction of total samples collected on which the organism is detectable (Butz et al., 1993). However, the positivity rate does not provide an indicator of infection risk, which depends on exposure magnitude (Haas et al., 1999) and therefore requires information about the quantity of virus on the surface. Virus quantity on a surface, expressed as number of virus or virus equivalents per unit area, has only been measured in a few studies (Bellamy et al., 1998; Russell et al., 2006; Piazza et al., 1987). Moreover, positivity rate is influenced by the sampling method and detection assay: CHAPTER 4. VIRUS RECOVERY FROM SURFACES 75 more sensitive sampling methods and detection assays will yield increases in positivity rates even though the actual level of virus contamination may be unchanged. Use of a sensitive, standard method would limit bias introduced by various sampling methods. Two previous studies have compared virus surface sampling methods and suggested that implement type (the tool used to collect the sample, such as a swab) and eluent type (the liquid used to aid in removal, such as saline solution) significantly influence virus recovery. Carducci et al. (2002) recovered a greater fraction of hepatitis C virus from a seeded surface using beef extract than using bovine serum albumin when swabbing with a cotton–tipped applicator. The study demonstrated that eluent type can significantly impact virus recovery from surfaces. Similarly, Taku et al. (2002) demonstrated the impact of implement type by comparing calicivirus recovery from food surfaces for four sampling methods. Rinsing a surface in 0.05 M glycine buffer, rubbing with a cell scraper, then aspirating the buffer was recommended over: 1) rinsing surface in buffer then aspirating, 2) swabbing surface with cotton–tipped applicator, or 3) swabbing surface with a nylon filter. However, Taku et al. (2002) recommended method is not easily adapted to the geometry of most fomites. Further research is needed to refine implement and eluent choice for sampling fomites to maximize virus recovery. In the present study, we systematically review the literature on virus sampling of fomites and use an extensive laboratory–based trial to compare methods of virus detection on surfaces. We identify, summarize, and analyze 45 articles that include unique data sets on virus detection on surfaces. The most commonly used and most effective sampling methods identified from the meta–analysis are compared in a laboratory–based study for removal of bacteriophage MS2, as measured using culture–based and quantitative reverse–transcription polymerase chain reaction (qRT–PCR), from plastic and stainless steel surfaces. Using both the literature review and experimental results, we identify polyester–tipped swabs prewetted in either 1/4 strength Ringer’s solution (heretofore referred to as “Ringer’s”) or saline solution as the most effective buffer and implement combination to remove virus from nonporous fomites. CHAPTER 4. VIRUS RECOVERY FROM SURFACES 4.3 4.3.1 76 Materials and Methods Review of Virus Surface Sampling Literature Relevant articles were identified by searching the PubMed database on 5 February 2010 with keywords: “virus” and one of the following: 1) “fomite(s)”, 2) “environmental contamination”, 3) “environmental surface(s)”, or 4) “environmental sample(s)” and “surface(s)”. Only articles written in English were considered. Of those identified, the articles included in the review fit the following criteria: 1) included original data collected from environmental surfaces, where clinical (e.g. skin, bodily fluids) and food (e.g. meat, vegetables) surfaces were not considered, and 2) tested samples for human pathogenic virus or fecal indicator bacteriophage (e.g. somatic, F+ bacteriophage). To identify articles not included in PubMed, the citations of the articles fitting the criteria were also reviewed. For analysis, data from the articles were separated into data sets according to the virus and the presence/absence of a clinically infected individual. That is, articles that reported positivity rates for multiple viruses were split into separate data sets for each virus. Similarly, articles that sampled surfaces during periods that encompassed both presence and absence of clinically infected individuals were split into separate data sets for each time period. Samples collected two weeks before and after at least one individual was identified as clinically infected were considered separately than samples collected where no clinically infected individual was present. In this manner, seventy–four data sets from forty–five articles were obtained. Positivity rate was determined, as the outcome variable, for each data set. Positivity rate was the only feasible outcome variable as most (96%) of the studies identified reported presence/absence of virus on surfaces. Only a few (3 of 74, or 4%) reported quantitative data. If the authors included clinical or food samples, those samples were removed. To allow for logit–transformation, the positivity rate for studies that detected the virus on none or all of the samples was adjusted to detection limits of 1/n or (n − 1)/n, respectively, where n is the study’s total number of samples collected. The positivity rate is an inherently biased outcome variable because the lower limit CHAPTER 4. VIRUS RECOVERY FROM SURFACES 77 of detection (LLOD) likely varies across studies for reasons described previously. As few studies (21%) reported either the quantitative concentration of the virus or the LLOD of the sampling method, the positivity rate could not be adjusted to account for the bias. We assessed the influence of the implement and eluent used to collect and analyze the samples on positivity rate. Similar implement and eluents were grouped for data analysis. Polyester and Dacron swabs were both categorized as polyester. The eluent used was categorized into one of four groups: media (defined here as any eluent with a carbon source, and includes Amies medium, beef extract, brain heart infusion broth, Letheen broth, minimum essential medium, RPMI–1640, and tryptose phosphate broth with 0.5% gelatin), saline (defined as any isotonic eluent without a carbon source, and includes phosphate buffered saline, 0.8% saline, and Ringer’s solution), water, or unreported. Additives and constituents of eluents, such as antibiotics, were ignored for data analysis with the lone exception of calcium. We examined whether or not the presence of calcium in the eluent influenced positivity rate, where calcium is present in Ringer’s solution, Amies medium, minimum essential medium, and RPMI–1640. Statistics The positivity rate for each study was logit–transformed, and normality of transformed data was assessed using the Shapiro–Wilk test to support use of parameteric statistics. Two bivariate linear models were used to determine the significance of implement and eluent choice, separately, on transformed positivity rate; positivity rate was weighted by total number of samples in each study. To determine effect of trace calcium in the eluent on positivity rate, a χ2 test for equal proportions was used. All statistics were performed using R (version 2.9.0, R Foundation for Statistical Computing, Vienna, Austria). CHAPTER 4. VIRUS RECOVERY FROM SURFACES 4.3.2 78 Laboratory–Based Surface Sampling Method Comparison In a laboratory–based trial, we compared fraction of virus recovered from surfaces using a subset of the implement and eluent choices identified in the literature as most commonly used and most effective. Virus and preparation of inoculum MS2 bacteriophage was obtained from the American Type Culture Collection (ATCC). MS2 (ATCC #15597–B1) is a +–sense RNA virus with a icosahedral, tailless, capsid of 27 nm in diameter. The isoelectric point (pI) of MS2 is 3.9. MS2 bacteriophage was chosen because of its prior use a surrogate for human viruses, such as norovirus, (Dawson et al., 2005) and the availability of plaque assay and qRT–PCR methods to enumerate both infective phage and copies of nucleic acids (USEPA, 2001; O’Connell et al., 2006). E. coli HS(pFamp)R (ATCC #700891) was used to propagate and enumerate MS2. The inoculum used in the study was prepared according to the polyethylene glycol precipitation method (Pecson et al., 2009). The propagated virus was then enumerated using the double agar layer method and diluted in dilution buffer (5 mM NaH2 PO4 , 10mM NaCl, pH = 7.4) to 1 × 104 PFU/ml to be used as virus stock. Immediately before being seeded on the surface, the virus stock was mixed with tryptic soy broth to form a 50% solution. Implement and Eluents Tested The two most commonly used, and the single most effective (highest mean positivity rate) implements and eluents were identified from the literature for use in the laboratory study. The implements tested included the cotton–tipped and polyester–tipped swabs as the most commonly used (used in 39 and 12, respectively, of the 74 studies) and antistatic cloth as the most effective (positivity rate of 0.408). Similarly, the eluents tested included saline and viral transport media (used in 18 and 9, respectively, CHAPTER 4. VIRUS RECOVERY FROM SURFACES 79 of the 74 studies) as the most common and Ringer’s solution (used in the same article as the antistatic cloth) as the most effective. A fourth eluent, termed acid/base, was added to assess a novel combination adapted from a method to concentrate virus from environmental water samples (Katayama et al., 2002). The acid/base eluent relies on knowledge of the virus surface charge to improve recovery from surfaces. This study is the first use of the eluent combination to remove virus from surfaces. Briefly, a weakly acidic (0.5 mM dihydrogen sulfate (H2 SO4 )) eluent is used to wet the implement prior to sampling. Viruses with low isoelectric points adsorb to negatively–charged surfaces (like cotton) under acidic conditions (Katayama et al., 2002). After sampling, the implement is placed into a weakly basic (1 mM sodium hydroxide, pH 10.5–10.8) eluent which reverses the surface charge of the virus to elute the virus from the implement. Surfaces Tested To determine the method most effective in removing virus from surfaces, we compared recovery from both high temperature polyvinyl chloride (PVC) plastic (Part No. 8748K21) and type 304 stainless steel with a mirror–like finish (Part No. 9785K11), both obtained from McMaster–Carr (Santa Fe Springs, CA, USA). Many of the surfaces identified in the literature review that were frequently contaminated (e.g. door knobs, faucet handles, drains, medical instruments, toys, playmats, computer parts, telephones) were composed of plastic or metal. PVC plastic and stainless steel, in 930cm2 square sheets, were chosen as representative samples as it was infeasible to test every potential surface. 10 replicates for each eluent and implement combination were tested on both surfaces. In total, 240 samples were collected (3 implements, 4 eluents, 2 surfaces, and 10 replicates). All were tested using the double agar layer plaque assay method, and a subset using qRT–PCR. Study Design For both plastic and stainless steel surfaces, a 5 µl inoculum of approximately 4900–5200 PFU bacteriophage was seeded in the center of 120 5 cm × 5 cm surface CHAPTER 4. VIRUS RECOVERY FROM SURFACES 80 swatches. The seeded aliquot was dried for 45 ± 1 minutes under ambient conditions (temperature 20–22◦ C and relative humidity of 45–60%, determined by a thermometer and hygrometer (Springfield Precision Instruments, Wood Ridge, NJ)). The order of implement and eluent combinations used to recover bacteriophage from the surfaces was randomized prior to the start of the study. The polyester and cotton–tipped swabs were obtained from Fisher Scientific (Thermo Fisher Scientific, Waltham, MA, USA). The antistatic cloths were obtained from Bel–Art Products (Pequannock, NJ, USA) and cut into single–ply square swatches of approximately 9 cm2 . The eluents used were Ringer’s solution (EMD Chemicals, Inc, Gibbstown, NJ, USA), 0.85% saline solution, virus transport media (Copan Diagnostics, Murietta, CA, USA), and acid/base. Centrifuge tubes (15ml, BD Biosciences, San Jose, CA) were filled with 1.5 ml of 0.85% saline, viral transport media, Ringer’s solution, or 1 mM sodium hydroxide. To sample, the polyester or cotton–tipped swabs were wetted in the eluent (or in 0.5 mM dihydrogen sulfate, for acid/base) and then rubbed with moderate and consistent pressure across the surface first horizontally, then vertically, then diagonally for a total of 10 s. The swab was then placed into the centrifuge tube, and the tube was capped and stored on ice for 4 hours to mimic typical transportation time. Antistatic cloth, otherwise following the same procedure, was not wetted prior to sampling. After storage, the samples were vortexed for 60 s. An aliquot of 100 µl was used to assay the samples for infective bacteriophage using the double agar layer method (USEPA, 2001). The remaining sample was stored at –80◦ C. qRT–PCR Viral recovery was determined using qRT–PCR from plastic and stainless steel surfaces for only two implement/eluent combinations: cotton–tipped swab in saline solution and polyester–tipped swab in Ringer’s. Cotton/saline was the most common implement/eluent combination used in studies reviewed in the meta–analysis; polyester/Ringer’s was the combination with highest efficacy of recovery measured using the culture–based assay (to be shown). Twenty–eight samples were assayed using qRT–PCR: seven samples for each combination of cotton/saline or polyester/Ringer’s CHAPTER 4. VIRUS RECOVERY FROM SURFACES 81 and plastic or stainless steel surface. RNA was extracted and quantified from 200 µl of sample volume, after storage at –80◦ C for 15–20 days following sample collection. qRT–PCR was performed on the extracts within 6 hours. To extract viral RNA, we used the Invitrogen PureLink Viral RNA/DNA extraction kit (Invitrogen, Carlsbad, CA) according to manufacturer’s instructions using 200 µl samples eluted in 20 µl of DNase/RNase–free water. Genomic RNA was enumerated using the Taqman quantitative reverse–transcption polymerase chain reaction (qRT–PCR) with reagents, primers, and cycling conditions of O’Connell et al. (2006) for a 25 µl reaction with 5 µl template (O’Connell et al., 2006). The location in the MS2 genome of the primers, probe, and target (the sequence of the qRT–PCR amplicon), is the RNA replicase β chain. The forward primer (5’–GCTCTGA- GAGCGGCTCTATTG–3’), reverse primer (5’–CGTTATAGCGGACCGCGT–3’), and probe (5’–[FAM]–CCGAGACCAATGTGCGCCGTG–[TAMRA]–3’) were obtained from Eurofins MWG Operon (Huntsville, AL) (O’Connell et al., 2006). RNA standards were created from total genomic RNA extracted without aid of transfer RNA from a high titer of purified MS2 bacteriophage, enumerated as 20 ng/µl using a NanoDrop ND–1000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and diluted to four standards, at 10–fold dilution, between 0.1 fg/µl (50 genome equivalents / µl) and 10000 fg / µl (50000 genome equivalents / µl). A genome of 3569 nucleotides with average molecular mass of 330 Da per nucleotide was assumed to convert RNA concentration to genome equivalents (O’Connell et al., 2006). qRT–PCR was performed using a StepOne Plus RealTime PCR System (Applied Biosystems, Carlsbad, CA) and all samples and standards were run in triplicate. Statistics Descriptive statistics (mean, median, standard deviation) are provided for recovery of infective bacteriophage from each surface using each method. To determine the implement and eluent choice that most effectively removes bacteriophage from surfaces, an n–way ANOVA with post–hoc Tukey’s test was performed on untransformed CHAPTER 4. VIRUS RECOVERY FROM SURFACES 82 data. Surface sampled, implement, and eluent are the independent variables; fraction of recovered infective bacteriophage is the dependent variable. Significance was determined if the p–value ≤ 0.05. Fraction of target RNA recovered was defined as target copies of RNA enumerated for each sample divided by the number of target copies seeded. Linear regression was used to model the relationship of the number of target copies estimated from qRT–PCR as a function of the number of infective bacteriophage estimated using the double agar layer method, with an intercept set at 0. Variables for surface swabbed (plastic or stainless steel) and for implement/eluent combination were included in the regression model. 4.4 4.4.1 Results Literature Review 310 articles were identified in the keyword search using PubMed. Of those, 40 fit the specified criteria. Review of the citations revealed an additional 5 relevant articles. In total, 45 relevant articles were identified. Eleven of the articles sampled for multiple pathogens and/or during time periods where a clinically infected individual was and was not present. As a result, the articles are treated as 74 separate studies. A summary of the articles, including the separation into studies, implement, eluent, and assay used, positivity rate, and locale are provided in Table B.1. Definitions of abbreviations used in Table B.1 are provided in Table B.2. Authors measured environmental contamination of 20 different etiologic agents, including causative agents of gastrointestinal, respiratory, bloodborne and/or sexually–transmitted diseases. A division of studies by virus, including the number of studies for each, total samples collected, number of samples with detectable virus, and fraction of samples with detectable virus are provided in Table B.3. The positivity rates of the studies, when logit–transformed, were normally–distributed (Shapiro–Wilk test for normality, W = 0.98, p =0.28). In total, 6804 samples were collected with detectable virus on 1105 for an overall positivity rate of 0.162. CHAPTER 4. VIRUS RECOVERY FROM SURFACES 83 In total, twelve different eluents (excluding additives) were used in the 74 identified studies. In 7 (9%) of the studies, the authors did not identify the eluent. Table 4.1 provides a summary of the studies, aggregated by eluent type, and includes the number of samples collected, number with detectable virus, and fraction with detectable virus for each eluent. The authors of thirty–four of the studies (46%) used media, while 29 (39%) used a saline solution. Studies where eluent was unreported were grouped into a “not–reported” category and included in the linear model. The linear model demonstrated no signficant influence of eluent category on positivity rate, when positivity rate was weighted by total samples collected (R2 = 0.05, p =0.33). Additionally, eluents with calcium were not associated with a significantly different fraction of samples with detectable virus (p > 0.05). Four implement types were used in the studies. Studies where implement was unreported or reported as unspecified swab type (23% of identified studies) were grouped into a “not reported” category and included in the linear model. A division of studies by implement type, including the number of studies for each, the total samples collected, number with detectable virus, and fraction with detectable virus for each implement type are provided in Table 4.2. Implement type explained 23% of the variation in positivity rate (R2 = 0.23, p < 0.001) according to the linear model using logit–transformed positivity rate weighted by total sample number as the dependent variable. Compared to cotton–tipped swabs, positivity rate was significantly higher for polyester swabs (p = 0.01) and significantly lower for rayon–tipped swabs (p = 0.02). We found no significant difference between cotton–tipped swabs and antistatic cloths, although antistatic cloths had the highest positivity rate (0.408), likely due to the small sample size. 4.4.2 Laboratory–based Surface Sampling Method Comparison Results of the recovery of MS2 bacteriophage from stainless steel and plastic for each implement/eluent combination are provided in Table 4.3 and Figure 4.1. The mean (µ̂) and standard deviation (σ̂) of the fraction recovered from stainless steel was 0.29 CHAPTER 4. VIRUS RECOVERY FROM SURFACES 84 and 0.17, respectively. Recovery from plastic was similar, with a mean and standard deviation of 0.30 and 0.24, respectively. As demonstrated by n–way ANOVA (Table 4.4), the surface sampled did not significantly influence the recovery of MS2 bacteriophage (p =0.63). Implement choice, however, was significant (p < 0.001). Post–hoc Tukey’s test revealed that antistatic cloths, with overall fraction of recovery of 0.09, were significantly lower than both polyester swabs (mean recovery fraction = 0.40, p < 0.001) and cotton swabs (mean recovery fraction = 0.38, p < 0.001). Recovery using polyester and cotton swabs were not significantly different (p =0.32). Similarly, eluent significantly influenced fraction of bacteriophage recovered (p =0.01). The largest fraction was recovered using Ringer’s solution, with a mean fraction recovered of 0.24, followed by saline (mean recovery fraction = 0.20) and acid/base (mean recovery fraction = 0.19). Viral transport media recovered the lowest fraction of virus (mean recovery fraction = 0.17). According to Tukey’s post–hoc test, the fraction recovered was significantly different only between Ringer’s solution and viral transport media (p =0.005). The interaction effect of implement and eluent combination was not significant (p =0.39). The combination of polyester swab and Ringer’s resulted in the largest mean fraction recovered (mean recovery fraction = 0.48), though it was not significantly different (p < 0.05) from polyester and any other eluent or Ringer’s and any other implement besides antistatic cloth. 