Download Physicians` Acceptable Treatment Failure Rates in Antibiotic

MAJOR ARTICLE
Physicians’ Acceptable Treatment Failure Rates
in Antibiotic Therapy for Coagulase-Negative
Staphylococcal Catheter-Associated Bacteremia:
Implications for Reducing Treatment Duration
Eli N. Perencevich,1,2 Anthony D. Harris,1,2 Keith S. Kaye,3 Douglas D. Bradham,1,2 David N. Fisman,4 Laura A. Liedtke,5
Larry J. Strausbaugh,6 and the Infectious Diseases Society of America Emerging Infections Network
1
Veterans Affairs Maryland Healthcare System and 2Department of Epidemiology and Preventive Medicine, University of Maryland School
of Medicine, Baltimore, Maryland; 3Department of Medicine, Duke University Medical Center, Durham, North Carolina; 4Drexel University School
of Public Health, Philadelphia, Pennsylvania; and 5Research Services and 6Infectious Disease Section, Medical Service, Veterans Affairs
Medical Center, Portland, Oregon
Background. Decreasing the duration of antimicrobial therapy is an attractive strategy for delaying the emergence of antimicrobial resistance. Limited data regarding optimal treatment durations for most clinical infections
hinder the adoption of this approach and impair optimal physician-patient communication under the shared
decision-making model. We aimed to identify acceptable failure rates among infectious disease consultants (IDCs)
for treatment of central venous catheter–associated bacteremia.
Methods. A case scenario involving a representative patient who developed central venous catheter–associated
bacteremia caused by coagulase-negative staphylococci and who received standard-of-care therapy was distributed
to all nonpediatric IDC members of the Infectious Diseases Society of America’s Emerging Infections Network in
August 2003. Each member was suggested 1 of 10 treatment failure rates and asked whether he or she would
accept or reject the given value. Logistic regression was used to evaluate the relationship between specific failure
rates offered to respondents and their willingness to accept them using a methodology derived from contingent
valuation.
Results. Among the 374 respondents (response rate, 54%), the median acceptable failure rate was 6.8%. Thus,
one-half of the IDCs would have found a failure rate of 6.8% to be acceptable. Seventy-five percent of IDCs would
have found a failure rate of 1.6% to be acceptable, and 25% of IDCs would have found a failure rate as high as
11.9% to be acceptable.
Conclusions. The quantified acceptable failure rates, when used to interpret clinical trial or cohort study results,
will help select optimal antimicrobial therapy durations for this specific condition. These findings are a critical
step in the development of effective shared decision-making models.
Concerns about the emergence of vancomycin-resistant
enterococci and staphylococci have prompted recommendations to reduce overall use of these agents. Some
researchers have proposed retarding the emergence of
resistance by decreasing the duration of antimicrobial
Received 25 May 2005; accepted 11 August 2005; electronically published 10
November 2005.
Presented in part: 14th Annual Scientific Meeting of the Society for Healthcare
Epidemiology of America, Philadelphia, Pennsylvania, 17–20 April 2004 (abstract 97).
Reprints or correspondence: Dr. Eli Perencevich, VA Maryland Health Care
System, 100 N. Greene St., Lower Level, Baltimore, MD 21201 (eperence
@epi.umaryland.edu).
Clinical Infectious Diseases 2005; 41:1734–41
2005 by the Infectious Diseases Society of America. All rights reserved.
1058-4838/2005/4112-0008$15.00
1734 • CID 2005:41 (15 December) • Perencevich et al.
therapy [1]. Unfortunately, very few clinical studies
have attempted to determine optimal treatment durations for conditions commonly treated with vancomycin. In the absence of such studies, many physicians
likely prescribe antibiotics for inordinate lengths of
time, because, as a group, physicians are risk averse [2,
3]. Few studies have addressed physician risk preference
for acceptable antimicrobial therapeutic failure rates or
their impact on treatment duration [4].
This study evaluated acceptable failure rates for vancomycin therapy of central venous catheter (CVC)–
associated coagulase-negative staphylococcal bacteremia among a cohort of infectious disease consultants
(IDCs). Physician preferences were elicited with a re-
alistic clinical vignette and assessed with methodology derived
from contingent valuation. The results delineate an acceptable
treatment failure threshold. Because decisions regarding changing the duration of antibiotic therapy are likely to be characterized by implicit trade-off between the likelihood of therapeutic success and the likelihood of complications, we sought
to establish a range of failure rates that might be considered
tolerable or “normative” by specialist physicians. These findings, in conjunction with results of clinical trials or cohort
studies, may facilitate selection of optimal durations of antimicrobial therapy for this condition.