4.4.3 qRT–PCR The surface inoculum, approximately 5×103 PFU, was determined to be approximately 2.9×105 target copies (58 target copies per plaque). Twenty–eight samples using implement/eluent combinations of cotton/saline and polyester/Ringer’s were assayed, and five (18%) were removed from analysis because the quantified target RNA exceeded 100% recovery, or 2.9×105 copies, likely due to laboratory error as suggested by high standard deviations in the triplicate samples. Mean and standard deviation of the fraction of target RNA recovered was 0.25 and CHAPTER 4. VIRUS RECOVERY FROM SURFACES 85 0.22, respectively. No product was detected in the blanks and no template controls used in the study. Linear regression of recovered target copies as a function of recovered infective bacteriophage, surface, and implement/eluent combination demonstrated no significant effect due to surface and sampling method on qRT–PCR recovery. Surface did not significantly influence recovery using qRT–PCR (p =0.62). Similarly, although polyester/Ringer’s recovered approximately 18000 greater target copies per sample than cotton/saline, the difference was not significant (p =0.62). The linear model also elucidated the relationship between recovery using qRT–PCR and recovery using a plaque assay. The linear relationship between recovered infective bacteriophage and target copies was significant (p < 0.001) and explained 79% of the variability (R2 = 0.79). Specifically, the ratio of recovered target copies to infective bacteriophage after a dessication step of 45 ± 1 minute was 59.8. This is consistent with previous estimates of the ratio of target copies to infective MS2 bacteriophage as determined using a plaque assay (O’Connell et al., 2006). 4.5 Discussion Indoor surface sampling is necessary to understand the role of fomites in disease transmission. However, studies employ many different sampling methods to recover virus from surfaces. In the present study, we demonstrate through a combined literature review and laboratory trial, that the sampling method significantly impacts virus recovery. In fact, sampling method may contribute to the wide range in positivity rates reported across studies. Standardization of a sampling method to the polyester–tipped swab and 1/4–strength Ringer’s solution or saline solution, may reduce variability and facilitate cross–study comparisons. To reduce influence of sampling method on positivity rate, polyester–tipped swabs should be used for virus detection on surfaces. Both the meta–analysis and laboratory trial demonstrated that polyester–tipped swabs improved recovery relative to other implements. In the meta–analysis the implement choice explained CHAPTER 4. VIRUS RECOVERY FROM SURFACES 86 24% of the variability in the positivity rate, a strong relationship that suggests that implement choice affects study outcomes. The recommendation to use polyester swabs is consistent with the recommendation of the United States Centers for Disease Control and Prevention to use synthetic fibers for clinical sample collection (http://www.cdc.gov/h1n1flu/specimencollection.htm). Cotton–tipped swabs are known to contain trace contaminants (Ellner and Ellner, 1966; Pollock, 1947) with demonstrated bacterial inhibition (Pollock, 1947). Similar interference with virus detection may be possible. Furthermore, the irregular arrangement of cotton fibers reduces elution of bacteria (Osterblad et al., 2003), and could contribute to the observed reduced virus recovery relative to polyester. Antistatic cloths recovered the lowest fraction of seeded virus in the methods comparison study. Antistatic cloths were not prewetted in this study, which likely contributed to the low recovery. As antistatic cloth is composed of synthetic fiber, prewetting may provide fractional recovery similar to polyester–tipped swabs. A potential benefit of using antistatic cloth is that larger surface areas can be sampled. However, this was not specifically addressed in our study. Ringer’s or saline solution should be used as an eluent for virus detection on surfaces. Although the meta–analysis demonstrated no significant differences in positivity rate attributable to eluent category, the laboratory trial demonstrated that Ringer’s recovered the greatest fraction of seeded virus using culture–based method, followed closely by saline solution. The laboratory trial is not meant to be all inclusive as testing all possible implement and eluent combinations, including additives to the eluent such as lecithin or Tween 80, is infeasible. However, based on the combinations tested, Ringer’s or saline solution should be used as eluent in future studies. Trace calcium in the eluent does not impact virus recovery on surfaces. Calcium impacts virus adsorption to surfaces (Fuhs et al., 1985) and reduces nucleic acid detection using PCR by inhibiting the polymerase enzyme (Bickley et al., 2008). However, based on the meta–analysis, there was no significant influence on positivity rate attributable to calcium. Similarly, Ringer’s solution (which differs from saline solution by the addition of potassium and calcium chloride) did not significantly differ from saline in recovery of infective bacteriophage using plaque assay or target RNA CHAPTER 4. VIRUS RECOVERY FROM SURFACES 87 using qRT–PCR. Use of the acid/base eluent method relied on the knowledge of the MS2 bacteriophage isoelectric point (3.9) to aid in recovery. Future studies intending to replicate this method need to consider the isoelectric point of the virus prior to assay development. When assessing infection exposure and risk from environmental contamination, the sampling method’s LLOD is needed (Herzog et al., 2009). Only 13 of the 45 articles reviewed included a quantitative assessment of the LLOD of their sampling method. The lack of a reported LLOD and the reliance on presence/absence data makes cross–comparison of studies and relating positivity rates to risk infeasible. In the present study, the mean and range of fractional recovery of infective bacteriophage using polyester/Ringer’s was 0.48 and (0.20, 0.98), respectively. Using the mean and range of fractional recovery, along with the assumption that the bacteriophage double agar layer method enumerates ≥ 1 PFU, the lower limit of detection is 2.1 with range (1.0, 5.0) PFU per area sampled. Similarly, assuming the qRT–PCR method has a lower limit of quantification of ≥ 250 genome equivalents (O’Connell et al., 2006), consistent with our standard curves, then the theoretical quantification limit is 892 with range (431, 2.5×104 ) genome equivalents based on a mean and range fractional recovery for RNA of 0.28 and (0.01, 0.58). In the future, reporting the LLOD would allow authors to combine dose–reponse curves and positivity rates to exposure and risk estimates (Haas et al., 1999; Julian et al., 2009). Although a standardized sampling method is recommended to allow cross–comparison of studies reporting positivity rates, there may be limitations. The recommendation to use polyester swabs in Ringer’s or saline solution is based on results from both a laboratory–scale study and a review of literature. The laboratory–scale study was based on recovery of one virus (MS2 bacteriophage) and two surfaces (high temperature PVC plastic and type 304 stainless steel with a mirror–like finish). Pathogenic viruses, however, have wide variation in physicochemical properties (such as size, shape, and isoelectric point) that may influence recovery by a standardized method. Similarly, the morphology and composition of the fomites’ surfaces may also influence recovery. PVC plastic and stainless steel are representative samples CHAPTER 4. VIRUS RECOVERY FROM SURFACES 88 of many potential fomites, as both are widely used in consumer products (Heudorf et al., 2007; Adams, 2009). As not all potential fomites are made of PVC plastic or stainless steel, the method recommended here may not be the most efficient recovery method for every virus / surface combination sampled. Nevertheless, a standardized method is recommended for cross–comparison of studies reporting positivity rates. Our findings suggest polyester–tipped swabs with Ringer’s or saline solution perform best. A priority in future research is linking surface contamination to adverse health outcomes. There is currently limited evidence that virus contamination on fomites is linked to increased risk of adverse health outcomes. To address this, longitudinal studies could simultaneously track health outcomes and surface contamination, similar to the work of (Gallimore et al., 2006; Bright et al., 2009; Boxman, Dijkman, Verhoef, Maat, van Dijk, Vennema and Koopmans, 2009), using the recommended sampling method. Additionally, quantifying virus concentrations on surfaces is a priority. Knowledge of virus quantity is an important step toward linking fomites to health risk, as exposures to greater concentrations result in greater risk of infection (Haas et al., 1999). Sampling surfaces with polyester/Ringer’s or polyester/saline, as evidenced by this study, is compatible with quantification of virus using plaque assay or qRT–PCR. Use of a standard method with a known recovery fraction will facilitate extrapolation of measured surface quantities to exposure and risk estimates. 4.6 Acknowledgments We thank the members of the Boehm Lab, who assisted with the work and/or provided suggestions for improving the manuscript. The research has been funded, in part, by the UPS Foundation and the United States Environmental Protection Agency (EPA) under the Science to Achieve Results (STAR) Graduate Fellowship Program, Assistance Agreement No. F07D30757. FT was supported by NSF awards BES–0641406 and SES–0827384. EPA has not officially endorsed this publication and the views expressed herein may not reflect the views of the EPA. CHAPTER 4. VIRUS RECOVERY FROM SURFACES 4.7 Figures 89 Fraction Recovered CHAPTER 4. VIRUS RECOVERY FROM SURFACES 1.0 0.8 0.6 0.4 0.2 0.0 Acid/Base Ringer's 90 Saline VTM ● ● ● ● ● ● ● ● ● A C P A C P A C P A C P Figure 4.1: Fraction of seeded MS2 bacteriophage recovered by implement/eluent combination using the double agar layer method to enumerate plaque forming units. Abbreviations used include: “VTM” for viral transport medium, “A” for antistatic cloth, “C” for cotton, and “P” for polyester–tipped swabs. Boxes depict the 25th, median, and 75th quartiles. Whiskers represent the 10th and 90th percentiles. Outliers are denoted by “•” CHAPTER 4. VIRUS RECOVERY FROM SURFACES 4.8 Tables 91 3 2 2 3 29 18 9 2 4 7 74 MEMc RPMI1640c LB Amiesc Saline Solutions Saline PBS Ringer’sc Water Not Reported Total 6804 120 278 125 420 538 99 218 123 2763 2218 1418 No. Samp. 3643 730 93 424 1105 11 98 51 88 37 29 83 33 367 228 332 No. Pos. Samp. 629 12 3 100 0.162 0.092 0.353 0.408 0.210 0.069 0.293 0.381 0.268 0.133 0.103 0.234 Frac. Pos. 0.173 0.016 0.032 0.236 Gallimore et al. (2006, 2008); Wu et al. (2005); Piazza et al. (1987); Froio et al. (2003); Boone and Gerba (2005); Ferenczy et al. (1989); Lederman et al. (2009); Bellamy et al. (1998); Lessa et al. (2009) Gallimore et al. (2005); Strauss et al. (2002); Butz et al. (1993); Kawahara and Yoshida (2009); Kuusi et al. (2002); Widdowson et al. (2002); Bausch et al. (2007); Lopez et al. (2008) Boxman, Dijkman, Verhoef, Maat, van Dijk, Vennema and Koopmans (2009); Boxman, Dijkman, te Loeke, Hagele, Tilburg, Vennema and Koopmans (2009) Runner (2007) Girou et al. (2008); Lyman et al. (2009); Hamada et al. (2008); CDC (2008) Akhter et al. (1995) Fischer et al. (2008); Carducci et al. (2002) Wilde et al. (1992); Winther et al. (2007); Pappas et al. (2009); Gwaltney (1982) Cheesbrough et al. (2000); Green et al. (1998); Goldhammer et al. (2006); Chen et al. (2004); Booth et al. (2005); Russell et al. (2006); Dowell et al. (2004) Keswick et al. (1983); Gurley et al. (2007); Soule et al. (1999) Asano et al. (1999); Yoshikawa et al. (2001) Boone and Gerba (2005) Jones et al. (2007); Bright et al. (2009) Ref. Table 4.1: Summary of the eluents used in the reviewed articles, including number of studies (“No. Studies.”), number of samples collected (“No. Samp.”), number of samples with detectable virus (“No. Pos. Samp.”), and fraction of samples with detectable virus (“Frac. Pos.”). Eluents with trace calcium are denoted by ’c ’ 9 No. Studies 34 5 2 8 VTM Eluent Media TPB gel BE BHIB CHAPTER 4. VIRUS RECOVERY FROM SURFACES 92 12 4 2 17 74 Polyester Rayon Antistatic Not Reported Total 6804 1165 571 125 1411 No. Samp. 3532 1105 128 36 51 388 No. Pos. Samp. 502 0.162 0.110 0.063 0.408 0.275 Frac. Pos. 0.142 Ref. Cheesbrough et al. (2000); Gallimore et al. (2005, 2006, 2008); Green et al. (1998); Keswick et al. (1983); Wilde et al. (1992); Winther et al. (2007); Gurley et al. (2007); Goldhammer et al. (2006); Wu et al. (2005); Chen et al. (2004); Strauss et al. (2002); Pappas et al. (2009); Asano et al. (1999); Butz et al. (1993); Fischer et al. (2008); Gwaltney (1982); Kawahara and Yoshida (2009); Kuusi et al. (2002); Soule et al. (1999); Widdowson et al. (2002); Piazza et al. (1987); Yoshikawa et al. (2001); Carducci et al. (2002); Froio et al. (2003) Bausch et al. (2007); Boone and Gerba (2005); Dowell et al. (2004); Ferenczy et al. (1989); Lederman et al. (2009); Lopez et al. (2008); Russell et al. (2006) Bellamy et al. (1998); Bright et al. (2009); Jones et al. (2007) Boxman, Dijkman, Verhoef, Maat, van Dijk, Vennema and Koopmans (2009); Boxman, Dijkman, te Loeke, Hagele, Tilburg, Vennema and Koopmans (2009) Akhter et al. (1995); Girou et al. (2008); Hamada et al. (2008); Lessa et al. (2009); Lyman et al. (2009); Runner (2007) Table 4.2: Summary of the implement types used in the reviewed articles, including the number of studies (“No. Studies.”), the total number of samples (“No. Samp.”), the samples with detectable virus (“No. Pos. Samp.”), and the fraction of samples with detectable virus (“Frac. Pos.”) are also provided No. Studies 39 Implement Cotton CHAPTER 4. VIRUS RECOVERY FROM SURFACES 93 CHAPTER 4. VIRUS RECOVERY FROM SURFACES Stainless Steel µ̂ median σ̂ Implement Eluent Cotton Saline 0.38 0.38 0.15 0.33 0.34 0.11 Ringer’s VTM 0.35 0.38 0.13 0.32 0.12 Acid/Base 0.32 Polyester Saline 0.39 0.38 0.17 0.39 0.38 0.13 Ringer’s VTM 0.29 0.30 0.13 0.38 0.17 Acid/Base 0.39 Antistatic Saline 0.15 0.13 0.15 0.16 0.10 0.18 Ringer’s 0.10 0.07 0.12 VTM Acid/Base 0.23 0.24 0.14 µ̂ 0.39 0.54 0.36 0.37 0.39 0.59 0.39 0.48 0.007 0.032 0.009 0.003 94 Plastic median σ̂ 0.45 0.16 0.56 0.16 0.34 0.07 0.33 0.15 0.41 0.12 0.56 0.21 0.37 0.13 0.48 0.11 0.003 0.01 0.003 0.08 0.003 0.01 0.003 0.001 Table 4.3: Summary of the fraction of MS2 bacteriophage recovered using each implement/eluent combination from stainless steel and plastic surfaces. The mean, median, and standard deviation are reported CHAPTER 4. VIRUS RECOVERY FROM SURFACES Effects Surface Implement Eluent Implement:Eluent Residuals d.f. 1 2 3 6 217 Sum of Squares 0.01 4.76 0.24 0.13 4.6 Mean Square 0.01 2.38 0.08 0.02 0.02 95 F–value 0.23 111.5 3.73 1.05 p–value 0.63 <0.001 0.01 0.40 Table 4.4: Fraction of virus recovered from a seeded surface as a function of the surface’s material and the implement and eluent used, based on statistical results of n-way ANOVA Chapter 5 Evidence for Causal Links between Respiratory Illness and Indicator Bacteria on Surfaces in Child Care Centers . The results presented in this chapter will be submitted to a peer reviewed journal in Winter 2011. Amy J. Pickering contributed extensively to the experimental design, data collection, data analysis, and manuscript preparation, and will be co-author on the resulting publication. James O. Leckie and Alexandria B. Boehm will also appear as co-authors, for their contributions to study design, data interpretation, and manuscript improvements. 96 CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 5.1 97 Abstract BACKGROUND: The link between microbial contamination on surfaces and health outcomes has not been fully established. Investigating temporal trends in health and environmental contamination may provide evidence for causal links between surface contamination and adverse health outcomes. OBJECTIVES: The objective is to investigate causal relationships between contamination on hands and surfaces and health in child care centers. METHODS: The present study tracks both respiratory and gastrointestinal disease incidence while monitoring weekly hand and environmental surface contamination over four months in child care centers. Microbial contamination was determined using quantitative densities of fecal indicator bacteria as well as presence/absence of viral pathogens. Health was monitored daily by childcare staff, who tracked adverse health outcomes, including respiratory illness. RESULTS: Symptomatic respiratory illness is significantly and positively associated with hand contamination and with environmental contamination. Detection of enterovirus on hands provides further support of the importance of surfaces in disease transmission. CONCLUSIONS: Symptomatic respiratory illness is both caused by, and causes increases in microbial contamination on hands. Specifically, increases in microbial contamination led to increases in symptomatic respiratory illness four to six days later, in agreement with typical incubation periods for respiratory illness. Respiratory illness also led to increases in microbial contamination on hands during presentation of symptoms. 5.2 Introduction Over 4.6 million children under the age of five years old are enrolled in center-based child care in the United States (Laughlin, 2010). Children attending center-based child care suffer 1.5-3 times more respiratory and 2-3.5 times more gastrointestinal CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 98 episodes per year than those in home-based child care (Fleming et al., 1987; Alexander et al., 1990; Lu et al., 2004). Attendees of center-based care are not the only ones with increased risk of illness. Evidence of the role of children in disseminating disease through communities abounds. Examples include the documented persistence of hepatitis A infections in areas with child care centers (Hadler et al., 1980; Desenclos and MacLafferty, 1993), increased disease prevalence in households with children attending child care (Garrett et al., 2006), and reductions in prevalence within communities following the vaccination of children (Dagan et al., 2005; Loeb et al., 2009). Therefore, reductions in infectious disease transmission within child care centers may influence burden in the community-at-large. Reductions in disease burden are often achieved through interventions tailored to interrupt known transmission routes. Studies have shown that in child care centers, the introduction of hygiene programs significantly reduce gastrointestinal disease by interrupting direct and indirect contact transmission (Krilov et al., 1996; Roberts, Jorm, Patel, Smith, Douglas and McGilchrist, 2000; Lennell et al., 2008; Sandora et al., 2008). However, the efficacy of hygiene programs in reducing respiratory illness is less certain. Hygiene intervention studies report conflicting results (Roberts, Smith, Jorm, Patel, Douglas and McGilchrist, 2000; Sandora et al., 2008) despite the documented importance of contact transmission for common respiratory pathogens like rhinovirus and respiratory syncytial virus (Gwaltney et al., 1978; Hall et al., 1980). Nevertheless, the success of hygiene programs is perceived to result from a reduction of the role of surfaces (e.g. hands and fomites) as mediators in transmission (Lennell et al., 2008). In support of this perception, multiple studies have demonstrated presence of pathogenic agents on hands and fomites during disease outbreaks, and suggest that the presence of agents is an indicator of infection risk (Green et al., 1998; Cheesbrough et al., 2000; Boone and Gerba, 2005; Wu et al., 2005). In fact, the link between microbial contamination on surfaces and health outcomes has not been fully established. Indoor environmental contamination may be endemic, or it may be an outcome from existing infectious disease while not contributing to further transmission. Among the first studies to establish a relationship between microbial contamination on surfaces and diarrheal illness is the work by Laborde et al. CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 99 (1993), which showed a significant association of hand and surface contamination with reported diarrheal illness in 37 child care centers. Contamination was quantified during a single survey for fecal coliform and was used as a representative sample for the child care center. Conversely, a similar study by Soule et al. (1999) focusing on rotavirus gastroenteritis in pediatric wards found no significant difference in rotavirus on surfaces when comparing rooms occupied by patients with symptomatic rotavirus gastroenteritis to those of uninfected patients. Although the cross–sectional analyses demonstrate potential relationships between environmental contamination and health, they can not elucidate causal links. Previous research has provided limited evidence of causal links between environmental contamination and adverse health outcomes. For example, laboratory studies have demonstrated that infected individuals handling fomites increase rhinovirus presence (Gwaltney, 1982; Winther et al., 2007). Additionally, in child care centers, Butz et al. (1993) demonstrated that presence of rotavirus contamination on surfaces followed two of five observed diarrheal outbreaks. However, none of the aforementioned studies investigated the role of surface contamination in precipitating outbreaks, as all investigated increases in fomite contamination after illness. In two other studies (Van et al. (1991) and Bright et al. (2009)), the authors tracked health and the presence of organisms on fomites and hands in child care centers, but did not incorporate temporal lags in analyses. Therefore, the studies acted more as cross–sectional analyses. Neither study found significant associations between illness and surface contamination (Van et al., 1991; Bright et al., 2009), although Van et al. (1991) supported findings that bacteria on hands is linked to diarrheal illness. In the present study, we investigated the temporal relationship between health and contamination of fomites and hands. The study is the first, to our knowledge, to track both respiratory and gastrointestinal disease incidence while monitoring weekly hand and environmental surface contamination. Microbial contamination in two child care centers was determined using quantitative densities of fecal indicator bacteria (e.g. Escherichia coli, enterococci, and fecal coliform) on hands and fomites as well as presence/absence of viral pathogens (e.g. enterovirus and norovirus). Health was CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 100 monitored daily by childcare staff, who tracked absenteeism, illness-related absenteeism, and symptomatic respiratory and gastrointestinal illness. Based on the data set, we investigate relationships between increases in symptomatic respiratory illness and increases in microbial contamination on hands. By incorporating temporal lags, we assess whether increases in hand or environmental contamination levels result in, or are caused by, increases in respiratory illness when accounting for incubation periods typical for common respiratory illness. 5.3 5.3.1 Methods and Materials Sites Permission from the Stanford University Research Compliance Office for Human Subjects Research was obtained prior to the study. 80 individuals were enrolled in the study at two child care centers in Northern California, USA: 8 child care center staff (100% female) and 72 children (36% female) aged 3-5 years. Hereafter, the sites are referred to as sites A and B. Each of the child care centers has a morning class from 8:00-11:30 with 17-20 enrolled children and a separate afternoon class from 12:30–16:00 with 13-17 enrolled children. The children are assigned one of the class times, and did not change times. Three staff members are on site at each facility. Child care staff or the children’s parents/guardians provided written consent at the start of the study. The child care centers were chosen because of similarities in: 1) geographic location, 2) admission requirements, 3) class schedule, 4) enrollment size, 5) facility layouts, 6) staff size, 7) janitorial service, and 8) food vendor. Additionally, the cleaning and hygiene regimens at the two centers were similar. The same janitorial service cleaned each facility nightly. Children were encouraged to wash their hands with soap and water upon arrival at the facility, following breakfast or lunch (at 8:45 or 13:15) and before snack (at 10:45 or 15:30). At Site B, staff encouraged children to use alcohol based hand sanitizer (ABHS) in addition to soap and water. Specifically, children upon arrival, before breakfast and lunch, before playtime, and when coming in from the outside. CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 5.3.2 101 Surveys / Demographic Data Collection In-person or telephone surveys lasting approximately 10-15 minutes were given to the child care center staff and the parents or guardians of the enrolled children to obtain information on study population demographics. The survey also included questions on hygiene habits (e.g., frequency of handwashing for staff or parent and child), and perceived health of the child such as previous gastrointestinal and respiratory illness, likelihood of future illness, and perception of overall health as rated on a scale of 1-10, with 10 being extremely healthy (referred to as the “healthy child index”). The survey was given at the beginning of the study; a shorter follow-up survey lasting approximately 5-10 minutes including questions on hygiene habits, knowledge, and perceived health of the child was administered at the end of the study as well. 5.3.3 Sampling Scheme Between 5 February 2009, and 1 June 2009, the morning and afternoon classes of both child care centers were visited by the research team weekly. A total of 64 sample events occurred over 16 weeks. Each visit lasted approximately one hour, typically starting between 8:30-10:30 for the morning class and 13:00-15:00 for the afternoon class. The visits were scheduled so as not to interrupt the children’s snacks or learning activities. During each visit, 2-3 research team members collected 8-12 hand rinse samples, collected 5 environmental surface samples, and verified that the health chart (see below) was filled out appropriately. 5.3.4 Health Data Collection Attendance and symptoms of infections among the children and staff were recorded, daily, by the child care center staff. The staff used standardized health charts that included check boxes for symptoms of respiratory and gastrointestinal illness including “stomach pain”, “3 or more bowel movements”, “vomit”, “bloody stool”, “diarrhea”, “runny/stuffy nose”, “fever”, “sore throat”, “cough” and “headache” as well as a comments section to allow for elaboration on reasons for absenteeism or descriptions of “other” symptoms. The staff were not aware of any of the microbiological laboratory CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 102 results until the end of the study. The research team examined completion of the health charts during sampling trips twice weekly, and collected the charts on the first sampling trip of the following week. Individuals with no recorded symptoms on a given day were classified as “nonsymptomatic”, otherwise the individual was classified as “symptomatic”. Symptomatic illness was further classified as “respiratory” (“runny nose”, “headache”, “cough”, or “sore throat”) or “gastrointestinal” (“stomach pain”, “diarrhea”, “bloody stool”, “more than 3 bowel movements”, or “recent vomiting”) illness. Consecutive days of symptomatic illness were classified as new episodes if they were preceded by six symptom-free days, similar to the description of new episodes of illness described elsewhere (Payment et al., 1991; Colford Jr et al., 2002). New episodes are hereafter referred to as “new illness episodes” and are described by the first day of symptomatic illness. The first day of illness was counted as the first observation of symptoms by child care staff or a reason provided by parents for an absence. The duration of an episode was calculated as the number of consecutive days with reported symptoms or illness-related absenteeism. As no data were collected on weekends, an episode was assumed to end on a Friday if no symptoms or illnessrelated absenteeism were reported on the following Monday, whereas if symptoms or illness-related absenteeism were reported on the following Monday, the episode was assumed to include the weekend. 5.3.5 Hand Rinse Sampling Between 8 and 12 hand rinse samples were collected during each visit, for a total of 616 samples over the duration of the study. Eight to ten children were sampled during each visit, and the three child care staff at each facility were sampled once per week. The children were asked to participate in an order determined through randomization of the list of enrolled children. If a child declined, the next child on the list was approached. Once assent was obtained, the researcher recorded whether or not there were visible signs of a runny nose, dirt on hands, and dirt under fingernails. The researcher then asked the subject how s/he was feeling, and the response was CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 103 recorded and later reclassified as “sick”,“fine”, or “no response”. Hand rinse sampling was performed using a modified glove-juice method as previously described (Pickering et al., 2010). The subject was asked to place first one hand, and then the other, into the same 69 oz. Whirl-pak bag (Nasco, Fort Atkinson, WI, USA) filled with 350 ml autoclaved Milli-Q grade water. The subject was encouraged to shake the hand vigorously for 15 s; the researcher then massaged the hand through the bag for an additional 15 s. After both hands were rinsed, the subject was provided a clean paper towel to dry her/his hands. The sample was placed on ice and transported to the laboratory, where it was processed within 6 hours. 5.3.6 Environmental Surface Sampling Surface samples were obtained immediately after the hand rinse samples. Between three and five fomites were sampled during each visit, for a total of 299, chosen based on the subjective classification as a surface with a high likelihood of contact based on children’s activities in the previous hour. For example, a toy block would be sampled if one or more children had been observed playing with the block. A summary of the fomites tested is reported in Table 5.1, with the most commonly sampled surfaces including toys, table tops, faucets, and doorknobs. The surface was sampled with a sterile cotton-tipped applicator wetted in 10 ml of 1/4 strength Ringer’s solution. The area sampled varied between approximately 25-100 cm2 , depending on the object tested. After sampling, the swab was replaced in the 1/4-strength Ringer’s solution and transported to the lab for bacterial assay (Kaltenthaler et al., 1995; Kyriacou et al., 2009). 5.3.7 Microbiological Methods Bacterial Assays All hand and surface samples were assayed for three fecal indicator bacteria: fecal coliform, enterococci and Escherchia coli. Membrane filtration was used to enumerate the bacteria. Fecal coliform was grown on mFC agar (BD Diagnostics, Inc, USA) at 44.5◦ C according to the U.S. EPA standard methods for water quality testing CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 104 (Bordner et al., 1978). Enterococci was grown on mEI agar (BD Diagnostics, Inc, USA) at 41.5◦ according to U.S. EPA Method 1600 (USEPA, 2002a). Escherchia coli was grown on modified mTEC agar (BD Diagnostics, Inc, USA) according to U.S. EPA Method 1603 (USEPA, 2002b) with incubation at 35◦ for two hours followed by incubation at 44.5◦ for an additional 22 hours. For the hand rinse samples, a volume of between 65-80 ml of the total 350 ml collected was filtered. For the surface samples, a volume of between 2-2.5 ml of the total 10 ml collected was filtered. To calculate bacterial concentration per two hands, the colony counts were multiplied by the ratio of total sample volume collected (350 ml) to sample volume filtered for each sample. For data analysis of hand rinse samples, the lower detection limit is 5.4 CFU per two hands, based on a 65 ml filtered sample volume. When no detectable bacteria was present, 1/2 the lower limit of detection (2.7 CFU per two hands) was used. For data analysis of surface samples, the lower detection limit is 5 CFU per surface, based on a 2 ml filtered sample volume. Surfaces samples were classified as either contaminated (≥5 CFU per 25-100 cm2 surface) or uncontaminated when no bacteria were detectable. Viral Assays A subset of sixty-seven hand rinse samples were also tested for the presence of enterovirus, norovirus genogroup I, and norovirus genogroup II. The sample volume remaining after bacterial assays, between 80-110 ml, was filtered through a 0.45um negatively-charged nitrocellulose filter (HA filter, Millipore, Billerica, MA, USA), placed in a 2 oz Whirl-pak bag (NASCO Corp., Fort Atkinson, WI) and stored at -80◦ . Both RNA and DNA were extracted using the Qiagen AllPrep DNA/RNA Mini Kit (Qiagen, Valencia, CA, USA) and eluted in 60 µl of RNAse/DNAse free water. RNAse/DNAse free water was used as an extraction blank with every set of ten samples extracted. 5 µl of template was then used in a 25 µl RT-PCR reaction, the details of which are included in Table 5.2. RT-PCR reagents used were the Qiagen One-Step RT-PCR kit (Qiagen, Valencia, CA, USA). Thermocycler conditions were obtained from RT-PCR kit manufacturer’s recommendations using the annealing temperatures listed in Table 5.2, and optimized on an Applied Biosystems Thermal Cycler 9700 CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 105 (Applied Biosystems, Foster City, CA). All PCR products were visualized on 1.5% agarose gels using a BioRad Gel Doc XR system (BioRad, Hercules, CA). 5.3.8 Statistics Most statistics were performed using PASW Statistics 18.0.2. (SPSS: An IBM Company, Chicago, IL, USA). Statistical methods are reported with the results, with additional details available in Appendix C. Analyses using generalized estimating equations (GEEs) were performed using the “geeglm” function in the “geepack” package in R (version 2.11.1, R Foundation for Statistical Computing, Vienna, Austria) (Zuur et al., 2009). Magnitude of the coefficients (β) and significance level (p) are reported, where the significance level used throughout the study is α =0.05. 5.4 Results 5.4.1 Surveys The population characteristics are presented in Table 5.3, including age of the parent/guardian respondent, ethnicity of the child, and percent of children with a chronic disease and taking medication. The only category in which the parents of children attending Site A and Site B differed significantly (p = 0.02) was the number of residents under 6 years old in their households. The parents of children at Site A had a mean 0.5 more residents under six years old than parents at Site B. Characteristics of hand hygiene habits, general health of the child, and previous and predicted respiratory and gastrointestinal illness are reported in Table 5.4. 5.4.2 Health Data A total of 5,651 person-days of health data were collected over the duration of the study. Attendance and symptomatic illness status were recorded for 5,619 (99.4%) and 5,503 (97.4%) days, respectively. The days without recorded attendance or symptomatic illness data were excluded from analysis. CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 106 Summaries of attendance and symptomatic illness for staff and children at each site are presented in Table 5.5. In total, children and staff were absent on 636 (11.3%) person-days. Symptomatic illness in children or staff was observed and reported by child care center staff on 1063 person-days, or 19.3%) of the total 5,303 person-days with recorded symptomatic illness status. On an additional 94 (1.7%) person-days, children were absent due to illness with symptoms that were not specified by the parents or guardians; these person-days were included in analyses of symptomatic illness but treated as missing data in analyses of respiratory or gastrointestinal illness. A time series of the absences and illness-related absences of children and staff at both of the centers is shown in Figure 5.1. Similarly, times series of respiratory illness, gastrointestinal illness, and new illness episodes are shown in Figure 5.2. Of the 5,503 person-days with recorded symptomatic illness status, respiratory symptoms were reported on 18.3% (or 1,010 person-days). The most common respiratory symptoms were runny nose (15.8% or 872 of the 5,503 person-days), cough (4.8% or 265 person-days), and sore throat (0.9% or 52 person-days). Gastrointestinal symptoms were reported on 38 (0.7%) person-days. The most common gastrointestinal symptoms were vomiting (0.3% or 18 person-days), stomach pain (0.3% or 18 person-days), and diarrhea (0.1% or 5 person-days). Fever was reported on 83 persondays (1.5%), 46 (0.8%) person-days were in conjunction with respiratory symptoms, 12 (0.2%) person-days with gastrointestinal symptoms, and 25 (0.5%) person-days as the only symptom. A total of 232 new illness episodes were identified during the study. Of those, 118 (50.8%) new illness episodes were first identified and reported by child care center staff as symptomatic illness on a day when the child or staff member was present. The remaining 114 (49.1%) new illness episodes were first reported as an absenteeism due to illness. The majority of the new illness episodes were respiratory (161 episodes, or 69.4%). The rest were gastrointestinal (15 episodes, or 6.4%), fever alone (14 episodes, or 6.0%), or an illness-related absences with unspecified symptoms (43 episodes, or 18.5%). The duration of episodes ranged between 1 and 48 days, with a mean and median of 6.7 and 3 days, respectively. Rates of health outcomes varied by site, as compared using Pearson’s χ2 . Site A CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 107 had significantly fewer absences (p = 0.02), illness-related absences (p = 0.001), and respiratory illness (p < 0.001) than Site B. There was no significant difference by site in gastrointestinal illness (p = 0.20) or new illness episodes (p = 0.82). Because of the significant differences in absences, illness-related absences, and respiratory-illness, all analyses of health outcomes using GEEs included a site-dependent variable. 5.4.3 Hand Rinse Samples Of the 616 hand rinse samples collected, enterococci, fecal coliform, and E. coli were detected in 208 (33%), 83 (14%), and 31 (5%), respectively. The range in concentrations of bacteria on hands was the same for all three bacteria: 8–≥1000 CFU per 2 hands. A time series of the fraction of hand rinse samples with detectable enterococci and fecal coliform is presented in Figure 5.3. Norovirus gI and gII were not detected in any of the 67 hand rinse samples tested. Enterovirus was detected in four of the 67 hand rinse samples tested (6%). Mean concentrations of all three fecal indicator bacteria were higher on the hands with visible dirt, visible dirt under nails, and on volunteers with visible runny noses. The concentrations of E. coli (p = 0.022), enterococci (p = 0.003), and fecal coliform (p = 0.006) were significantly higher when dirt on hands was visible, with mean effect sizes of 0.052, 0.113, and 0.141 log10 CFU per two hands, respectively. Only the concentration of fecal coliform (p = 0.035), but not E. coli (p = 0.206) or enterococci (p = 0.239), was significantly higher when dirt under nails was visible with a mean effect size of 0.077 log10 CFU per two hands. A visible runny nose was associated with significantly higher concentrations of enterococci (p = 0.001) and fecal coliform (p = 0.021), with mean effect sizes of 0.240 and 0.095 log10 CFU per two hands, respectively, but not with concentrations of E. coli (p = 0.089). A volunteer’s response on how s/he was feeling at the time of the sampling was not associated with the concentration of E. coli (p = 0.459), enterococci (p = 0.100), or fecal coliform (p = 0.511) on the hands. The low prevalence of enterovirus on hands precluded statistical analysis with presentation of symptoms or presence of fecal indicator bacteria. Two of the four children with detectable enterovirus also had reported symptomatic respiratory CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 108 illness. 5.4.4 Environmental Samples In total, 299 environmental samples were collected. A summary of the objects tested and the number of samples with detectable enterococci and fecal coliforms are presented in Table 5.1. Enterococci and fecal coliform were detected in 19 (6%) and 9 (3%) of the samples, respectively. E. coli were not detected in any of the samples. Concentrations on fomites for both enterococci and fecal coliform ranged from ≤5–≥1000 CFU per 100 cm2 . The presence of enterococci on a surface was significantly correlated to the presence of fecal coliform (McNemar χ2 test, p = 0.041). The presence of enterococci or fecal coliform was not significantly associated with site (Pearson χ2 test, enterococci p = 0.59, fecal coliform p = 0.82) or time of class (Pearson χ2 test, enterococci p = 0.89, fecal coliform p = 0.79). Data on the fraction of fomites sampled in a given week with detectable enterococci or fecal coliform were used in all time series analyses, as presented in Figure 5.4. 5.4.5 Health Associations with Hand and Surface Contamination Fifteen separate GEE were used to examine the associations between respiratory illness and microbial contamination on hands and surfaces for daily lags of up to -7 through +7 days (Table 5.6). Data were clustered by child, and a variable was included to control for site. Symptomatic respiratory illness is significantly and positively associated with hand contamination on the same day (β = 0.40, p = 0.003), one day before (β = 0.31, p = 0.041)) and one day later (β = 0.41, p = 0.005), four days before (β = 0.99, p =< 0.001), and seven days later (β = 0.41, p = 0.015). In addition, there is a significant and positive association with environmental contamination as measured four (β = 1.37, p = 0.42) and five (β = 0.79, p = 0.014) days before as well as two (β = 0.69, p = 0.023) and three (β = 0.091, p = 0.030) days later. CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 109 No significant associations between hand and environmental contamination with either illness-related absences or new episodes of illness were found (See Tables C.1 and C.2). Gastrointestinal illness prevalence was low, and therefore insufficient to analyze associations between gastroenteritis and microbial contamination in a manner similar to respiratory illness. 5.5 Discussion In child care centers, hygiene plays an important role in reducing transmission of both gastrointestinal and respiratory illness. Previous field studies in child care centers have demonstrated significant correlations between microbial contamination and adverse health outcomes (Van et al., 1991; Laborde et al., 1993), but this study is the first to our knowledge to infer causality based on analysis of temporal trends. As such, the study provides insight into the timing of microbial contamination relative to symptomatic illness. Specifically, increases in detectable enterococci on hands and fomites precedes symptomatic respiratory illness by a four- to six- day period consistent with incubation periods for respiratory diseases (Long et al., 1997). Furthermore, the study demonstrates that the occurrence of enterococci on hands and fomites increases in the two days following symptoms. These findings suggest that respiratory illness can contribute to, and result from, microbial contamination on hands and fomites. The illness rates and microbial contamination in this study used to infer causal relationships are consistent with previously observed values. The estimated rates for total illness (0.76 per child per month) and respiratory illness (0.63 per child per month) are similar to the rates reported for children of similar ages attending child care (Wald et al., 1991; Dahl et al., 1991; Krilov et al., 1996). Microbial contamination is also similar to previous studies. In classrooms with children from infancy to under five years old, fecal coliform and fecal streptococci (a group of organisms of which enterococci are a subset) were observed on 4-10% and 16% of fomites sampled (Weniger et al., 1983; Holaday et al., 1990; Kyriacou et al., 2009), consistent and CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 110 expectedly greater than the 3% and 6% detection rates in the present study. Similarly, our detection rates for fecal coliform (14%) and enterococci (33%) on hands are consistent with reported detection rate of 6-20% for fecal coliform (Van et al., 1991; Holaday et al., 1990), and 52.9% for fecal streptococci (Kyriacou et al., 2009). Respiratory symptoms increase with enterococci occurrence on hands on the same day, one day before, and one day after. Although other studies have demonstrated significant correlations between bacteria on hands and health (Van et al., 1991; Pickering et al., 2010), this is the first study to demonstrate significant associations with daily lags. This finding suggests enterococci acts as a superior indicator for respiratory illness relative to both fecal coliform or E. coli. Enterococci, while commonly isolated in feces, have also been isolated from the mouth (Murray, 1990) and the nose (Crossley and Ross, 1985). In the present study, runny/stuffy nose or coughing were reported as the majority of symptoms, providing a possible source of enterococci to the environment. In support, individuals with visible runny noses had significantly higher concentrations of enterococci on hands. Furthermore, (Pickering et al., 2011) demonstrated, in Africa, significant associations between enterococci density on mothers’ hands and time since last handwashing. The same relationship was not significant for E. coli (Pickering et al., 2011). Respiratory symptoms significantly lag enterococci occurrence on fomites two to three days later. This finding is consistent with asymptomatic excretion of microorganisms persisting after the conclusion of symptoms, as has been reported for respiratory viruses (Long et al., 1997). However, the median duration of symptomatic illness is three days, so significant associations within three days of symptoms may not necessarily imply associations occurred in the absence of symptoms. Similarly, the majority of fomites with detectable enterococci were toys that were not cleaned regularly by the nightly janitorial staff. Significant associations, therefore, may be a result of enterococci persistence. Pinfold (1990) suggests that fecal streptococci, a group of microorganisms that include enterococci, are more likely to survive crosscontamination than E. coli because enterococci survival on fingertips is 4-8 times longer (Pinfold, 1990). Nevertheless, the study suggests that hygiene interventions CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 111 targeting recovered individuals for up to three days following the conclusion of observable symptoms may reduce illness transmission. Enterococci occurrence on fomites preceded increases in respiratory illness by four to five days. As the incubation period for common respiratory viruses is 2-5 days (Long et al., 1997), this finding is among the first field evidence of a causal role of microbial contamination on fomites in respiratory disease transmission. As enterococci may be shed with nasal secretions, the findings suggest that the increased presence on fomites may be indicative of increased presence of pathogens responsible for outbreaks, as well. The detection of enterovirus on hand rinse samples demonstrates the potential role of hands in pathogen transmission. The pan-enterovirus primers used in this study to detect enterovirus are capable of detecting some serotypes of rhinovirus, a common respiratory virus in child care centers (Rotbart, 1995). The low detection rate (6% of hand rinse samples tested) is consistent with a previous study conducted in Africa (Pickering et al., 2011). A higher detection rate may have been possible with a more efficient virus recovery method. We compared direct extraction of sewage to the hand rinse sample method by spiking hand rinse water with sewage and found an approximate ten–fold increase in the lower limit of detection (data not shown). Accounting for the ten–fold increase, PCR template volume, hand rinse sample volume, and a PCR reaction lower limit of detection equal to the published 2.5 PFU of poliovirus (Jaykus et al., 1996), the detection limit for virus is at least 1500 PFU poliovirus per two hands. Efforts to reduce the detection limit may yield higher pathogen detection rates. The only health outcomes demonstrating significant associations with microbial contamination were respiratory symptoms. The other health outcomes modeled, illness–related absences and new illness episodes, did not demonstrate significant trends when modeled as a function of enterococci contamination on surfaces (See Appendix C). Modeling absences, including illness-related absences, as a function of microbial contamination was confounded by the inability to collect data for children who are not present during sampling trips. The lack of data on microbial contamination on hands of absent children likely contributed to bias in the illness-related CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 112 absences model. New illness episodes are similarly impacted as absences accounted for almost half (49%) of all defined illnesses. In fact, symptomatic respiratory illness is well–suited to be modeled as a function of microbial contamination as shedding, particularly during symptomatic illness, is perceived to be a cause of transmission and contamination (Hall et al., 1980). For a further discussion of the health outcome model used, specifically the use of multiple comparisons, refer to the Appendix C. The present study suggests that respiratory illness both lags and leads increased environmental surface and hand contamination. However, data collection methods and statistical analysis used in the study may bias the findings. Symptomatic illness in this study, as it was reported by child care center staff, is a subjective measure. Evidence of the influence of subjectivity of the child care center staff includes the significantly different respiratory incidence rates between Site A and Site B, despite similarity between the two centers (See Table 5.3) and a similar number of new episodes of illness (See Table 5.5). At the end of the study the child care center staff were prompted on the recording frequency, and the resulting quality of the child health charts. Between the choices of “highly”, “somewhat”, and “not accurate”, the child care center staff reported that the quality of the data was “somewhat accurate” and that the charts were filled out daily. Identification of health outcomes, specifically the presentation of symptoms, by trained professionals combined with collection and analysis of clinical samples for specific etiological agents would likely improve reliability of health measurements. Similarly, health data were only collected on weekdays. Therefore, samples collected on days where the associated lag time corresponded to a weekend were not included in the analysis. However, the frequency of illness relative to the number of observations for every lag remained consistent (See Table 5.6) suggesting bias may be minimized. Finally, as the study covered only four months, seasonal trends in illness may be confounded with seasonal trends in microbial contamination. Future studies investigating the relationship between fomites and respiratory illness could incorporate indicator bacteria with specificity to nasal secretions or saliva by identifying organisms from, for example, recent microbiota studies (Frank et al., 2010). The bacteria used as indicators of microbial contamination in this study are CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 113 typically used as indicators of fecal contamination, and are therefore more typically associated with gastrointestinal illness than with respiratory illness. However, the significant findings of our results suggest that presence of enterococci may be an appropriate predictor of respiratory symptoms. 5.6 Acknowledgments The authors acknowledge Lauren Sassoubre, Isaias Espinoza, and Elfego Felix for their assistance on site and in the laboratory, as well as Todd Russell, Thienan Nguyen, and Francisco Tamayo for their assistance in the laboratory. The Boehm Research Group provided helpful suggestions for study design and data analysis. The authors also acknowledge Gojo Industries, Inc, for providing Purell Alcohol Based Hand Sanitizer to the participating child care centers. Gojo Industries was not otherwise involved in the study. The research has been funded, in part, by the UPS Foundation Endowment Fund at Stanford University and the United States Environmental Protection Agency (EPA) under the Science to Achieve Results (STAR) Graduate Fellowship Program. EPA has not officially endorsed this publication and the views expressed herein may not reflect the views of the EPA. CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 5.7 Tables 114 CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS Surface Tested ball 7 block 11 book 4 chair 10 computer 5 doorknob 17 faucet 21 floor 3 glue container 1 kitchen surfaces 3 marker 5 mirror 1 playground 29 sandbox 10 shelf 3 soap dispenser 1 storage bin 4 table 59 toilet 6 toothbrush 1 toy 92 tray 1 water table 5 Total 299 Enterococci 1 (5%) 1 (33%) 1 (20%) 4 (14%) 2 (20%) 1 (2%) 9 (10%) 19 (6%) 115 Fecal Coliform 2 (10%) 1 (10%) 1 (2%) 5 (5%) 9 (3%) Table 5.1: Summary of environmental fomites sampled along with the number and corresponding percent of samples with detectable (≥5 CFU per 100cm2 ) enterococci and fecal coliform. No E. coli were detected on fomites 371 344 Entero- Noro- gI Noro- gII TA Primers Reference ◦ ( C) 55 F: 5’- ACCGGATGGCCAATCCAA -3’ Jaykus et al., 1996 R: 5’- CCTCCGGCCCCTGAATG -3’ 50 F: 5’- CTGCCCGAATTYGTAAATGA -3’ Kojima et al., 2002 R: 5’- CCAACCCARCCATTRTACA -3’ Lyman et al., 2009 50 F: 5’- CNTGGGAGGGCGATCGCAA -3’ Kojima et al., 2002 R: 5’- CCRCCNGCATRHCCRTTRTACAT -3’ Lyman et al., 2009 Table 5.2: Amplicon size, annealing temperature, and primers for PCR detection of enterovirus (Entero-) and norovirus (Noro-) gI and gII Size (bp) 192 Target CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 116 CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS Characteristics Number of parents or guardians interviewed Site A Site B p − value 28 34 30 22-44 29 22-48 0.91 86 7 7 97 0 3 0.28 43 21 36 44 15 41 0.77 5.2 2.2 4.9 1.6 0.45 0.02 21 38 0.24 18 7 25 32 15 9 24 15 0.97 Response rate (interviewed / enrolled) Age of parent or guardian respondent (years) Mean Range Self-reported ethnicity (%) Hispanic African American Other Type of family residence %) Single Family Duplex Apt. Complex Mean no. residents in child’s household Total Under 6 yo Children from families with pets (%) Children with chronic disease (%) Asthma Other Total Children on any medications (%) 0.84 0.2 Table 5.3: Child Care Center Population Demographics 117 CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS Characteristic Baseline No. gastrointestinal illness in past 6 mo. (%) 0 1 2 or more No. respiratory illness in past 6 mo. (%) 0 1 2 or more Site A Site B 89 4 7 56 26 9 32 32 36 3 21 74 8.24 8.04 7.80 6.80 7.40 6.60 4 71 21 18 6 56 24 3 7.86 7.03 9.40 11.20 11.10 7.30 4 48 32 16 0 26 59 15 16 44 0 44 Healthy child index (mean) Handwashing rate per day (mean) Parent Child Likelihood of child illness in next month (%) High Some None Don’t Know Follow-up Healthy child index (mean) Handwashing rate per day (mean) Parent Child Likelihood of child illness in next month (%) High Some None Don’t Know Child was ill during study (%) Gastrointestinal Respiratory Table 5.4: Child Care Center Population Health and Hygiene Knowledge 118 CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS Site A Category Total No. Person-Days Absences Total Illness-Related Illness Total Illness∗ Symptomatic Illness Respiratory Total Gastrointestinal Total New Episodes Total Respiratory Gastrointestinal Fever Only Unspecified 119 Site B Children Staff Children Staff Total n = 37 2560 n=3 229 n = 33 2618 n = 3-5 212 5619 281 135 7 2 339 196 9 2 636 335 368 306 57 57 714 669 31 31 1170 1063 288 57 635 30 1010 14 1 22 1 38 106 67 5 5 31 7 7 0 0 0 113 82 9 9 13 6 5 0 0 0 232 161 14 14 44 Table 5.5: Number of person-days with recorded attendance and symptomatic illness subset by child care facility and site. ∗ Total Illness is the combination of recorded symptomatic illness and illness-related absenteeism with unspecified symptoms. Obs. 407 373 294 205 188 274 443 534 446 338 223 184 223 359 420 Ill. 89 80 61 41 34 53 89 101 72 53 33 38 59 77 82 Coef. -1.95 -1.88 -2.24 -1.91 -2.65 -2.60 -2.52 -2.10 -2.11 -2.27 -2.29 -2.63 -2.22 -2.01 -1.99 Intercept SE Pr(>|W|) 0.29 <0.001 0.32 <0.001 0.39 <0.001 0.30 <0.001 0.44 <0.001 0.39 <0.001 0.33 <0.001 0.30 <0.001 0.33 <0.001 0.34 <0.001 0.37 <0.001 0.38 <0.001 0.42 <0.001 0.33 <0.001 0.30 <0.001 Coef. 0.41 0.34 0.25 0.14 0.34 0.34 0.41 0.40 0.31 0.15 0.38 0.99 0.32 0.19 0.00 Hands SE Pr(>|W|) 0.17 0.015 0.18 0.053 0.20 0.213 0.21 0.492 0.36 0.343 0.22 0.112 0.15 0.005 0.14 0.003 0.15 0.041 0.18 0.393 0.24 0.107 0.22 <0.001 0.20 0.116 0.16 0.250 0.16 0.987 Coef. 0.33 -0.12 0.05 0.37 0.91 0.69 0.34 0.27 0.28 0.23 0.15 1.37 0.79 0.25 0.01 Fomites SE Pr(>|W|) 0.21 0.106 0.22 0.580 0.25 0.843 0.31 0.227 0.42 0.030 0.30 0.023 0.19 0.069 0.17 0.122 0.17 0.104 0.22 0.300 0.33 0.649 0.42 0.001 0.32 0.014 0.22 0.257 0.23 0.956 Coef. 0.42 0.75 1.28 0.47 1.25 1.22 1.16 0.53 0.21 0.73 0.39 0.55 1.31 0.81 1.00 Site Correlation SE Pr(>|W|) Estimate SE 0.33 0.199 0.15 0.05 0.34 0.027 0.16 0.06 0.41 0.002 0.32 0.16 0.40 0.237 0.07 0.07 0.42 0.003 -0.01 0.07 0.40 0.002 0.04 0.05 0.38 0.002 0.18 0.07 0.33 0.102 0.16 0.05 0.38 0.587 0.19 0.07 0.39 0.058 0.11 0.05 0.47 0.414 0.08 0.04 0.46 0.235 0.10 0.10 0.43 0.002 0.12 0.07 0.37 0.029 0.16 0.06 0.34 0.003 0.11 0.04 Table 5.6: Parameters of generalized estimating equation for respiratory illness as function of enterococci on hands and fomites. “Lag” is the number of days between collection of health data and the collection of data for contamination on surfaces where a positive lag implies that health outcomes preceded microbial contamination. “Obs.” is the number of observations with both health measurements and microbial contamination data. “Ill.” is the number of observations when respiratory illness was observed,“Hands” is the number of enterococci detected on hands, “Fomites” is the fraction of fomites sampled with detectable enterococci, “Site” is the facility, with the coefficients representing the difference ifor Site B relative to Site A, “Coef.” and “SE” are the coefficient and standard error of the correlation coefficient. “Pr(>|W|)” is the significance, with values less than 0.05 considered significant and highlighted in bold. Lag 7 6 5 4 3 2 1 0 -1 -2 -3 -4 -5 -6 -7 CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 120 CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 5.8 Figures 121 CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS Proportion Absent Per Day 1.0 (a) Absences 0.8 Children - Site A (n = 37) Children - Site B (n = 33) Staff - Site A (n = 3) Staff - Site B (n = 3-5) 0.6 0.4 0.2 n Ju 1- ay ar pr -A M 8- 14 -M 21 b -Fe 1.0 25 b Fe 2- 0.0 Proportion Absent due to Illness Per Day 122 (b) Illness-Related Absences 0.8 Children - Site A (n = 37) Children - Site B (n = 33) Staff - Site A (n = 3) Staff - Site B (n = 3-5) 0.6 0.4 0.2 n ay Ju 1- M 8- ar pr -A 14 -M 21 b -Fe 25 b Fe 2- 0.0 Figure 5.1: Time series of the proportion of children and staff who (a) are absent, and (b) are absent due to illness. The shaded portion of the figures represents five days (April 13–April 17) when no child care classes were held and no data were collected Proportion with Respiratory Symptoms Per Day CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 1.0 (a) Respiratory Symptoms 0.8 Children - Site A (n = 37) Children - Site B (n = 33) Staff - Site A (n = 3) Staff - Site B (n = 3-5) 0.6 0.4 0.2 (b) Gastrointestinal Symptoms 0.8 Children - Site A (n = 37) Children - Site B (n = 33) Staff - Site A (n = 3) Staff - Site B (n = 3-5) 0.6 0.4 0.2 ay n Ju 1- M 8- ar pr -A 14 -M b b Fe -Fe 21 25 0.0 2- Proportion with Gastrointestinal Symptoms Per Day n ay Ju 1- M 8- ar b pr -A 14 -M 21 -Fe 25 b Fe 2- 0.0 12 No. New Illness Episodes Per Day 123 (c) New Illness Episodes 10 Children - Site A (n = 37) Children - Site B (n = 33) Staff - Site A (n = 3) Staff - Site B (n = 3-5) 8 6 4 2 n ay Ju 1- M 8- ar pr -A 14 -M 21 b -Fe 25 b Fe 2- 0 Figure 5.2: Time series of the proportion of children and staff who (a) have respiratory symptoms, and (b) have gastrointestinal symptoms. Also presented is a time series of the first day of (c) new illness episodes. The shaded portion of the figures represents five days (April 13–April 17) when no child care classes were held and no data were collected Proportion of Hands with Detectable Bacteria CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 1.0 (a) Hand Contamination at Site A 0.8 enterococci - AM enterococci - PM fecal coliform - AM fecal coliform - PM enterovirus neg / pos 0.6 0.4 0.2 Under Limit of Detection 0.0 n ay Ju 1- M 8- ar b pr -A -M 14 21 -Fe 1.0 25 b Fe 2- Proportion of Hands with Detectable Bacteria 124 (b) Hand Contamination at Site B 0.8 enterococci - AM enterococci - PM fecal coliform - AM fecal coliform - PM enterovirus neg / pos 0.6 0.4 0.2 Under Limit of Detection 0.0 n ay Ju 1- M 8- ar b pr -A -M 14 21 -Fe 25 b Fe 2- Figure 5.3: Time series of the proportion of hand samples with detectable bacteria at (a) Site A and (b) Site B. The shaded portion of the graph represents five days (April 13–April 17) when no child care classes were held and no data were collected. Sampling visits when no samples had bacterial densities above the lower limit of detection on hands (≥ 5 CFU per two hands) are marked by columns with heights equal to the line marked “Limit of Detection”. Samples tested for enterovirus are under the abscissa corresponding to the sample’s date. Each ◦ represents a negative sample and • represents a positive sample Proportion of Surfaces with Detectable Bacteria CHAPTER 5. HEALTH AND SURFACES IN CHILD CARE CENTERS 1.0 (a) Environmental Contamination at Site A 0.8 enterococci - AM enterococci - PM fecal coliform - AM fecal coliform - PM 0.6 0.4 0.2 Under Limit of Detection ay n Ju 1- M 8- ar pr -A 14 M 21 b b Fe 1.0 -Fe 25 0.0 2- (b) Environmental Contamination at Site B 0.8 enterococci - AM enterococci - PM fecal coliform - AM fecal coliform - PM 0.6 0.4 0.2 Under Limit of Detection n ay Ju 1- M 8- ar pr -A 14 -M 21 b b Fe 2- 0.0 -Fe 25 Proportion of Surfaces with Detectable Bacteria 125 Figure 5.4: Time series of the proportion of fomites sampled with detectable bacteria on hands at (a) Site A and (b) Site B, and on fomites at (a) Site A and (b) Site B. The shaded portion of the graph represents five days (April 13–April 17) when no child care classes were held and no data were collected. Sampling visits when no samples had bacterial densities above the lower limit of detection on fomites (≥ 2.5 CFU per 25 cm2 ) are marked by columns with heights equal to the line marked “Limit of Detection”. Chapter 6 Conclusions and Future Directions 6.1 Conclusions Conclusion 1: Virus transfers readily between surfaces Chapter 2 investigates the fraction of virus that transfers between a fingerpad and a glass surface. From the study, we demonstrate that the mean, median, and standard deviation of the fraction of virus transferred between a fingerpad and glass surface is 0.23, 0.18, and 0.22, respectively. These findings are of similar order of magnitude as findings in previous literature on virus transfer between nonporous surfaces and fingerpads (Ansari et al., 1991; Mbithi et al., 1992; Rusin et al., 2002). The findings demonstrate that the amount of virus transferred, on a single contact, to the fingerpad from a contaminated fomite is at a similar order of magnitude as the original level of contamination. Chapter 2 also demonstrated that specific factors investigated (e.g., hand washing, direction of transfer, and virus species) significantly influenced the fraction transferred. However, the small effect size (5-10% of total fraction transferred) of the factors on viral transfer suggests the factors likely have little impact on infection risk from fomites. Conclusion 2: Density of microorganisms on surfaces is both increased by, and leads to, adverse health outcomes 126 CHAPTER 6. CONCLUSIONS 127 A major finding of this dissertation is evidence of a causal link between density of microorganisms on surfaces and risk of infection. Chapter 3 demonstrates that in a model of child-fomite interaction, the concentration of virus on the child’s previously uncontaminated hands will equal the concentration on the fomite within minutes due to frequent, repetitive hand-surface contacts (Figure 3.3). At that time, hand-mouth contacts will contribute to ingested dose, and therefore risk of illness. This will occur even if the fomite is removed. In the model, the initial concentration on the hands is directly linked to the likelihood of infection. As Table 3.2 demonstrates, a tenfold increase in the concentration of virus on the model fomite increased the child’s dose one hundred fold, even after only 10 minutes of child-fomite interaction. Therefore, the model suggests that high concentrations of virus on surfaces are indicative of increased risk of illness. If microbial contamination on surfaces are indicative of increased risk of illness, it is likely that a field scale study investigating temporal trends in contamination and health outcomes would demonstrate associations. Multiple previous studies have suggested that a significant correlation exists between microbial contamination and health (Van et al., 1991; Butz et al., 1993; Laborde et al., 1994). However, these studies have shown correlations based on sampling fomites only once, or based on multiple samplings that all occur after outbreaks. Without temporal sampling during both outbreak and non-outbreak periods, causation between microbial contamination and illness can not be inferred. Rather, increases in microbial contamination need to precede increases in disease burden to demonstrate causation, as suggested in Chapter 3. As there is a lag between dose and response for gastrointestinal and respiratory viral disease of between 12 and 120 hours (See Table 1.1), increases in illness due to microbial contamination need to be tracked with daily resolution. This was the motivation for Chapter 5, the field-based study investigating temporal trends in contamination and health. Chapter 5, demonstrates that increases in microbial contamination lead to increases in adverse health outcomes. Using enterococci as an indicator of bacterial contamination, respiratory disease is significantly and positively associated with contamination four and five days prior to illness. The lag of four or five days is consistent CHAPTER 6. CONCLUSIONS 128 with incubation periods (typically between 1-4 days) for respiratory illness. Not only does the work in Chapter 5 suggest that microbial contamination contributes to respiratory disease burden, but it also demonstrates that microbial contamination is caused by respiratory disease burden. Specifically, enterococci on the hands is significantly associated with symptomatic illness on the same day as symptoms, as well as on both the day before and the day after. This finding holds not only for the symptoms reported by the child care center staff, but also for the symptom of visible runny nose as reported by the research team during hand sample collection. Conclusion 3: Virus sampling methods on fomites should be standardized Evidence that microbial contamination contributes to respiratory illness supports the need for standardized fomite sampling (Chapter 4). In the study in child care centers (Chapter 5), enterococci was used as an indicator of microbial contamination. The next research step is to sample hands and fomites for etiological agents in addition to indicators. Significance of an association between etiological agents on surfaces and increased illness (and accounting for the three to five day incubation period) would provide stronger evidence of the role of fomites in respiratory disease transmission in child care centers. Indicators such as enterococci are more easier to detect in the environment than pathogens. Therefore, etiological agents may be detectable on fewer than the 6% of samples with detectable enterococci identified in Chapter 5. To improve detection of virus on surfaces, Chapter 4 suggests an effective sampling method. The use of polyester-tipped swabs in conjunction with 1/4 strength Ringer’s (hereafter referred to as “Ringer’s”) or saline solution resulted in significantly increased detection of infective virus relative to other methods tested. Using MS2 bacteriophage as a model virus, polyester-tipped swab with Ringer’s recovered 60% more than the mean total virus recovered using other methods tested in the study. The literature review included in the chapter also demonstrated that polyester-tipped swabs were significantly associated with higher fraction of samples with detectable virus. Use of a standardized method will also allow cross comparisons of fomite-sampling studies. In the 45 studies identified in Chapter 4 that sampled surfaces for pathogenic CHAPTER 6. CONCLUSIONS 129 virus, the authors used 12 different implements and 4 different eluents. As implement choice significantly influences recovery of virus from surfaces, comparison of outcomes across studies is difficult unless the same sampling method is used. Standardizing the sampling method, specifically through use of the polyester-tipped swabs in Ringer’s or saline solution, would reduce this bias. Alternatively, quantifying the lower limit of detection for the assay, or quantifying the virus detected on surfaces, would allow cross comparison of studies. 6.2 Future Directions My dissertation explored the role of fomites in disease transmission. I anticipate my future research will build upon this knowledge, expand to incorporate additional transmission routes, and continue investigating additional environmental reservoirs of infectious disease. The goal of my future research will be to contribute to a holistic understanding of human-environment interactions in infectious disease transmission. In this section I describe several areas of future research following on my dissertation work that will lead to new discoveries concerning the role of fomites in communicable infectious disease. Linking physicochemical properties of etiological agents to survival and transmission The movement and fate of virus through the environment via fomites may be influenced by virus physicochemical properties. In support, Chapter 2 demonstrates that virus species significantly influences virus transfer to and from fomites. Similarly, the work by Abad et al. (1994) demonstrates that virus species also influence virus persistence on fomites. The physicochemical properties of virus, therefore, are influential in both virus transfer between surfaces and virus persistence. This is consistent with work demonstrating the importance of physicochemical properties of virus movement and fate in the subsurface Dowd et al. (1998). CHAPTER 6. CONCLUSIONS 130 In Chapter 2, the characteristics that differed between the three bacteriophage tested are the isoelectric point and hydrophobicity. Both characteristics have been demonstrated to influence transport through the subsurface (Shields and Farrah, 2002). Unfortunately, as transfer of only three viruses was investigated, no trends between fraction transferred and either isoelectric point or hydrophobicity were elucidated. In the estimate of the inactivation rate for MS2 bacteriophage, discussed in Chapter 3, only one virus was studied and therefore no conclusions relating physicochemical properties to persistence can be drawn. Therefore, questions remain concerning the cause of the difference in transfer between the viruses and whether or not the difference would be applicable to transfer of animal virus. Similar questions concerning whether or not physicochemical properties of virus influence viral persistence on surfaces, and whether or not the differences in either persistence or transfer contribute to increased efficacy in fomite-mediated transmission. Incorporating secondary transmission into quantitative microbial risk assessments of fomite-mediated transmission Chapter 3 models the risk of infection for a single individual interacting with a contaminated fomite. The model is among the first to incorporate complex humanfomite interactions, including sporadic, sequential contact events, into a quantitative microbial risk assessment. However, the model only focuses on half of the fomitemediated transmission route (steps 2-4 of Figure 1.2). The model ignores the steps leading to contamination of the fomite. Therefore, questions remain concerning the influence of shedding of virus to fomites on resulting infectious disease transmission, and the role of fomites in secondary infections (person-to-person spread). Agent-based modeling provides a framework for modeling secondary transmission rates due to indirect contact. Recently, infectious disease modeling has incorporated environmental reservoirs into compartmental modeling (Li et al., 2009; Stilianakis and Drossinos, 2010). In compartmental modeling, rate parameters drive movement of individuals between compartments. In this manner, compartmental modeling assumes homogeneity and perfect mixing within compartments, as well rates of transfer CHAPTER 6. CONCLUSIONS 131 between compartments (Rahmandad and Sterman, 2008). Fomite-mediated transmission, however, is an inherently spatial phenomenon: infection from a fomite can not occur unless both an infected individual and a susceptible individual contact the same fomite in that order. Therefore, an alternative method, such as agent-based modeling, might prove more useful in understanding fomite-mediated transmission. Agent based modeling allows for heterogeneity across individuals, as well as among the network of their interactions (Rahmandad and Sterman, 2008). Defining both individuals as agents capable of moving within predefined ranges, and fomites as stationary agents, a framework for agent-based modeling of infectious disease transmission is suggested. The work of Chapter 3 lays the groundwork for agent-based modeling of fomitemediated transmission. The child modeled in Chapter 3 is provided a defined set of parameters (e.g., frequency and sequence of fomite contacts, likelihood of infection given a dose). Similarly, the fomite parameters are predefined (e.g., inactivation rate of virus, fraction virus transferred on contact). Replicating those agents, and defining additional parameters (e.g., frequency of contacts between agents, shedding of agents to others), would be among the first steps toward development of an agent-based model. Using a representative closed system, such as a nursing home, office, or child care center, could then provide an opportunity to validate the results by comparing predicted outbreak patterns to documented patterns, as seen in (Bartlett III et al., 1988; Iizuka, 2006). Once fomite-mediated transmission is modeled, characteristics of fomite-mediated transmission could be explored. Additionally, interventions could be implemented in the model to characterize likelihood of success in laboratory or field settings. Relative contribution of transmission routes to total respiratory and gastrointestinal disease burden Chapter 5 identified a significant association between respiratory illness and microbial contamination on surfaces. The daily resolution of health data provided an opportunity to investigate causal links between illness and fomites. However, no data were collected on direct contact, common vehicle, or airborne transmission. CHAPTER 6. CONCLUSIONS 132 Although there is a significant association between health outcomes and microbial contamination, the proportion of total respiratory illness attributable to indirect contact transmission is unknown. Therefore, the question of the relative contribution of transmission routes, and fomite-mediated transmission in particular, remains. Efforts to develop a systematic, evidence-based approach, to understanding relative contributions of transmission routes will contribute to development of interventions to reduce overall disease burden. Based on the sampling method of the work in Chapter 5, a multi-route exposure study could be performed. Airborne and common vehicle transmission could be monitored by incorporating personal air monitoring and replicate food/water diets. Simultaneously, data on subjects’ behaviors (e.g., contacts with other subjects, movement within the facility, contact with surfaces) could be collected via third-person or videographic observations (Ferguson et al., 2006). An individual’s likelihood of illness could then be modeled as a function of both their behaviors and the presence of etiological agents in the environment. The modeling could provide insight into the relative contributions of transmission routes to disease burden. Estimating the contribution of heterogeneous fomite use to variability in infection risk Current sampling protocol for estimating virus contamination on fomites relies on subjective sampling choice. In Chapter 4, we identified over 40 unique publications investigating virus contamination on fomites. In the publications, as in the fomites sampled in Chapter 5, the fomite choice for sampling was subjectively chosen by the research staff. No publication included a sampling protocol that identified the fomites that should be sampled prior to the study. However a fomite’s contribution to disease transmission is likely a function of its use as well as the presence of microbial contamination. For example, in Chapter 5 we modeled the interaction of a child with a contaminated toy ball. If the same child was in a room of multiple toys, and only one or a couple were contaminated, the child’s choice would increase the variability in the likelihood of infection. By choosing CHAPTER 6. CONCLUSIONS 133 to interact with an uncontaminated toy, the child would reduce his risk of infection to zero. Additional evidence is provided by Jiang et al. (1998), who suggested that surfaces likely to be contacted by children were more likely contaminated with a DNA marker seeded into a child care center. Understanding the heterogeneity of fomite use would improve understanding of fomite-mediated tranmsission. Fomite use could be identified through, for example, sensors or videographic techniques. The data gleaned could then be incorporated in sample choice for future fomite contamination studies, as well as in the interpretation of results. Additionally, identification of highly used fomites could be used to tailor environmental hygiene interventions. Extension of models to nosocomial bacterial infections Fomite-mediated transmission is an important route of nosocomial bacterial infections, and future research should extend the presented work to investigate transmission of bacterial infections in hospitals. The focus of the dissertation is on indoor transmission of viral respiratory and gastrointestinal disease. Motivation for the focus is provided in Chapter 1, and includes the notion that viral infections, unlike bacterial infections, can not readily be treated with antibiotics. Antibacterialresistant bacteria, however, are more frequently responsible for nosocomial infections. The most common examples, methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococci (VRE), are believed to be readily transmitted via fomites. Evidence of detection and persistence of MRSA and VRE on surfaces supports the likelihood of fomite-mediated transmission . There are both clinical and economic incentives for curbing nosocomial bacterial infections. Approximately 1.7 million hospital acquired infections occurred in the U.S. hospitals alone in 2002 (Klevens et al., 2007). This comes at an estimated cost of $700$2000 per case (Graves et al., 2008). Reductions in total healthcare expenditures, therefore, may be achieved through increased infection control targeted to effectively interrupting transmission routes (Graves et al., 2008). Future work extending the dissertation to understanding transmission of hospital-acquired infections may aid in the design and implementation of interventions for infection control. Appendix A Supplemental Material for Chapter 3: Equations Used in Discrete-Time Model Equations used to represent fomes-mouth contacts: Change in concentration on both hands: CH(tc ) = CH(tc−1 ) e(−kh ∆t) (A.1) Change in concentration on fomes: CF (tc ) = (1 − T EF M SF )CF (tc−1 ) e(−kf ∆t) (A.2) Increase in dose: DOSE = T EF M SF AF CF (tc−1 ) e(−kf ∆t) 134 (A.3) APPENDIX A. SUPPLEMENTAL MATERIAL FOR CHAPTER 3 135 Equations used to represent hand-fomes contacts: Change in concentration on hand in contact with fomes: CH(tc ) = CH(tc−1 ) e(−kh ∆t) − T EF H SH (CH(tc−1 ) e(−kh ∆t) − CF (tc−1 ) e(−kf ∆t) ) (A.4) Change in concentration on hand not in contact with fomes: CH(tc ) = CH(tc−1 ) e(−kh ∆t) (A.5) Change in concentration on fomes: CF (tc ) = CF (tc−1 ) e(−kf ∆t) − T EF H SH AH (CF (tc−1 ) e(−kf ∆t) − CH(tc−1 ) e(−kh ∆t) ) AF (A.6) Increase in dose: Dose = 0 (A.7) Equations used to represent hand-mouth contacts: Change in concentration on hand in contact with mouth: CH(tc ) = (1 − T EHM SM )CH(tc−1 ) e(−kh ∆t) (A.8) Change in concentration on hand not in contact with mouth: CH(tc ) = CH(tc−1 ) e(−kh ∆t) (A.9) Change in concentration on fomes: CF (tc ) = CF (tc−1 ) e(−kf ∆t) (A.10) DOSE = T EHM SM AH CH(tc−1 ) e(−kh ∆t) (A.11) Increase in dose: APPENDIX A. SUPPLEMENTAL MATERIAL FOR CHAPTER 3 Variables tc tc−1 ∆t = tc - tc−1 CH CF kh kf T EF M T EF H T EHM SF SH SM AF AH DOSE = = = = = = = = = = = = = = = = time of the current contact time of the previous contact time between successive contacts concentration of virus on surface of hand, virus/cm2 concentration of virus on surface of fomes, virus/cm2 inactivation rate of virus on hand, s−1 , base e inactivation rate of virus on fomes, s−1 , base e fraction of virus transferred from fomes to mouth fraction of virus transferred between fomes and hand fraction of virus transferred from hand to mouth fraction of surface area of fomes in contact with mouth fraction of surface area of hand in contact with fomes fraction of surface area of hand in contact with mouth surface area of fomes, cm2 surface area of hand, cm2 number of viral particles ingested 136 Appendix B Supplemental Material for Chapter 4: Virus Recovery from Fomites B.1 Tables 137 Year 1995 1999 2007 1998 First Author Akhter Asano Bausch Bellamy Yes Yes Yes AdenoEnteroAdeno- No Yes Ebola Entero- Yes Ebola Yes Yes Influenza VZV Yes Ill Rhino- Virus RTPCR RTPCR culture nPCR EIA culture culture culture EIA Assay Rayon Polyester Polyester Cotton Swab Swab Swab Swab Swab Implem Saline PBS PBS RPMI1640 TPB + Anti TPB + Anti TPB + Anti TPB + Anti TPB + Anti Eluent toilet bowl telephone, tap handle, above same as cryoprobes forceps, gloves, filter conditioner button, air push- channel television table, handle, door N/A N/A N/A N/A sign chart toys, vital televisions, handles, toilet Surfaces 3 1 0 18 0 0 0 0 12 Pos. No. 448 28 28 72 146 146 146 146 146 Total Frac 0.01 0.04 <0.04 0.25 <0.01 <0.01 <0.01 <0.01 0.08 Pos No No No Yes No No No No No LOD Home Hospital Hospital Home Hospital Hospital Hospital Hospital Hospital Locale APPENDIX B. SUPPLEMENTAL MATERIAL FOR CHAPTER 4 138 Year 2005 2005 2009 2009 First Author Boone Booth Boxman Boxman No Influenza Noro- NoroYes Yes Yes No Influenza SARS-CoV Yes Ill Influenza Virus qRTPCR qRTPCR RTPCR RTPCR RTPCR RTPCR Assay Antistatic Antistatic Polyester Polyester Polyester Polyester Implem Ringer’s Ringer’s VTM Saline LB LB Eluent door button, elevator handrail, telephone, cash desk, knife grips toilet seats, cntrl remote table, controls remote switches, toilets light keyboards, phones, dooknobs, handles, above same as areas changing diaper floors, counters, drains, dishcloths, handles, seats, fauct toys, toilet Surfaces 48 3 3 54 25 58 Pos. No. 119 6 85 92 109 109 Total Frac 0.40 0.50 0.04 0.59 0.23 0.53 Pos Yes Yes No No No No LOD Ship Ship Hospital Home DCC DCC Locale APPENDIX B. SUPPLEMENTAL MATERIAL FOR CHAPTER 4 139 Year 2009 1993 2002 2000 First Author Bright Butz Carducci Cheesebrough Yes No Noro- Yes Noro- HCV Yes Yes NoroRota- Yes Ill Influenza Virus nRTPCR nRTPCR nRTPCR RTPCR RTPCR RTPCR Assay Cotton Cotton Cotton Cotton Rayon Rayon Implem VTM VTM BE PBS Amies Amies Eluent above same as cushions phones, tables, toilet, carpet, holder test tube test tube, plastic toys handle, handle, sink toilet fountain, telephone, above same as dispensers tops, towel counter- handles, doorknobs, computers, desks, Surfaces 0 61 2 14 9 13 Pos. No. 144 144 42 91 55 54 Total Frac <0.01 0.42 0.05 0.15 0.16 0.24 Pos No No Yes No No No LOD Hotel Hotel Hospital DCC Class Class Locale APPENDIX B. SUPPLEMENTAL MATERIAL FOR CHAPTER 4 140 Year 2004 2008 2004 1989 2008 2003 2005 First Author Chen Diggs Dowell Ferenczy Fischer Froio Gallimore Yes Yes HCV Astro- Yes Yes HBV HCV Yes Yes SARS-CoV HPV Yes Yes Yes Ill SARS-CoV Noro- SARS-CoV Virus hnRTPCR RTPCR antigen RTPCR DBH culture RTPCR RTPCR RTPCR Assay Cotton Cotton Cotton Cotton Polyester Polyester Polyester Swab Cotton Implem PBS Saline Saline BE Saline VTM VTM N/R VTM Eluent equip. medical toilet taps, console, cuff pressure blood machine dialysis (glass) crack pipes cryoprobes forceps, gloves, N/A handrail mouse, computer N/R (plug) outlet bedding, table, bookshelf, chair, bedside fountain, water drinking Surfaces 4 1 1 1 36 0 22 1 9 Pos. No. 12 64 64 51 100 90 90 25 100 Total Frac 0.33 0.02 0.02 0.02 0.36 <0.01 0.24 0.04 0.09 Pos No No Yes Yes Yes No No No Yes LOD Hospital Hospital Hospital Outside Hospital Hospital Hospital Class Hospital Locale APPENDIX B. SUPPLEMENTAL MATERIAL FOR CHAPTER 4 141 Year 2006 2008 2008 2006 First Author Gallimore Gallimore Girou Goldhammer Rhino- HCV No No No No Rota- Rota- Yes Rota- No No Noro- Noro- Yes Noro- No No Astro- Astro- Yes Ill Astro- Virus culture nRTPCR nRTPCR hnRTPCR hnRTPCR nRTPCR nRTPCR nRTPCR nRTPCR hnRTPCR hnRTPCR Assay Cotton N/R Cotton Cotton Cotton Cotton Cotton Cotton Cotton Cotton Cotton Implem VTM N/R Saline Saline Saline Saline Saline Saline Saline Saline Saline Eluent equipment exercise table machine, dialysis telephone switch, light toilet tap, above same as light switch toilet tap, above same as above same as above same as above same as above same as equip. medical phone, toilet taps, console, game Surfaces 63 6 28 12 3 17 7 7 21 2 4 Pos. No. 100 82 242 242 242 121 33 99 55 66 88 Total Frac 0.63 0.07 0.12 0.05 0.01 0.14 0.21 0.07 0.38 0.03 0.05 Pos No Yes No No No No No No No No No LOD Gym Hospital Hospital Hospital Hospital Hospital Hospital Hospital Hospital Hospital Hospital Locale APPENDIX B. SUPPLEMENTAL MATERIAL FOR CHAPTER 4 142 Year 1998 2007 1982 2008 2007 2009 1983 First Author Green Gurley Gwaltney Hamada Jones Kawahara Keswick Rota- MCV Noro- Adeno- Rhino- Nipah Noro- Virus No Yes Yes Yes Yes No Yes Ill antigen nPCR RTPCR RTPCR culture RTPCR RTPCR Assay Cotton Cotton Rayon N/R Cotton Cotton Cotton Implem MEM + Anti PBS Amies Gel N/R BHIB MEM + Anti VTM Eluent sink, hands door knob, diaper pail, keyboard chair, drawers, toy pianaca, locker, desk, doorknobs surfaces, kitchen surfaces, bathroom lamp equiprment, medical frame, lens, glasses tiles frame wall, bed commodes curtains, lockers, Surfaces 4 8 11 17 20 11 11 Pos. No. 25 9 14 21 47 468 36 Total Frac 0.16 0.89 0.79 0.81 0.43 0.02 0.31 Pos No No No No No No No LOD DCC Home Boat Hospital Home Hospital Ward Locale APPENDIX B. SUPPLEMENTAL MATERIAL FOR CHAPTER 4 143 Year 2002 2009 2009 2008 2009 2009 First Author Kuusi Lederman Lessa Lopez Lyman Pappas No No Influenza Yes RotaRSV Yes NoroNo Yes Astro- Picorna- Yes Yes Yes Yes Yes Ill Adeno- VZV Adeno- Orthopox- Noro- Virus RTPCR RTPCR RTPCR RTPCR RTPCR RTPCR PCR qPCR qPCR qPCR RTPCR Assay Cotton Cotton Cotton N/R N/R N/R N/R Polyester Swab Polyester Cotton Implem BHIB + BSA BHIB + BSA BHIB + BSA N/R N/R N/R N/R PBS Saline Saline PBS Eluent toys toys toys N/R N/R N/R N/R fixtures carts, and frames, chairs, bed dust from keyboard cabinet, table, bedside bedrain, seat, cup seat, car booster nighstand, toys, slipper, washcloth, ointment, toilet seat handle, door bathroom handle, ultrasound Surfaces 0 0 10 38 11 9 16 18 7 8 4 Pos. No. 18 17 52 38 40 45 27 26 37 25 30 Total Frac <0.06 <0.06 0.19 >0.97 0.28 0.20 0.59 0.69 0.19 0.32 0.13 Pos Yes Yes Yes No No No No Yes No No No LOD Hospital Hospital Hospital DCC DCC DCC DCC Hospital Hospital Home Hotel Locale APPENDIX B. SUPPLEMENTAL MATERIAL FOR CHAPTER 4 144 Year 1987 2007 2006 1999 First Author Piazza Runner Russell Soule Yes Yes HAV HCV Rota- Yes No Yes HBV Adeno- Yes HIV No No Influenza HBV Ill Virus RTPCR TIGER PCR PCR PCR PCR antigen RTPCR Assay Cotton Polyester Swab Swab Swab Swab Cotton Cotton Implem MEM VTM Water Water Water Water + BSA Saline BHIB + BSA Eluent washbasins struments, medical in- tables, cloths, cleaning playmats, handles, rifles lockers, pillows, above same as above same as above same as container sharps disposable arms headrest, dental chair benches, work toys Surfaces 22 163 4 2 2 3 12 1 Pos. No. 45 629 30 30 30 30 190 18 Total Frac 0.49 0.26 0.13 0.07 0.07 0.10 0.06 0.06 Pos No No No No No No No Yes LOD Hospital Military Hospital Hospital Hospital Hospital Dentist Hospital Locale APPENDIX B. SUPPLEMENTAL MATERIAL FOR CHAPTER 4 145 Year 2002 2002 1992 2007 2005 2001 First Author Strauss Widdowson Wilde Winther Wu Yoshikawa VZV Noro- Yes Yes No Yes Rota- Rhino- No Yes No Ill Rota- Rota- HPV Virus nPCR RTPCR RTPCR RTPCR RTPCR RTPCR nPCR Assay Cotton Cotton Cotton Cotton Cotton Cotton Cotton Implem RPMI1640 Saline BHIB + BSA BHIB BHIB PBS PBS Eluent door chair, table, back of rail seat, bed table, toilet button, elevaot phones, etc. faucet, pens, handles, Door area floor, child toy balls, child area toy balls, lightswitch cupboard, linen door handle bathroom lamp, nation bed, exami- amination panel, ex- bed control Cyroguns, Surfaces 11 5 52 15 2 2 37 Pos. No. 27 10 150 57 65 94 102 Total Frac 0.41 0.50 0.35 0.26 0.03 0.02 0.36 Pos Yes No No No No No Yes LOD Home LTC Hotel DCC DCC Hospital Hospital Locale APPENDIX B. SUPPLEMENTAL MATERIAL FOR CHAPTER 4 146 First Author Year Virus Assay Implem Eluent Surfaces Pos. No. viations provided in Table B.2. Table B.1: A summary of the articles included in the analysis using the abbre- Ill Total Frac Pos LOD Locale APPENDIX B. SUPPLEMENTAL MATERIAL FOR CHAPTER 4 147 APPENDIX B. SUPPLEMENTAL MATERIAL FOR CHAPTER 4 Amies Anti BE BHIB BSA Culture DBH DCC EIA Eluent HAV HBV HCV HIV hnPCR HPV Ill Implem LB LOD LTC MCV MEM nPCR / nRTPCR N/R PBS PCR qPCR / qRTPCR Ringer’s RSV RT-PCR SARS-CoV Surfaces Swab Total TPB VTM VZV 148 Amies medium Antibiotics Beef extract Brain heart infusion broth Bovine serum albumin Cell culture Dot blot hybridization Day care center Enzyme immunoassay Eluent type used Hepatitis A virus Hepatitis B virus Hepatitis C virus Human immunodeficiency virus Hemi-nested PCR Human papillomavirus Clinically infected individual was present Implement type used Letheen broth Lower limit of detection is reported in article Long Term Care Molluscum contagiosum virus Minimal essential medium Nested PCR / RTPCR Not reported by author Phosphate buffered saline Polymerase chain reaction quantitative PCR / RTPCR 1/4 strength Ringer’s solution Respiratory syncytial virus Reverse transcription PCR Severe acute respiratory syndrome - corona virus Type of surfaces with detectable target Unreported type of swab Total number of surfaces sampled Tryptose phosphate broth Viral transport medium Varicella zoster virus Table B.2: The abbreviations, and corresponding definitions, used in Table B.1 and Table B.3. APPENDIX B. SUPPLEMENTAL MATERIAL FOR CHAPTER 4 Virus No. Studies Adeno6 Astro5 Ebola 2 2 EnteroHAV 1 HBV 3 HCV 5 1 HIV HPV 2 Influenza 7 MC 1 Nipah 1 14 NoroOrthopox1 Picorna1 Rhino3 Rota11 1 RSV SARS-CoV 4 VZV 3 Total 74 Samples Collected 1006 453 56 594 30 284 269 30 202 546 9 468 1019 25 52 297 957 17 365 125 6804 No. Pos. 203 22 1 3 2 15 14 3 73 151 8 11 204 8 10 135 161 0 34 47 1105 149 Frac. Pos. .202 .049 .018 .005 .067 .053 .052 .100 .361 .277 .889 .024 .200 .320 .192 .455 .168 .000 .093 .376 .162 Table B.3: A summary of the articles included in the analysis using the abbreviations provided in Table B.2 Appendix C Supplemental Material for Chapter 5: Fomites and Health in Child Care Centers C.1 Methods C.1.1 Statistics Most statistics were performed using PASW Statistics 18.0.2. (SPSS: An IBM Company, Chicago, IL, USA). A significance level of α <0.05 was used throughout the study. Survey. The results of the surveys were compared across two sites using a Z-test for two proportions for percentage data (pet ownership, medication, and chronic illness), Pearson’s χ2 for categorical data (ethnicity, type of household), and Mann-Whitney U test or Kruskal-Wallis one-way analysis of variance for two or more independent samples of ordinal data( Changes in parent and child handwashing rates and healthy child index between the baseline and follow-up surveys were examined using a matched-pair t-test. 150 APPENDIX C. SUPPLEMENTAL MATERIAL FOR CHAPTER 5 151 Microbial Contamination. Significance of self-reported health on microbial contamination was determined using Kruskal-Wallis one-way analysis of variance. Mann-Whitney U tests were used to assess significance of associations of hand contamination with visible signs (dirt on hands, dirt under fingernails, and runny nose), time of class (morning or afternoon), site (A or B), location of sampling (indoor or outdoor). Mann-Whitney U tests were also used to assess significance of associations of environmental contamination with time of class (morning or afternoon), site (A or B), and location of sampling (indoor or outdoor). Bivariate Correlations Bivariate Spearman rank correlations were used to investigate correlations between environmental contamination (weekly fraction of samples with /geq5 CFU enterococci or fecal coliform), hand contamination (daily and weekly fraction of samples with /geq5.4 CFU enterococci or fecal coliform per two hands), and health (daily and weekly respiratory illness, gastrointestinal illness, absences, illness-related absences, and number of unique illness episodes). Results from the bivariate correlations were then used to identify fecal indicator bacteria (e.g., enterococci, fecal coliform, or E. coli ) and health outcome (e.g. respiratory illness, gastrointestinal illness, absences, illness-related absences or unique illness episodes) to include in model of health outcome as a function of microbial contamination. Health Outcome as a Function of Contamination To investigate associations between surface contamination and health, health outcomes were modeled as functions of the density of enterococci on hands, the presence / absence of enterococci on at least one sampled environmental surface, and site. Preliminary bivariate correlations suggest enterococci is a more appropriate indicator of hand and surface contamination then fecal coliform and E. coli, so enterococci was used as the dependent variable. Inter - individual correlation in the longitudinal data was accounted for using generalized estimating equations (GEE) with a logit link APPENDIX C. SUPPLEMENTAL MATERIAL FOR CHAPTER 5 152 function clustered on individual St Sauver et al. (1998). We assumed data correlation was independent in time, and use a compound correlation structure. The GEE analysis was performed using the “geeglm” function in the “geepack” package in R (version 2.11.1, R Foundation for Statistical Computing, Vienna, Austria). To infer casual links between surface (hand and environmental) contamination and health (where respiratory illness has an incubation time of 2-5 days Long et al. (1997)), we used separate GEEs to model health outcomes as a function of microbial contamination at daily lags of up to plus and minus seven days. Only the subset of data with measured health outcomes and corresponding microbial contamination data at the specified lag was included (i.e., missing values were removed from analysis). Health outcomes explored included illness-related absences, respiratory illness, and onset of unique illness episodes. The low prevalence of gastrointestinal illness during the study (see Results) precluded analysis of gastroenteritis. C.2 Results C.2.1 Bivariate Correlations C.2.2 Hand Contamination and Health Data. Correlations between hand contamination and health data were performed by aggregating each individual’s data over the duration of the study and investigating associations between the fraction of total days individuals experienced an adverse health outcome to the mean density of bacteria on their hands over the duration of the study. The sample size was, therefore, seventy-seven, equal to the number of individuals who were both enrolled in the study and assented to at least one hand sample. Enterococci on hands was significantly correlated to respiratory illness (ρs = 0.276, p = 0.015), but not gastrointestinal illness(ρs = 0.034, p = 0.771), absences (ρs = −0.047, p = 0.688), illness-related absences (ρs = 0.046, p = 0.693), or number of unique illnesses (ρs = −0.093, p = 0.422). Neither the density of fecal coliform nor E. coli on hands were significantly correlated to any health outcomes. Specifically, fecal APPENDIX C. SUPPLEMENTAL MATERIAL FOR CHAPTER 5 153 coliform was not significantly correlated to respiratory illness (ρs = 0.128, p = 0.267), gastrointestinal illness(ρs = 0.148, p = 0.199), absences (ρs = −0.041, p = 0.726), illness-related absences (ρs = −0.022, p = 0.850), or number of unique illnesses (ρs = −0.031, p = 0.786). Similarly, E. coli was not significantly correlated to respiratory illness (ρs = 0.122, p = 0.290), gastrointestinal illness(ρs = −0.056, p = 0.626), absences (ρs = 0.034, p = 0.769), illness-related absences (ρs = 0.127, p = 0.270), or number of unique illnesses (ρs = −0.140, p = 0.224). C.2.3 Hand Contamination and Environmental Contamination. The fraction of environmental fomites with detectable enterococci on a given day was not correlated with corresponding fraction of hand samples with detectable enterococci (ρs = 0.206, p = 0.108). Similarly, there was no significant correlation for fecal coliform (ρs = 0.069, p = 0.591). C.2.4 Environmental Contamination and Health Data. The fraction of samples with detectable enterococci at each facility is significantly correlated with the fraction of individuals with respiratory illness symptoms during the same week (ρs = 0.297, p = 0.018), but is not significantly correlated with gastrointestinal illness (ρs = 0.157, p = 0.218), absences (ρs = −0.151, p = 0.237), illness-related absences (ρs = 0.039, p = 0.761), or number of unique illness episodes (ρs = 0.126, p = 0.326). There are no significant correlations for fecal coliform and respiratory illness, (ρs = −0.023, p = 0.859), gastrointestinal illness, (ρs = 0.045, p = 0.728), absences (ρs = 0.066, p = 0.606), illness-related absences (ρs = 0.088, p = 0.490), or number of unique illness episodes (ρs = 0.170, p = 0.184). APPENDIX C. SUPPLEMENTAL MATERIAL FOR CHAPTER 5 C.2.5 154 Health Associations with Hand and Surface Contamination Using generalized estimating equations to model illness-related absences as a function of hand and environmental contamination, while controlling for site and incorporating daily lags did not elucidate clear trends. Illness-related absences is significantly associated with hand and environmental contamination only sporadically (See Table C.2) and is likely a result of false positive detection. Specifically, illness-related absences is negatively associated with hand contamination as measured five days previously (β = −0.958, p = 0.039) and positively associated with environmental surface contamination as measured three (β = 2.34, p < 0.001) and seven days later (β = 0.546, p = 0.021). New episodes of illness as a health outcome was also significantly associated with hand and environmental contamination and hand hygiene only sporadically (See Table C.1). False positive detection likely explains significant associations. Specifically, new episodes of illness were positively associated with hand contamination as measured on the same day (β = 0.613, p = 0.019) and with environmental surface contamination as measured two days later (β = 1.49, p = 0.0136). C.3 C.3.1 Discussion Use of Multiple Comparisons The use of fifteen GEEs to model health outcome at daily lags requires use of multiple comparisons. The expected number of false positives is two, as estimated for fifteen models with two variables (excluding intercept and adjustment for site) and a significance threshold of 0.05. In the model of respiratory illness, 9 of the 30 total variables are significant. The likelihood of a false positivity rate of 8 or more variables, relying on the assumption that all tests are independent, is less than 0.001% Storey (2003). Therefore, most of the 9 significant variables are likely true positives. The clustering of significant associations around specific lags (e.g. hands are significantly associated with respiratory illness on -1,0, and +1 days, fomites on -4, and -5 days as well as +2 APPENDIX C. SUPPLEMENTAL MATERIAL FOR CHAPTER 5 155 and +3 days) provides evidence that the findings are likely true. Random significant associations, such as hands at a lag of +7 days are more likely false. Conversely, the likelihood of 3 or more significant associations out of 30 significant tests, as observed in the two models for new episodes of illness and illness-related absences is approximately 19%, suggesting that the majority of significant findings for the two models are false positives. APPENDIX C. SUPPLEMENTAL MATERIAL FOR CHAPTER 5 C.4 Tables. 156 Obs. 439 395 328 214 204 284 455 572 472 365 233 193 231 370 440 Ill. 18 10 8 7 10 9 19 14 13 11 8 15 13 18 15 Coef. -3.01 -2.89 -4.41 -2.79 -4.24 -4.75 -3.34 -4.03 -3.03 -2.84 -3.39 -3.14 -2.72 -2.71 -3.71 Intercept SE Pr(>|W|) 0.32 <0.001 0.40 <0.001 0.68 <0.001 0.50 <0.001 0.66 <0.001 1.02 <0.001 0.45 <0.001 0.50 <0.001 0.45 <0.001 0.49 <0.001 0.51 <0.001 0.47 <0.001 0.42 <0.001 0.39 <0.001 0.55 <0.001 Coef. -0.30 0.10 0.28 -0.32 0.53 -0.60 -0.33 0.62 0.11 -0.80 -0.08 0.48 -0.73 -0.71 0.01 Hands SE Pr(>|W|) 0.33 0.361 0.32 0.747 0.41 0.484 0.38 0.402 0.47 0.255 0.67 0.372 0.40 0.409 0.25 0.014 0.40 0.784 0.61 0.187 0.42 0.850 0.31 0.118 0.39 0.060 0.44 0.104 0.41 0.990 Coef. 0.66 -1.61 0.23 0.62 1.11 0.58 0.02 0.54 -1.71 -0.11 -0.12 -0.16 0.19 0.47 -0.20 Fomites Site Correlation SE Pr(>|W|) Coef. SE Pr(>|W|) Estimate SE 0.42 0.120 -0.58 0.56 0.307 0.00 0.04 1.12 0.151 -1.97 1.10 0.075 0.00 0.02 0.70 0.736 0.74 0.69 0.284 -0.02 0.05 0.86 0.468 -42.98 0.42 <0.001 -0.08 0.13 0.57 0.050 1.00 0.69 0.147 0.03 0.06 0.52 0.267 1.97 1.10 0.073 0.00 0.03 0.45 0.966 0.66 0.55 0.226 0.03 0.04 0.48 0.258 -0.99 0.63 0.117 0.00 0.01 0.94 0.071 -0.70 0.54 0.200 -0.02 0.04 0.55 0.845 -0.55 0.63 0.381 0.00 0.02 0.60 0.837 0.32 0.81 0.693 0.03 0.06 0.76 0.835 0.72 0.56 0.204 -0.06 0.07 0.76 0.800 0.37 0.59 0.537 -0.03 0.07 0.43 0.280 -0.15 0.48 0.756 -0.02 0.03 0.66 0.766 0.70 0.61 0.248 0.03 0.06 Table C.1: New episodes of illness model parameters for generalized estimating equation as function of enterococci on hands and enterococci on surfaces while controlling for site. Lags of up to plus and minus seven days between illness and contamination are modelled separately. “Lag” is the number of days between collection of health data and the collection of data for contamination on hands, and fomites, so a positive lag implies that health outcomes preceeded microbial contamination use and negative lags imply that health data succeeded microbial contamination. “Obs.” refers to the number of observations with both health measurements and microbial contamination data at the specified lag. ”Ill.” refers to the number of observations in which new episodes were observed,“Hands” is the number of enterococci detected on an individual’s hands, “Fomites” is the fraction of surfaces sampled with detectable enterococci, “Site” is the facility, with the coefficients representative of the difference in the model for Site B relative to Site A, “Coef.” is the coefficient on the variable, “SE” is the standard error of the coefficent, and “Pr(>|W|)” is the significance of the coefficient, with values less than 0.05 considered significant and highlighted in bold. Lag 7 6 5 4 3 2 1 0 -1 -2 -3 -4 -5 -6 -7 APPENDIX C. SUPPLEMENTAL MATERIAL FOR CHAPTER 5 157 Obs. 434 391 326 212 203 283 454 469 361 228 193 231 368 438 Ill. 31 19 20 7 13 17 12 20 18 9 12 17 27 25 Coef. -2.89 -2.64 -3.05 -2.99 -3.15 -4.05 -4.03 -3.06 -2.64 -3.73 -2.39 -3.27 -2.86 -3.23 Intercept SE Pr(>|W|) 0.37 <0.001 0.39 <0.001 0.40 <0.001 0.62 <0.001 0.57 <0.001 0.67 <0.001 0.58 <0.001 0.42 <0.001 0.43 <0.001 0.84 <0.001 0.40 <0.001 0.56 <0.001 0.33 <0.001 0.42 <0.001 Hands Coef. SE Pr(>|W|) -0.13 0.29 0.664 0.08 0.35 0.811 -0.05 0.32 0.876 -0.64 0.54 0.237 -0.62 0.52 0.239 0.29 0.34 0.398 -0.03 0.36 0.929 0.39 0.26 0.133 -0.93 0.51 0.068 0.47 0.39 0.228 -0.36 0.48 0.452 -0.59 0.29 0.043 0.31 0.22 0.152 -0.12 0.39 0.758 Coef. 0.53 -0.53 0.33 0.47 2.17 0.43 0.28 -0.41 -0.86 -0.26 0.83 0.09 0.45 0.34 Surfaces SE Pr(>|W|) 0.24 0.023 0.46 0.249 0.41 0.411 0.74 0.523 0.56 <0.001 0.31 0.171 0.47 0.558 0.35 0.241 0.59 0.142 0.54 0.632 0.69 0.230 0.55 0.871 0.35 0.198 0.41 0.413 Coef. 0.37 -0.58 0.45 -0.81 -0.15 1.44 0.58 -0.54 0.59 0.65 -0.79 1.49 -0.13 0.60 Site Correlation SE Pr(>|W|) Estimate SE 0.44 0.401 0.05 0.04 0.51 0.254 -0.02 0.03 0.40 0.255 -0.07 0.04 0.78 0.296 0.06 0.24 0.58 0.792 -0.03 0.10 0.63 0.022 -0.01 0.06 0.62 0.356 -0.01 0.03 0.59 0.353 0.09 0.13 0.50 0.235 -0.01 0.05 0.78 0.406 0.10 0.34 0.58 0.173 -0.04 0.06 0.66 0.024 0.01 0.03 0.34 0.704 -0.07 0.05 0.40 0.136 -0.02 0.02 Table C.2: Illness-related absences model parameters for generalized estimating equation as function of enterococci on hands and enterococci on surfaces, while controlling for site. Lags of up to plus and minus seven days between illness and contamination are modelled separately. “Lag” is the number of days between collection of health data and the collection of data for contamination on hands and fomites, so a positive lag implies that health outcomes preceeded microbial contamination and negative lags imply that health data succeeded microbial contamination. “Obs.” refers to the number of observations with both health measurements and microbial contamination data at the specified lag. ”Ill.” refers to the number of observations in which illness-absences was observed,“Hands” is the number of enterococci detected on an individual’s hands, “Fomites” is the fraction of surfaces sampled with detectable enterococci, “Site” is the facility, with the coefficients representative of the difference in the model for Site B relative to Site A, “Coef.” is the coefficient on the variable, “SE” is the standard error of the coefficent, and “Pr(>|W|)” is the significance of the coefficient, with values less than 0.05 considered significant and highlighted in bold. Lag 7 6 5 4 3 2 1 0 -1 -2 -3 -4 -5 -6 -7 APPENDIX C. SUPPLEMENTAL MATERIAL FOR CHAPTER 5 158 Bibliography Abad, F. X., R.M., P. and Bosch, A. (1994) Survival of enteric viruses on environmental fomites. Appl. Environ. Microbiol. 60, 3704–3710. Abad, F. X., Villena, C., Guix, S., Caballero, S., R.M., P. and Bosch, A. (2001) Potential role of fomites in the vehicular transmission of human astroviruses. Appl. Environ. Microbiol. 67, 3904–3907. Abbaszadegan, M., Mayer, B. K., Ryu, H. and Nwachuku, N. (2007) Efficacy of Removal of CCL Viruses under Enhanced Coagulation Conditions. Environ. Sci. Technol. 41, 971–977. Abbey, H. (1952) An examination of the Reed-Frost theory of epidemics. Hum. Biol. 24, 201–233. Adams, R. O. (2009) A review of the stainless steel surface. J. Vac. Sci. Technol. A 1, 12–18. Aiello, A. E., Coulborn, R. M., Perez, V. and Larson, E. L. (2008) Effect of hand hygiene on infectious disease risk in the community setting: a meta-analysis. Am. J. Public Health 98, 1372–1381. Akhter, J., al Hajjar, S., Myint, S. and Qadri, S. M. (1995) Viral contamination of environmental surfaces on a general paediatric ward and playroom in a major referral centre in Riyadh. Eur. J. Epidemiol. 11, 587–90. Alexander, C. S., Zinzeleta, E. M., Mackenzie, E. J., Vernon, A. and Markowitz, 159 BIBLIOGRAPHY 160 R. K. (1990) Acute gastrointestinal illness and child care arrangements. Am. J. Epidemiol. 131, 124. Ansari, S. A., Sattar, S. A., Springthorpe, V. S., Wells, G. A. and Tostowaryk, W. (1988) Rotavirus survival on human hands and transfer of infectious virus to animate and nonporous inanimate surfaces. J. Clin. Microbiol. 26, 1513–1518. Ansari, S. A., Springthorpe, V. S., Sattar, S. A., Rivard, S. and Rahman, M. (1991) Potential role of hands in the spread of respiratory viral infections: studies with human parainfluenza virus 3 and rhinovirus 14. J. Clin. Microbiol. 29, 2115–9. Asano, Y., Yoshikawa, T., Kanematsu, M., Ihira, M., Suzuki, K. and Suga, S. (1999) Rapid contamination with varicella-zoster virus DNA to the throat of a daycare attendee and environmental surfaces from a child with varicella. Pediatr. Int. 41, 233– 6. Ashburn, P. M. and Caraig, C. F. (1907) Experimental investigations regarding the etiology of dengue fever. J. Infect. Dis. 4, 440–475. Atkinson, M. P. and Wein, L. M. (2008) Quantifying the routes of transmission for pandemic influenza. B. Math. Biol. 70, 820–67. AuYeung, W. (2007)Analysis and Simulation of Young Children’s Hand and Mouthing Contact Behaviors.. Ph. d. thesis. Stanford University. AuYeung, W., Canales, R. and Leckie, J. (2006) Contact-specific fractional surface area of hand-to-mouth immersions. Epidemiol. 17, S468–S469. Barel, A. O., Paye, M. and Maibach, H. I. (2009) Handbook of Cosmetic Science and Technology. Informa HealthCare. Barker, J. (2001)The role of viruses in gastrointestinal disease in the home. J. Infection 43, 42–44. Baron, S. and Jennings, P. M. (1991) Medical Microbiology. Churchill Livingstone New York. BIBLIOGRAPHY 161 Bartlett III, A. V., Reves, R. R. and Pickering, L. K. (1988) Rotavirus in infanttoddler day care centers: Epidemiology relevant to disease control strategies1. J. Pediatr. 113, 435–441. Bausch, D. G., Towner, J. S., Dowell, S. F., Kaducu, F., Lukwiya, M., Sanchez, A., Nichol, S. T., Ksiazek, T. G. and Rollin, P. E. (2007)Assessment of the risk of Ebola virus transmission from bodily fluids and fomites. J. Infect. Dis. 196, 142–147. Beamer, P. (2007) Development of a Model to Estimate Aggregate and Cumulative Exposure and Dose in Young Children. Ph.d. thesis. Stanford University. Bean, B., Moore, B. M., Sterner, B., Peterson, L. R., Gerding, D. N. and Balfour, H. H., J. (1982) Survival of influenza viruses on environmental surfaces. J. Infect. Dis. 146, 47–51. Bell, A. N. (1901) Fomites and Yellow Fever. Public Health Pap. Rep. 27, 144. Bell, D. M. (2006)Non-pharmaceutical interventions for pandemic influenza, international measures. Emerg. Infect. Dis. 12, 81–7. Bellamy, K., Laban, K. L., Barrett, K. E. and Talbot, D. C. (1998) Detection of viruses and body fluids which may contain viruses in the domestic environment. Epidemiol. Infect. 121, 673–80. Bickley, J., Short, J. K., McDowell, D. G. and Parkes, H. C. (2008)Polymerase chain reaction (PCR) detection of Listeria monocytogenes in diluted milk and reversal of PCR inhibition caused by calcium ions. Lett. Appl. Microbiol. 22, 153–158. Bidawid, S., Malik, N., Adegbunrin, O., Sattar, S. A. and Farber, J. M. (2004) Norovirus cross-contamination during food handling and interruption of virus transfer by hand antisepsis: experiments with feline calicivirus as a surrogate. J. Food Protect. 67, 103–9. Boone, S. A. and Gerba, C. P. (2005)The occurrence of influenza A virus on household and day care center fomites. J. Infection 51, 103–9. BIBLIOGRAPHY 162 Boone, S. A. and Gerba, C. P. (2007)Significance of fomites in the spread of respiratory and enteric viral disease. Appl. Environ. Microbiol. 73, 1687–96. Boone, S. A. and Gerba, C. P. (2010) The Prevalence of Human Parainfluenza Virus 1 on Indoor Office Fomites. Food Env. Virology 2, 41–46. Booth, T. F., Kournikakis, B., Bastien, N., Ho, J., Kobasa, D., Stadnyk, L., Li, Y., Spence, M., Paton, S., Henry, B., Mederski, B., White, D., Low, D. E., McGeer, A., Simor, A., Vearncombe, M., Downey, J., Jamieson, F. B., Tang, P. and Plummer, F. (2005) Detection of airborne severe acute respiratory syndrome (SARS) coronavirus and environmental contamination in SARS outbreak units. J. Infect. Dis. 191, 1472–7. Bordner, R., Winter, J. and Scarpino, P. (1978)Microbiological methods for monitoring the environment. Water and wastes. Technical Report EPA 600-8-78-017. U.S. Environmental Protection Agency. Boxman, I., Dijkman, R., Verhoef, L., Maat, A., van Dijk, G., Vennema, H. and Koopmans, M. (2009) Norovirus on swabs taken from hands illustrate route of transmission: a case study. J. Food Protect. 72, 1753–5. Boxman, I. L., Dijkman, R., te Loeke, N. A., Hagele, G., Tilburg, J. J., Vennema, H. and Koopmans, M. (2009)Environmental swabs as a tool in norovirus outbreak investigation, including outbreaks on cruise ships. J. Food Protect. 72, 111–9. Brady, M. T., Evans, J. and Cuartas, J. (1990) Survival and disinfection of parainfluenza viruses on environmental surfaces. Am. J. Infect. Control 18, 18–23. Brankston, G., Gitterman, L., Hirji, Z., Lemieux, C. and Gardam, M. (2007) Transmission of influenza A in human beings. Lancet Infect. Dis. 7, 257–65. Bresee, J. S., Widdowson, M. A., Monroe, S. S. and Glass, R. I. (2002) Foodborne viral gastroenteritis: challenges and opportunities. Clin. Infect. Dis. 35, 748–753. BIBLIOGRAPHY 163 Bright, K. R., Boone, S. A. and Gerba, C. P. (2009)Occurrence of bacteria and viruses on elementary classroom surfaces and the potential role of classroom hygiene in the spread of infectious diseases. J. Sch. Nurs. 26, 33–41. Bures, S., Fishbain, J. T., Uyehara, C. F. T., Parker, J. M. and Berg, B. W. (2000) Computer keyboards and faucet handles as reservoirs of nosocomial pathogens in the intensive care unit. Am. J. Infect. Control 28, 465–471. Butz, A. M., Fosarelli, P., Dick, J., Cusack, T. and Yolken, R. (1993) Prevalence of rotavirus on high-risk fomites in day-care facilities. Pediatrics 92, 202–5. Carducci, A., Verani, M., Casini, B., Giuntini, A., Mazzoni, F., Rovini, E., Passaglia, A., Giusti, L., Valenza, A. and Lombardi, R. (2002)Detection and potential indicators of the presence of hepatitis C virus on surfaces in hospital settings. Lett. Appl. Microbiol. 34, 189–93. Casewell, M. W. and Desai, N. (1983) Survival of multiply-resistant Klebsiella aerogenes and other Gram-negative bacilli on finger-tips. J. Hosp. Infect. 4, 350–360. CDC (2008) Norovirus Outbreak in an Elementary School—District of Columbia, February 2007. Morb. Mortal. Wkly. Rep. 56, 1341–1343. Cheesbrough, J. S., Barkess-Jones, L. and Brown, D. W. (1997) Possible prolonged environmental survival of small round structured viruses. J. Hosp. Infect. 35, 325– 326. Cheesbrough, J. S., Green, J., Gallimore, C. I., Wright, P. A. and Brown, D. W. (2000) Widespread environmental contamination with Norwalk-like viruses (NLV) detected in a prolonged hotel outbreak of gastroenteritis. Epidemiol. Infect. 125, 93–8. Chen, Y. C., Huang, L. M., Chan, C. C., Su, C. P., Chang, S. C., Chang, Y. Y., Chen, M. L., Hung, C. C., Chen, W. J., Lin, F. Y. and Lee, Y. T. (2004) SARS in hospital emergency room. Emerg. Infect. Dis. 10, 782–8. BIBLIOGRAPHY 164 Chen, Y., Jackson, K. M., Chea, F. P. and Schaffner, D. W. (2001)Quantification and variability analysis of bacterial cross-contamination rates in common food service tasks. J. Food Protect. 64, 72–80. Clay, S., Maherchandani, S., Malik, Y. S. and Goyal, S. M. (2006)Survival on uncommon fomites of feline calicivirus, a surrogate of noroviruses. Am. J. Infect. Control 34, 41–43. Clendening, L. (1960) Source Book of Medical History. Dover Publications, Inc.. Mineola, NY. Cohen Hubal, E. A., Nishioka, M. G., Ivancic, W. A., Morara, M. and Egeghy, P. P. (2008) Comparing surface residue transfer efficiencies to hands using polar and nonpolar fluorescent tracers. Environ. Sci. Technol. 42, 934–9. Colford Jr, J. M., Rees, J. R., Wade, T. J., Khalakdina, A., Hilton, J. F., Ergas, I. J., Burns, S., Benker, A., Ma, C. and Bowen, C. (2002) Participant blinding and gastrointestinal illness in a randomized, controlled trial of an in-home drinking water intervention. Emerg. Infect. Dis. 8, 29. Covello, V. and Merkhofer, M. (1993) Risk Assessment Methods: Approaches for Assessing Health and Environmental Risks. Plenum Press. New York, NY. Craun, M. F., Craun, G. F., Calderon, R. L. and Beach, M. J. (2006) Waterborne outbreaks reported in the United States. J. Water Health 4, 19–30. Crossley, K. B. and Ross, J. (1985)Colonization of hospitalized patients by Staphylococcus aureus, Staphylococcus epidermidis and enterococci. J. Hosp. Infect. 6, 179– 186. Curtis, V., Cairncross, S. and Yonli, R. (2000) Review: Domestic hygiene and diarrhoea- pinpointing the problem. Trop. Med. Int. Health 5, 22–32. Dagan, R., Leventhal, A., Anis, E., Slater, P., Ashur, Y. and Shouval, D. (2005) Incidence of hepatitis A in Israel following universal immunization of toddlers. JAMA - J. Am. Med. Assoc. 294, 202. BIBLIOGRAPHY 165 Dahl, I. L., Grufman, M., Hellberg, C. and Krabbe, M. (1991)Absenteeism because of illness at daycare centers and in three family systems. Acta Paediatr. 80, 436–445. Das, I., Lambert, P., Hill, D., Noy, M., Bion, J. and Elliott, T. (2002) Carbapenemresistant Acinetobacter and role of curtains in an outbreak in intensive care units. J. Hosp. Infect. 50, 110–114. Dawson, D. J., Paish, A., Staffell, L. M., Seymour, I. J. and Appleton, H. (2005) Survival of viruses on fresh produce, using MS2 as a surrogate for norovirus. J. Appl. Microbiol. 98, 203–209. Desenclos, J. C. and MacLafferty, L. (1993) Community wide outbreak of hepatitis A linked to children in day care centres and with increased transmission in young adult men in Florida 1988-9. J. Epidemiol. Commun. H. 47, 269. Dick, E. C., Jennings, L. C., Mink, K. A., Wartgow, C. D. and Inhorn, S. L. (1987) Aerosol transmission of rhinovirus colds. J. Infect. Dis. 156, 442–448. Donowitz, L. G. (1999) Infection Control in the Child Care Center and Preschool. Lippincott Williams and Wilkins. Dowd, S. E., Pillai, S. D., Wang, S. and Corapcioglu, M. Y. (1998) Delineating the specific influence of virus isoelectric point and size on virus adsorption and transport through sandy soils. Appl. Environ. Microbiol. 64, 405–10. Dowell, S. F., Simmerman, J. M., Erdman, D. D., Wu, J. S., Chaovavanich, A., Javadi, M., Yang, J. Y., Anderson, L. J., Tong, S. and Ho, M. S. (2004)Severe acute respiratory syndrome coronavirus on hospital surfaces. Clin. Infect. Dis. 39, 652–7. Duizer, E., Bijkerk, P., Rockx, B., De Groot, A., Twisk, F. and Koopmans, M. (2004) Inactivation of caliciviruses. Appl. Environ. Microbiol. 70, 4538. Ekanem, E. E., Dupont, H. L., Pickering, L. K., Selwyn, B. J. and Hawkins, C. M. (1983) Transmission dynamics of enteric bacteria in day-care-centers. Am. J. Epidemiol. 118, 562–572. BIBLIOGRAPHY 166 Elkhyat, A., Agache, P., Zahouani, H. and Humbert, P. (2001) A new method to measure in vivo human skin hydrophobia. Int. J. Cosmetic Sci. 23, 347–352. Ellner, P. D. and Ellner, C. J. (1966) Survival of bacteria on swabs. J. Bacteriol. 91, 905. Elveback, L., Ackerman, E., Gatewood, L. and Fox, J. P. (1971) Stochastic twoagent epidemic simulation models for a community of families. Am. J. Epidemiol. 93, 267–280. Evans, M. R., Meldrum, R., Lane, W., Gardner, D., Ribeiro, C. D., Gallimore, C. I. and Westmoreland, D. (2002) An outbreak of viral gastroenteritis following environmental contamination at a concert hall. Epidemiol. Infect. 129, 355–60. Fenn, E. A. (2000) Biological warfare in eighteenth-century North America: beyond Jeffery Amherst. J Am. Hist. 86, 1552–1580. Ferenczy, A., Bergeron, C. and Richart, R. M. (1989) Human papillomavirus DNA in fomites on objects used for the management of patients with genital human papillomavirus infections. Obstet. Gynecol. 74, 950–4. Ferguson, A. C. (2003)Estimating Percutaneous Absorption Using Micro-level Activity Time Series (MLATS). Ph.d. thesis. Stanford University. Ferguson, A. C., Biddle, D., Coleman, S., Bursac, Z. and Johnson, W. (2009)In-vitro soil adherence for dermal exposure using a controlled mechanical chamber. J. Appl. Sci. Res. 5, 232–243. Ferguson, A. C., Canales, R. A., Beamer, P., AuYeung, W., Key, M., Munninghoff, A., Lee, K. T. W., Robertson, A. and Leckie, J. O. (2006) Video methods in the quantification of children’s exposures. J. Expo. Sci. Env. Epid. 16, 287–298. Fewtrell, L., Kaufmann, R. B., Kay, D., Enanoria, W., Haller, L. and Colford, J. M. (2005) Water, sanitation, and hygiene interventions to reduce diarrhoea in less developed countries: a systematic review and meta-analysis. Lancet Infect. Dis. 5, 42–52. BIBLIOGRAPHY 167 Fischer, B., Powis, J., Firestone Cruz, M., Rudzinski, K. and Rehm, J. (2008) Hepatitis C virus transmission among oral crack users: viral detection on crack paraphernalia. Eur. J. Gastroenterol. Hepatol. 20, 29–32. Fleming, D. W., Cochi, S. L., Hightower, A. W. and Broome, C. V. (1987)Childhood upper respiratory tract infections: to what degree is incidence affected by day-care attendance?. Pediatrics 79, 55. Frank, D. N., Feazel, L. M., Bessesen, M. T., Price, C. S., Janoff, E. N. and Pace, N. R. (2010)The human nasal microbiota and Staphylococcus aureus carriage. PLOS One 5, e10598. Friedman, C. and Petersen, K. H. (2004)Infection Control in Ambulatory Care. Jones and Bartlett Learning. Froio, N., Nicastri, E., Comandini, U. V., Cherubini, C., Felicioni, R., Solmone, M., Di Giulio, S. and Petrosillo, N. (2003)Contamination by hepatitis B and C viruses in the dialysis setting. Am. J. Kidney. Dis. 42, 546–50. Fuhs, G. W., Chen, M., Sturman, L. S. and Moore, R. S. (1985) Virus adsorption to mineral surfaces is reduced by microbial overgrowth and organic coatings. Microb. Ecol. 11, 25–39. Galasso, G. J. and Sharp, D. G. (1962) Virus particle aggregation and the plaqueforming unit. J. Immunol. 88, 339. Gallimore, C. I., Taylor, C., Gennery, A. R., Cant, A. J., Galloway, A., IturrizaGomara, M. and Gray, J. J. (2006) Environmental monitoring for gastroenteric viruses in a pediatric primary immunodeficiency unit. J. Clin. Microbiol. 44, 395– 9. Gallimore, C. I., Taylor, C., Gennery, A. R., Cant, A. J., Galloway, A., Lewis, D. and Gray, J. J. (2005) Use of a heminested reverse transcriptase PCR assay for detection of astrovirus in environmental swabs from an outbreak of gastroenteritis in a pediatric primary immunodeficiency unit. J. Clin. Microbiol. 43, 3890–4. BIBLIOGRAPHY 168 Gallimore, C. I., Taylor, C., Gennery, A. R., Cant, A. J., Galloway, A., Xerry, J., Adigwe, J. and Gray, J. J. (2008) Contamination of the hospital environment with gastroenteric viruses: comparison of two pediatric wards over a winter season. J. Clin. Microbiol. 46, 3112–5. Garrett, V., Bornschlegel, K., Lange, D., Reddy, V., Kornstein, L., Kornblum, J., Agasan, A., Hoekstra, M., Layton, M. and Sobel, J. (2006)A recurring outbreak of Shigella sonnei among traditionally observant Jewish children in New York City: the risks of daycare and household transmission. Epidemiol. Infect. 134, 1231–1236. Gerba, C. P. (1984) Applied and theoretical aspects of virus adsorption to surfaces. Adv. Appl. Microbiol. 30, 133–68. Gerba, C. P. and Haas, C. N. (1988) Assessment of risks associated with enteric viruses in contaminated drinking water. Chemical and Biological Characterization of Municipal Sludges, Sediments, Dredge Spoils, and Drilling Muds. ASTM International. Gfatter, R., Hackl, P. and Braun, F. (1997) Effects of soap and detergents on skin surface pH, stratum corneum hydration and fat content in infants. Dermatology 195, 258–62. Gibbons, J. D. (1985) Nonparametric Statistical Inference (Statistics: Textbooks & Monographs). Marcel Dekker, Inc.. New York, NY. Gibson, L. L., Rose, J. B. and Haas, C. N. (1999) Use of quantitative microbial risk assessment for evaluation of the benefits of laundry sanitation. Am. J. Infect. Control 27, S34–S39. Gibson, L. L., Rose, J. B., Haas, C. N., Gerba, C. P. and Rusin, P. A. (2002)Quantitative assessment of risk reduction from hand washing with antibacterial soaps. J. Appl. Microbiol. 92, 136–143. Giraldo, J. O. and Palacio, D. H. (2008) Deterministic SIR (SusceptibleInfectedRemoved) models applied to varicella outbreaks. Epidemiol. Infect. 136, 679–687. BIBLIOGRAPHY 169 Girard, M., Ngazoa, S., Mattison, K. and Jean, J. (2010) Attachment of noroviruses to stainless steel and their inactivation using household disinfectants. J. Food Prot. 73, 400–404. Girou, E., Chevaliez, S., Challine, D., Thiessart, M., Morice, Y., Lesprit, P., TkoubScheirlinck, L., Soing-Altrach, S., Cizeau, F., Cavin, C., Andre, M., Dahmanne, D., Lang, P. and Pawlotsky, J. M. (2008) Determinant roles of environmental contamination and noncompliance with standard precautions in the risk of hepatitis C virus transmission in a hemodialysis unit. Clin. Infect. Dis. 47, 627–33. Glass, R. J., Glass, L. M., Beyeler, W. E. and Min, H. J. (2006) Targeted social distancing design for pandemic influenza. Emerg. Infect. Dis. 12, 1671–1681. Goldhammer, K. A., Dooley, D. P., Ayala, E., Zera, W. and Hill, B. L. (2006)Prospective study of bacterial and viral contamination of exercise equipment. Clin. J. Sport. Med. 16, 34–8. Goldmann, D. A. (2000) Transmission of viral respiratory infections in the home. Pediatr. Infect. Dis. J. 19, S97. Graves, N., Halton, K., Doidge, S., Clements, A., Lairson, D. and Whitby, M. (2008) Who bears the cost of healthcare-acquired surgical site infection?. J. Hosp. Infect. 69, 274–282. Green, J., Wright, P. A., Gallimore, C. I., Mitchell, O., Morgan-Capner, P. and Brown, D. W. (1998) The role of environmental contamination with small round structured viruses in a hospital outbreak investigated by reverse-transcriptase polymerase chain reaction assay. J. Hosp. Infect. 39, 39–45. Gurley, E. S., Montgomery, J. M., Hossain, M. J., Bell, M., Azad, A. K., Islam, M. R., Molla, M. A., Carroll, D. S., Ksiazek, T. G., Rota, P. A., Lowe, L., Comer, J. A., Rollin, P., Czub, M., Grolla, A., Feldmann, H., Luby, S. P., Woodward, J. L. and Breiman, R. F. (2007) Person-to-person transmission of Nipah virus in a Bangladeshi community. Emerg. Infect. Dis. 13, 1031–7. BIBLIOGRAPHY 170 Gwaltney, J. M. (1982)Transmission of experimental rhinovirus infection by contaminated surfaces. Am. J. Epidemiol. 116, 828–833. Gwaltney, J. M., J., Moskalski, P. B. and Hendley, J. O. (1978) Hand-to-hand transmission of rhinovirus colds. Ann. Intern. Med. 88, 463–467. Haas, C. N. (1983)Estimation of risk due to low doses of microorganisms: a comparison of alternative methodologies. Am. J. Epidemiol. 118, 573. Haas, C. N., Marie, J. R., Rose, J. B. and Gerba, C. P. (2005)Assessment of benefits from use of antimicrobial hand products: Reduction in risk from handling ground beef. Int. J. Hyg. Envir. Heal. 208, 461–466. Haas, C., Rose, J. and Gerba, C. (1999) Quantitative Microbial Risk Assessment.. John Wiley and Sons, Inc.. New York, NY. Hadler, S. C., Webster, H. M., Erben, J. J., Swanson, J. E. and Maynard, J. E. (1980) Hepatitis A in day-care centers. A community-wide assessment. New England J. Med. 302, 1222–1227. Hahn, S., Kim, Y. J. and Garner, P. (2001) Reduced osmolarity oral rehydration solution for treating dehydration due to diarrhoea in children: systematic review. Brit. Med. J. 323, 81. Hall, C. (1997) External pressure at the hand during object handling and work with tools. Int. J. Ind. Ergonom. 20, 191–206. Hall, C. B. (1983) The nosocomial spread of respiratory syncytial viral infections. Annu. Rev. Med. 34, 311–319. Hall, C. B. and Douglas Jr, R. G. (1981)Modes of transmission of respiratory syncytial virus. J. Pediatr. 99, 100. Hall, C. B., Douglas Jr, R. G., Schnabel, K. C. and Geiman, J. M. (1981)Infectivity of respiratory syncytial virus by various routes of inoculation. Infect. Immun. 33, 779. BIBLIOGRAPHY 171 Hall, C. B., Douglas, R. G. and Geiman, J. M. (1980)Possible transmission by fomites of respiratory syncytial virus. J. Infect. Dis. 141, 98–102. Hamada, N., Gotoh, K., Hara, K., Iwahashi, J., Imamura, Y., Nakamura, S., Taguchi, C., Sugita, M., Yamakawa, R., Etoh, Y., Sera, N., Ishibashi, T., Chijiwa, K. and Watanabe, H. (2008)Nosocomial outbreak of epidemic keratoconjunctivitis accompanying environmental contamination with adenoviruses. J. Hosp. Infect. 68, 262– 8. Heijne, J. C. M., Teunis, P., Morroy, G., Wijkmans, C., Oostveen, S., Duizer, E., Kretzschmar, M. and Wallinga, J. (2009)Enhanced hygiene measures and norovirus transmission during an outbreak. Emerg. Infect. Dis. 15, 24. Heikkinen, T. and Jrvinen, A. (2003) The common cold. Lancet 361, 51–59. Hendley, J. O., Wenzel, R. P. and Gwaltney Jr, J. M. (1973)Transmission of rhinovirus colds by self-inoculation. New England J. Med. 288, 1361–1364. Herath, G., Yamamoto, K. and Urase, T. (1999)Removal of viruses by microfiltration membranes at different solution environments. Water Sci. Technol. 40, 331–338. Herzog, A. B., McLennan, S. D., Pandey, A. K., Gerba, C. P., Haas, C. N., Rose, J. B. and Hashsham, S. A. (2009) Implications of Limits of Detection of Various Methods for Bacillus anthracis in Computing Risks to Human Health. Appl. Environ. Microbiol. 75, 6331. Heudorf, U., Mersch-Sundermann, V. and Angerer, J. (2007) Phthalates: toxicology and exposure. Int. J. Hyg. Envir. Heal. 210, 623–634. Holaday, B., Pantell, R., Lewis, C. and Gilliss, C. L. (1990)Patterns of fecal-coliform contamination in day-care-centers. Public Health Nurs. 7, 224–228. Hota, B. (2004) Contamination, disinfection, and cross-colonization: are hospital surfaces reservoirs for nosocomial infection?. Clin. Infect. Dis. 39, 1182–1189. BIBLIOGRAPHY 172 Hurst, C. (1996) Modeling Disease Transmission and Its Prevention by Disinfection. Cambridge University Press. Great Britian. Iizuka, S. (2006) An outbreak of group C rotavirus infection in an elementary school in Shimane Prefecture, Japan, February 2006. Jpn. J. Infect. Dis. 59, 350. ILSI (1996) A conceptual framework to assess the risks of human disease following exposure to pathogens. Risk Anal. 16, 841–848. ILSI (1999)Revised Framework for Microbial Risk Assessment. An ILSI Risk Science Institute Workshop Report pp. 1–29. Iwamoto, M., Jernigan, D. B., Guasch, A., Trepka, M. J., Blackmore, C. G., Hellinger, W. C., Pham, S. M., Zaki, S., Lanciotti, R. S. and Lance-Parker, S. E. (2003) Transmission of West Nile virus from an organ donor to four transplant recipients. New England J. Med. 348, 2196. James, C. T. and David, J. W. (2001) Epidemiologic Methods for the Study of Infectious Diseases. Oxford University Press. New York, NY. Jaykus, L. A., De Leon, R. and Sobsey, M. D. (1996)A virion concentration method for detection of human enteric viruses in oysters by PCR and oligoprobe hybridization. Appl. Environ. Microbiol. 62, 2074. Jefferson, T., Del Mar, C., Dooley, L., Ferroni, E., Al-Ansary, L. A., Bawazeer, G. A., van Driel, M. L., Foxlee, R. and Rivetti, A. (2009) Physical interventions to interrupt or reduce the spread of respiratory viruses: systematic review. Brit. Med. J. 339, b3675. Jennings, L. C. and Dick, E. C. (1987) Transmission and control of rhinovirus colds. Eur. J. Epidemiol. 3, 327–35. Jennings, L. C., Dick, E. C., Mink, K. A., Wartgow, C. D. and Inhorn, S. L. (1988) Near disappearance of rhinovirus along a fomite transmission chain. J. Infect. Dis. pp. 888–892. BIBLIOGRAPHY 173 Jiang, X., Dai, X., Goldblatt, S., Buescher, C., Cusack, T. M., Matson, D. O. and Pickering, L. K. (1998) Pathogen transmission in child care settings studied by using a cauliflower virus DNA as a surrogate marker. J. Infect. Dis. 177, 881–888. Jones, E. L., Kramer, A., Gaither, M. and Gerba, C. P. (2007) Role of fomite contamination during an outbreak of norovirus on houseboats. Int. J. Environ. Health Res. 17, 123–31. Jones, G., Steketee, R. W., Black, R. E., Bhutta, Z. A. and Morris, S. S. (2003) How many child deaths can we prevent this year?. Lancet 362, 65–71. Julian, T. R., Canales, R. A., Leckie, J. O. and Boehm, A. B. (2009) A model of exposure to rotavirus from nondietary ingestion iterated by simulated intermittent contacts. Risk Anal. 29, 617–32. Julian, T. R., Leckie, J. O. and Boehm, A. B. (2010)Virus transfer between fingerpads and fomites. J. Appl. Microbiol. 109, 1868–1874. Kaltenthaler, E. C., Elsworth, A. M., Schweiger, M. S., Mara, D. D. and Brauxholtz, D. A. (1995) Faecal contamination on children’s hands and environmental surfaces in primary schools in Leeds. Epidemiol. Infect. 115, 527–534. Kapikian, A. Z., Wyatt, R. G., Dolin, R., Thornhill, T. S., Kalica, A. R. and Chanock, R. M. (1972) Visualization by immune electron microscopy of a 27-nm particle associated with acute infectious nonbacterial gastroenteritis. J. Virol. 10, 1075. Katayama, H., Shimasaki, A. and Ohgaki, S. (2002) Development of a virus concentration method and its application to detection of enterovirus and Norwalk virus from coastal seawater. Appl. Environ. Microbiol. 68, 1033. Kawahara, A. and Yoshida, M. (2009) Detection of viral DNA in nonlesional skin of patients with molluscum contagiosum and on environmental fomites. Br. J. Dermatol. 160, 1357–9. BIBLIOGRAPHY 174 Keswick, B. H., Pickering, L. K., DuPont, H. and Woodward, W. E. (1983) Survival and detection of rotaviruses on environmental surfaces in day-care-centers. Appl. Environ. Microbiol. 46, 813–816. Khan, A. J., Luby, S. P., Fikree, F., Karim, A., Obaid, S., Dellawala, S., Mirza, S., Malik, T., Fisher-Hoch, S. and McCormick, J. B. (2000) Unsafe injections and the transmission of hepatitis B and C in a periurban community in Pakistan. Bull. World Health Organ. 78, 956–963. Klevens, R. M., Edwards, J. R., Richards, C. L., Horan, T. C., Gaynes, R. P., Pollock, D. A. and Cardo, D. M. (2007) Estimating health care-associated infections and deaths in US hospitals, 2002. Public Health Rep. 122, 160. Kojima, S., Kageyama, T., Fukushi, S., Hoshino, F. B., Shinohara, M., Uchida, K., Natori, K., Takeda, N. and Katayama, K. (2002) Genogroup-specific PCR primers for detection of Norwalk-like viruses. J. Virol. Methods 100, 107–114. Kownatzki, E. (2003) Hand hygiene and skin health. J. Hosp. Infect. 55, 239–245. Krilov, L. R., Barone, S. R., Mandel, F. S., Cusack, T. M., Gaber, D. J. and Rubino, J. R. (1996) Impact of an infection control program in a specialized preschool* 1. Am. J. Infect. Control 24, 167–173. Kuusi, M., Nuorti, J. P., Maunula, L., Minh Tran, N. N., Ratia, M., Karlsson, J. and von Bonsdorff, C. H. (2002) A prolonged outbreak of Norwalk-like calicivirus (NLV) gastroenteritis in a rehabilitation centre due to environmental contamination. Epidemiol. Infect. 129, 133–8. Kyriacou, A., Drakopoulou, S., Georgaki, I., Fountoulakis, M., Mitsou, E., Lasaridi, K. E., Manios, Y. and Manios, T. (2009) Screening for faecal contamination in primary schools in Crete, Greece. Child Care Hlth Dev. 35, 159–163. Laborde, D. J., Weigle, K. A., Weber, D. J. and Kotch, J. B. (1993) Effect of fecal contamination on diarrheal illness rates in day-care centers. Am. J. Epidemiol. 138, 243. BIBLIOGRAPHY 175 Laborde, D. J., Weigle, K. A., Weber, D. J., Sobsey, M. D. and Kotch, J. B. (1994) The frequency, level, and distribution of fecal contamination in day-care-center classrooms. Pediatrics 94, 1008–1011. Laughlin, L. (2010) Whos minding the kids? Childcare arrangements: spring 2005 / summer 2006. Curr. Popul. Rep. p70-121. Law, A. M. and Kelton, W. D. (1997) Simulation Modeling and Analysis. McGrawHill. New York, NY. Lawniczak, A. T., Fuks, H. and Di Stefano, B. (2006) Individually based SIR models, their motivation, and their spatio-temporal dynamics. Dyn. Contin. Discret. I 13, 437–455. Lazzerini, M. and Ronfani, L. (2008) Oral zinc for treating diarrhoea in children. Cochrane Database Syst. Rev. 3. Lederman, E., Miramontes, R., Openshaw, J., Olson, V. A., Karem, K. L., Marcinak, J., Panares, R., Staggs, W., Allen, D. and Weber, S. G. (2009)Eczema vaccinatum resulting from the transmission of vaccinia virus from a smallpox vaccinee: An investigation of potential fomites in the home environment. Vaccine 27, 375–377. Lemon, S., Sparling, P., Hamburg, M., Relman, D., Choffnes, E. and Mack, A. (2008) Vector-borne Diseases: Understanding the Environmental, Human Health, and Ecological Connections.. The National Academies Press. Washington, D.C. Lennell, A., Khlmann-Berenzon, S., Geli, P., Hedin, K., Petersson, C., Cars, O., Mannerquist, K., Burman, L. G. and Fredlund, H. (2008) Alcohol-based handdisinfection reduced children’s absence from Swedish day care centers. Acta Paediatr. 97, 1672–1680. Lessa, F. C., Gould, P. L., Pascoe, N., Erdman, D. D., Lu, X., Bunning, M. L., Marconi, V. C., Lott, L., Widdowson, M. A. and Anderson, L. J. (2009)Health Care Transmission of a Newly Emergent Adenovirus Serotype in Health Care Personnel at a Military Hospital in Texas, 2007. J. Infect. Dis. 200, 000–000. BIBLIOGRAPHY 176 Li, S., Eisenberg, J. N. S., Spicknall, I. H. and Koopman, J. S. (2009) Dynamics and control of infections transmitted from person to person through the environment. Am. J. Epidemiol. 170, 257. Liao, C., Chen, S. and Chang, C. (2008)Modelling respiratory infection control measure effects. Epidemiol. Infect. 136, 299–308. Liljas, L., Fridborg, K., Valegard, K., Bundule, M. and Pumpens, P. (1994) Crystal structure of bacteriophage fr capsids at 3.5 A resolution. J. Mol. Biol. 244, 279–290. Link, L., Lukens, S. and Bush, M. A. (1995) Spherical grip strength in children 3 to 6 years of age. Am. J. Occup. Ther. 49, 318. Loeb, M., Dafoe, N., Mahony, J., John, M., Sarabia, A., Glavin, V., Webby, R., Smieja, M., Earn, D. J. D. and Chong, S. (2009) Surgical mask vs N95 respirator for preventing influenza among health care workers: a randomized trial. J. Am. Med. Assoc. 302, 1865–1871. Long, S. S., Pickering, L. K. and Prober, C. G. (1997) Principles and Practice of Pediatric Infectious Diseases. first edn. Churchill Livingstone. Philadelphia. Lopez, A. S., Burnett Hartman, A., Nambiar, R., Ritz, L., Owens, P., Loparev, V. N., Guris, D. and Schmid, D. S. (2008)Transmission of a newly characterized strain of varicella zoster virus from a patient with herpes zoster in a long term care facility, West Virginia, 2004. J. Infect. Dis. 197, 646–653. Lowen, A. C., Mubareka, S., Steel, J. and Palese, P. (2007) Influenza virus transmission is dependent on relative humidity and temperature. PLOS Pathog. 3, 1470– 1476. Lu, N., Samuels, M. E., Shi, L., Baker, S. L., Glover, S. H. and Sanders, J. M. (2004) Child day care risks of common infectious diseases revisited. Child Care Hlth Dev. 30, 361–368. BIBLIOGRAPHY 177 Luijt, D. S., Schirm, J., Savelkoul, P. H. M. and Hoekstra, A. (2001)Risk of infection by reprocessed and resterilized virus-contaminated catheters; an in-vitro study. Eur. Heart J. 22, 378. Lyman, W. H., Walsh, J. F., Kotch, J. B., David, J. W., Elizabeth, G. and Vinjè, J. (2009)Prospective study of etiologic agents of acute gastroenteritis outbreaks in child care centers. J. Pediatr. 154, 253–257. Malek, M. A., Curns, A. T., Holman, R. C., Fischer, T. K., Bresee, J. S., Glass, R. I., Steiner, C. A. and Parashar, U. D. (2006)Diarrhea- and rotavirus-associated hospitalizations among children less than 5 years of age: United States, 1997 and 2000. Pediatrics 117, 1887–1892. Mangili, A. and Gendreau, M. A. (2005) Transmission of infectious diseases during commercial air travel. Lancet 365, 989–996. Mathers, C. D., Boerma, T. and Ma Fat, D. (2008) The Global Burden of Disease: 2004 Update. World Health Organization. Geneva, Switzerland. Mbithi, J. N., Springthorpe, V. S., Boulet, J. R. and Sattar, S. A. (1992) Survival of hepatitis-A virus on human hands and its transfer on contact with animate and inanimate surfaces. J. Clin. Microbiol. 30, 757–763. McCluskey, R., Sandin, R. and Greene, J. (1996) Detection of airborne cytomegalovirus in hospital rooms of immuncompromised patients. J. Virol. Methods 56, 115–118. Michen, B. and Graule, T. (2010) Isoelectric points of viruses. J. Appl. Microbiol. 109, 388–397. Monto, A. S. (2002)Epidemiology of viral respiratory infections. Am. J. Med. 112, 4– 12. Montville, R. and Schaffner, D. W. (2003) Inoculum size influences bacterial cross contamination between surfaces. Appl. Environ. Microbiol. 69, 7188–7193. BIBLIOGRAPHY 178 Morens, D. M. and Rash, V. M. (1995) Lessons from a nursing home outbreak of influenza A. Infect. Cont. Hosp. Ep. 16, 275–280. Mubareka, S., Lowen, A. C., Steel, J., Coates, A. L., Garcia-Sastre, A. and Palese, P. (2009)Transmission of influenza virus via aerosols and fomites in the guinea pig model. J. Infect. Dis. 199, 858–865. Murray, B. E. (1990) The life and times of the Enterococcus. Clin. Microbiol. Rev. 3, 46. Musa, E. K., Desai, N. and Casewell, M. W. (1990) The survival of Acinetobacter calcoaceticus inoculated on fingertips and on formica. J. Hosp. Infect. 15, 219–27. Nasell, I. (2002)Stochastic models of some endemic infections. Math. Biosci. 179, 1– 19. Nazaroff, W. W., Nicas, M. and Miller, S. L. (1998) Framework for evaluating measures to control nosocomial tuberculosis transmission. Indoor Air 8, 205–218. Neely, A. N. and Sittig, D. F. (2002) Basic microbiologic and infection control information to reduce the potential transmission of pathogens to patients via computer hardware. Brit. Med. J. 9, 500. Nicas, M. and Best, D. (2008) A study quantifying the hand-to-face contact rate and its potential application to predicting respiratory tract infection. J. Occup. Environ. Hyg. 5, 347–52. Nicas, M. and Jones, R. M. (2009) Relative contributions of four exposure pathways to influenza infection risk. Risk Anal. 29, 1292–303. Nicas, M. and Sun, G. (2006) An integrated model of infection risk in a health-care environment. Risk Anal. 26, 1085–1096. NRC (1983) Risk Assessment in the Federal Government: Managing the Process. National Resource Council, National Academy Press. Washington, D.C. BIBLIOGRAPHY 179 O’Connell, K. P., Bucher, J. R., Anderson, P. E., Cao, C. J., Khan, A. S., Gostomski, M. V. and Valdes, J. J. (2006) Real-time fluorogenic reverse transcription-PCR assays for detection of bacteriophage MS2. Appl. Environ. Microbiol. 72, 478. Osterblad, M., Jarvinen, H., Lonnqvist, K., Huikko, S., Laippala, P., Viljanto, J., Arvilommi, H. and Huovinen, P. (2003)Evaluation of a new cellulose sponge-tipped swab for microbiological sampling: a laboratory and clinical investigation. J. Clin. Microbiol. 41, 1894. Osterholm, M. T., Max, B. J., Hanson, M. and Polesky, H. F. (1979)Potential risk of salivary-mediated viral hepatitis type B transmission from oral exposure to fomites. J. Hyg. 83, 487–90. Otake, S., Dee, S. A., Rossow, K. D., Deen, J., Joo, H. S., Molitor, T. W. and Pijoan, C. (2002) Transmission of porcine reproductive and respiratory syndrome virus by fomites (boots and coveralls). J. Swine Health Prod. 10, 59–66. Pẽrez Rodrı́guez, F., Valero, A., Todd, E. C. D., Carrasco, E., Garcı́a-Gimeno, R. M. and Zurera, G. (2007)Modeling transfer of Escherichia coli O157: H7 and Staphylococcus aureus during slicing of a cooked meat product. Meat Sci 76, 692–699. Pappas, D. E., Hendley, J. O. and Schwartz, R. H. (2009) Respiratory Viral RNA on Toys in Pediatric Office Waiting Rooms. Pediatr. Infect. Dis. J. 29, 1–3. Payment, P., Richardson, L., Siemiatycki, J., Dewar, R., Edwardes, M. and Franco, E. (1991) A randomized trial to evaluate the risk of gastrointestinal disease due to consumption of drinking water meeting current microbiological standards. Am. J. Public Health 81, 703. Pecson, B. M., Martin, L. V. and Kohn, T. (2009)Quantitative PCR for determining the infectivity of bacteriophage MS2 upon inactivation by heat, UVB and singlet oxygen: advantages and limitations of an enzymatic treatment to reduce falsepositive results. Appl. Environ. Microbiol. 75, 5544–5554. BIBLIOGRAPHY 180 Piazza, M., Valletta, G., Picciotto, L., Guadagnino, V., Borgia, G., Nappa, S., Battagliese, G., Matarasso, S., Sammartino, G., Laino, A. and et al. (1987) HBsAg (hepatitis B virus surface antigen) contamination in the dental environment. Minerva Stomatol. 36, 903–7. Pickering, A. J., Davis, J., Walters, S. P., Horak, H. M., Keymer, D. P., Mushi, D., Strickfaden, B., Chynoweth, J. S., Liu, J., Blum, A., Rogers, K. and Boehm, A. B. (2010)Hands, Water, and Health: Fecal Contamination in Tanzanian Communities with Improved, Non-Networked Water Supplies. Environ. Sci. Technol. 44, 3267– 3272. Pickering, A. J., Julian, T., Mamuya, S., Boehm, A. B. and Davis, J. (2011)Hand fecal contamination among Tanzanian mothers varies temporally and following household activities. Trop. Med. Int. Health In Press. Pinfold, J. V. (1990) Faecal contamination of water and fingertip-rinses as a method for evaluating the effect of low-cost water supply and sanitation activities on faecooral disease transmission. I. A case study in rural north-east Thailand. Epidemiol. Infect. 105, 363–375. Pirtle, E. C. and Beran, G. W. (1996)Stability of porcine reproductive and respiratory syndrome virus in the presence of fomites commonly found on farms. Journal of the American Veterinary Medical Association 208, 390. Plunket, J. D. (1879) Cotton as a Fomite. Public Health Pap. Rep. 5, 84. Pollock, M. R. (1947)The growth of /emphH. pertussis on media without blood.. Br. J. Exp. Pathol. 28, 295–307. Rahmandad, H. and Sterman, J. (2008) Heterogeneity and network structure in the dynamics of diffusion: Comparing agent-based and differential equation models. Manage. Sci. 54, 998–1014. Ravenel, M. P. (1931)Hieronymus Fracastorius. Contagion, Contagious Diseases, and Their Treatment. Am. J. Public Health 21, 106. BIBLIOGRAPHY 181 Reddy, C. A., Beveridge, T. J., Breznak, J. A., Snyder, L., Schmidt, T. M. and Marzluf, G. A. (2006) Methods for General and Molecular Microbiology. 3rd, ed edn. American Society of Microbiology Press. Washington, D.C. Reed, S. E. (1975) An investigation of the possible transmission of Rhinovirus colds through indirect contact. J. Hyg. 75, 249–258. Regli, S., Rose, J. B., Haas, C. N. and Gerba, C. P. (1991) Modeling the risk from Giardia and viruses in drinking water. J. Am. Water Works Ass. 83, 76–84. Roberts, L., Jorm, L., Patel, M., Smith, W., Douglas, R. M. and McGilchrist, C. (2000) Effect of infection control measures on the frequency of diarrheal episodes in child care: a randomized, controlled trial. Pediatrics 105, 743. Roberts, L., Smith, W., Jorm, L., Patel, M., Douglas, R. M. and McGilchrist, C. (2000) Effect of infection control measures on the frequency of upper respiratory infection in child care: a randomized, controlled trial. Pediatrics 105, 738. Rogers, M., Weinstock, D. M., Eagan, J., Kiehn, T., Armstrong, D. and Sepkowitz, K. A. (2000)Rotavirus outbreak on a pediatric oncology floor: possible association with toys. Am. J. Infect. Control 28, 378–380. Rose, J. B., Haas, C. N. and Regli, S. (1991)Risk assessment and control of waterborne giardiasis. Am. J. Public Health 81, 709. Rotbart, H. A. (1995) Human Enterovirus Infections. American Society for Microbiology. Washington, D.C. Roth, D. E., Richard, S. A. and Black, R. E. (2010) Zinc supplementation for the prevention of acute lower respiratory infection in children in developing countries: meta-analysis and meta-regression of randomized trials. Int. J. Epidemiol. 39, 795– 808. Rudnik, S., McDevitt, J., First, M. and Spengler, J. (2009) Inactivating influenza viruses on surfaces using hydrogen peroxide or triethylene glycol at low vapor concentrations. Am. J. Infect. Control 37, 813–819. BIBLIOGRAPHY 182 Runner, J. C. (2007) Bacterial and viral contamination of reusable sharps containers in a community hospital setting. Am. J. Infect. Control 35, 527–30. Rusin, P., Maxwell, S. and Gerba, C. (2002) Comparative surface-to-hand and fingertip-to-mouth transfer efficiency of Gram-positive bacteria, Gram-negative bacteria, and phage. J. Appl. Microbiol. 93, 585–592. Russell, K. L., Broderick, M. P., Franklin, S. E., Blyn, L. B., Freed, N. E., Moradi, E., Ecker, D. J., Kammerer, P. E., Osuna, M. A., Kajon, A. E., Morn, C. B. and Ryan, M. A. (2006)Transmission dynamics and prospective environmental sampling of adenovirus in a military recruit setting. J. Infect. Dis. 194, 877–85. Sakaguchi, H., Wada, K., Kajioka, J., Watanabe, M., Nakano, R., Hirose, T., Ohta, H. and Aizawa, Y. (2010)Maintenance of influenza virus infectivity on the surfaces of personal protective equipment and clothing used in healthcare settings. Environ. Health Prev. Med. 15. Sandora, T. J., Shih, M. C. and Goldmann, D. A. (2008) Reducing Absenteeism From Gastrointestinal and Respiratory Illness in Elementary School Students: A Randomized, Controlled Trial of an Infection-Control Intervention. Pediatrics 121, e1555–e1567. Scott, E. and Bloomfield, S. F. (1990)The survival and transfer of microbial contamination via cloths, hands and utensils. J. Appl. Microbiol. 68, 271–278. Shields, P. A. and Farrah, S. R. (2002) Characterization of virus adsorption by using DEAE-sepharose and octyl-sepharose. Appl. Environ. Microbiol. 68, 3965. Siegel, J. D., Rhinehart, E. and Jackson, M. (2007) Guideline for Isolation Precautions: Preventing Transmission of Infectious Agents in Healthcare Settings 2007. Centers for Disease Control and Prevention. Atlanta, GA. Siegel, S. (1988) Nonparametric Statistics for the Behavioral Sciences. McGraw-Hill. New York, NY. BIBLIOGRAPHY 183 Soule, H., Genoulaz, O., Gratacap-Cavallier, B., Mallaret, M. R., Morand, P., Francois, P., Luu Duc Bin, D., Charvier, A., Bost-Bru, C. and Seigneurin, J. M. (1999) Monitoring rotavirus environmental contamination in a pediatric unit using polymerase chain reaction. Infect. Cont. Hosp. Ep. 20, 432–4. St Sauver, J., Khurana, M., Kao, A. and Foxman, B. (1998) Hygienic practices and acute respiratory illness in family and group day care homes. Public Health Rep. 113, 544. Stilianakis, N. I. and Drossinos, Y. (2010)Dynamics of infectious disease transmission by inhalable respiratory droplets. J. R. Soc. Interface 7, 1355–1366. Stone, L., Olinky, R. and Huppert, A. (2007) Seasonal dynamics of recurrent epidemics. Nature 446, 533–536. Storey, J. D. (2003) The positive false discovery rate: A Bayesian interpretation and the q-value. Ann. Stat. 31, 2013–2035. Strauss, S., Sastry, P., Sonnex, C., Edwards, S. and Gray, J. (2002) Contamination of environmental surfaces by genital human papillomaviruses. Sex. Transm. Infect. 78, 135–8. Taku, A., Gulati, B. R., Allwood, P. B., Palazzi, K., Hedberg, C. W. and Goyal, S. M. (2002) Concentration and detection of caliciviruses from food contact surfaces. J. Food Protect. 65, 999–1004. Tellier, R. (2009) Aerosol transmission of influenza A virus: a review of new studies. J. R. Soc. Interface 6, S783–S790. Teunis, P. F. M., Moe, C. L. and Liu, P. (2008) Norwalk virus: How infectious is it?. J. Med. Virol. 80, 1468–1476. Teunis, P. F. M., Nagelkerke, N. J. D. and Haas, C. N. (1999) Dose response models for infectious gastroenteritis. Risk Anal. 19, 1251–1260. BIBLIOGRAPHY 184 Thomas, Y., Vogel, G., Wunderli, W., Suter, P., Witschi, M., Koch, D., Tapparel, C. and Kaiser, L. (2008) Survival of influenza virus on banknotes. Appl. Environ. Microbiol. 74, 3002–3007. Thompson, W. (1847)Remarks on the climate of lower egypt, particularly on alexandria, and the plaque that appeared there in 1835, 1840, and 1841: Illustrated by official returns for ten years.. The Medical Times: A Journal 16, 31. Traore, O., Springthorpe, V. S. and Sattar, S. A. (2002) A quantitative study of the survival of two species of Candida on porous and non-porous environmental surfaces and hands. J. Appl. Microbiol. 92, 549–555. Tulve, N. S., Suggs, J. C., McCurdy, T., Hubal, E. A. C. and Moya, J. (2002) Frequency of mouthing behavior in young children. J. Expo. Anal. Env. Epid. 12, 259– 264. USEPA (2001) EPA method 1602: Male-specific (F+) and Somatic Coliphage in Water by Single Agar Layer (SAL) Procedure. Technical Report EPA 821-R-01029. U.S. Environmental Protection Agency. USEPA (2002a)EPA method 1600: Enterococci in water by membrane filtration using membrane-enterococcus indoxyl-β-D-glucoside agar (mEI).. Technical Report EPA 821-R-02-022. U.S. Environmental Protection Agency. USEPA (2002b) EPA method 1603: Escherichia coli (E. coli) in water by membrane filtration using modified membrane-thermotolerant Escherichia coli agar (modified mTEC).. Technical Report EPA 821-R-02-023. U.S. Environmental Protection Agency. USEPA (2006) Child-Specific Exposure Factors Handbook (External Review Draft). Technical Report EPA 600-R-06-096A. U.S. Environmental Protection Agency. van de Vijver, L., Boelsma, E., Bausch-Goldbohm, R. A. and Roza, L. (2003)Subjective skin condition and its association with objective skin measurements. Cosmet. Toiletries 118, 45–54. BIBLIOGRAPHY 185 Van, R., Marrow, A. L., Reves, R. R. and Pickering, L. K. (1991) Environmental contamination in child day-care centers. Am. J. Epidemiol. 133, 460–470. Wald, E. R., Guerra, N. and Byers, C. (1991) Frequency and severity of infections in day care: three-year follow-up. J. Pediatr. 118, 509–514. Ward, R. L., Bernstein, D. I., Young, E. C., Sherwood, J. R., Knowlton, D. R. and Schiff, G. M. (1986) Human rotavirus studies in volunteers: determination of infectious dose and serological response to infection. J. Infect. Dis. 154, 871–880. Weber, T. and Stilianakis, N. (2008) Inactivation of influenza A viruses in the environment and modes of transmission: A critical review. J. Infect. 57, 361–367. Wein, L. M. and Atkinson, M. P. (2009) Assessing infection control measures for pandemic influenza. Risk Anal. 29, 949–62. Weniger, B. G., Ruttenber, A. J., Goodman, R. A., Juranek, D. D., Wahlquist, S. P. and Smith, J. D. (1983) Fecal coliforms on environmental surfaces in two day care centers. Appl. Environ. Microbiol. 45, 733. Widdowson, M. A., van Doornum, G. J., van der Poel, W. H., de Boer, A. S., van de Heide, R., Mahdi, U., Haanen, P., Kool, J. L. and Koopmans, M. (2002) An outbreak of diarrhea in a neonatal medium care unit caused by a novel strain of rotavirus: investigation using both epidemiologic and microbiological methods. Infect. Cont. Hosp. Ep. 23, 665–70. Wilde, J., Van, R., Pickering, L., Eiden, J. and Yolken, R. (1992) Detection of rotaviruses in the day-care environment by reverse-transcriptase polymerase chainreaction. J. Infect. Dis. 166, 507–511. Winther, B., Alper, C. M., Mandel, E. M., Doyle, W. J. and Hendley, J. O. (2007) Temporal relationships between colds, upper respiratory viruses detected by polymerase chain reaction, and otitis media in young children followed through a typical cold season. Pediatrics 119, 1069. BIBLIOGRAPHY 186 Wu, H. M., Fornek, M., Schwab, K. J., Chapin, A. R., Gibson, K., Schwab, E., Spencer, C. and Henning, K. (2005) A norovirus outbreak at a long-term-care facility: the role of environmental surface contamination. Infect. Cont. Hosp. Ep. 26, 802–10. Xue, J., Zartarian, V. G., Oezkaynak, H., Dang, W., Glen, G., Smith, L. and Stallings, C. (2006) A probabilistic arsenic exposure assessment for children who contact chromated copper arsenate(CCA)-treated playsets and decks, part 2: sensitivity and uncertainty analyses. Risk Anal. 26, 533–541. Xue, J., Zartarian, V., Moya, J., Freeman, N., Beamer, P., Black, K., Tulve, N. and Shalat, S. (2007) A meta-analysis of children’s hand-to-mouth frequency data for estimating nondietary ingestion exposure. Risk Anal. 27, 411–420. Yoshikawa, T., Ihira, M., Suzuki, K., Suga, S., Tomitaka, A., Ueda, H. and Asano, Y. (2001) Rapid contamination of the environments with varicella-zoster virus DNA from a patient with herpes zoster. J. Med. Virol. 63, 64–6. Zartarian, V. G., Ozkaynak, H., Burke, J. M., Zufall, M. J., Rigas, M. L. and Furtaw, E. J. (2000) A modeling framework for estimating children’s residential exposure and dose to chlorpyrifos via dermal residue contact and nondietary ingestion. Environ. Health Perspect. 108, 505–514. Zartarian, V. G., Streicker, J., Rivera, A., Cornejo, C. S., Molina, S., Valadez, O. F. and Leckie, J. O. (1995) A pilot-study to collect micro-activity data of 2-year-old to 4-year-old farm-labor children in Salinas Valley, California. J. Expo. Anal. Env. Epid. 5, 21–34. Zuur, A. F., Ieno, E. N., Walker, N. J., Saveliev, A. A. and Smith, G. M. (2009)Mixed Effects Models and Extensions in Ecology with R. Springer Science and Business Media, LLC. New York, NY.