METHODS
Pilot studies. We developed a realistic case scenario, which
included the history, physical examination findings, test results,
and treatment plan for a 50-year-old patient who developed
CVC-associated bacteremia caused by coagulase-negative
staphylococci after undergoing cardiac surgery. Management
included removal of the line and a standard course of vancomycin therapy.
We distributed the clinical vignette to 5 nationally recognized
experts in infectious diseases, asking for their opinions about
the case and an estimate of the treatment failure rate that would
be tolerable or acceptable, assuming “standard of care” treatment. Treatment failure was defined as “the patient having
blood cultures obtained from a peripheral vein positive for
coagulase-negative staphylococci 3–5 days after completion of
the vancomycin therapy.”
After receipt of comments and estimated failure rates from
the expert physicians, we modified the clinical description and
sent it to 55 IDCs who practiced in both academic medical
centers and community hospitals. We asked these 55 IDCs to
provide the treatment failure rate that they would find acceptable, given the treatment standard of care plan that was
outlined. In this pilot study, 21 IDCs received an open-ended
question and a request to determine the acceptable failure rate,
and 34 IDCs received 5 ranges of failure rates in a multiplechoice format.
The 21 IDCs who received the open-ended questionnaires
reported a mean acceptable failure rate (SD) of 6.3% 6.7% and a median acceptable failure rate of 5% (range, 0%–
25%). The 34 IDCs who completed the multiple-choice portion
of the pilot study reported a median acceptable failure rate of
10.1% but !1%, with 91% of responses indicating an acceptable
failure rate of !2% and none accepting a failure rate of 15%.
We used these pilot study results to determine the 10 possible
acceptable failure rates proffered in the formal Infectious Diseases Society of America (IDSA) Emerging Infections Network
(EIN) survey.
Study subjects. The EIN is a provider-based sentinel network, established by a Cooperative Agreement Program Award
from the Centers for Disease Control and Prevention (CDC)
in 1995. It consists of IDCs who belong to either the IDSA or
the Pediatric Infectious Diseases Society, who regularly engage
in clinical activity, and who volunteer to participate in network
activities.
Questionnaire. In August 2003, the EIN distributed questionnaires titled “Failure Rates for Therapy of Bacteremia Associated with Central Venous Catheters” via facsimile and email to 687 members in North America who practice as adult
IDCs. Nonresponders received a second and third survey 2 and
4 weeks later, respectively.
The surveys included a 1-page introduction to the topic, a
1-page clinical vignette, and the questionnaire (appendix 1,
figure A1) Treatment failure was defined as it was in the pilot
study. However, instead of offering an open-ended or multiplechoice question, the EIN questionnaire offered its members a
specific failure rate to accept or reject, assuming “standard of
care” therapy. In all 3 distributions, individual members received the same acceptable failure rate, which they could accept
or reject. Results from the pilot study established the range of
acceptable failure rates.
To capture all likely values, we set the upper range of failure
rates at 30%, because 1 IDC in the open-ended pilot questionnaire reported a 25% failure rate as acceptable. For the
lower range, we selected 0.01%, because 2 IDCs in the openended question group reported 0% and 0.5% as acceptable
failure rates, and 4 in the multiple-choice group selected “less
than !0.1%” as an acceptable failure rate. We then established
a range of 10 acceptable failure rates that were used in the EIN
survey instrument as possible choices. Because the median failure rate in the open-ended portion of the pilot study was 5%,
and because the most frequently selected responses in the multiple-choice portion of the pilot study were “0.1% to 1%” and
“1% to 2%,” we selected the following 10 choices for acceptable
failure rates in the EIN survey: 0.01%, 0.1%, 1%, 2%, 3%, 5%,
10%, 15%, 20%, and 30%. Furthermore, we anticipated that
300 members would respond to the survey. Thus, on average,
we expected that 30 IDCs would respond to each questionnaire
with 1 of the 10 choices of the possible failure rates.
The EIN randomized distribution of questionnaires with the
10 failure rates inserted in the last question to members practicing within the 9 regions of the United States distinguished
by the CDC. Each EIN member received only 1 potential acceptable failure rate to which they were asked to respond yes
or no (e.g. 10%). The EIN survey also asked its members to
provide the number of years that they had been in clinical
practice.
Elicitation of preferences. The approach used for elicitation of preferences in this study was closely related to economic
studies of willingness-to-pay or willingness-to-accept payment
performed using “contingent valuation” methodology [5]. We
Acceptable Treatment Failure Rates • CID 2005:41 (15 December) • 1735
Table 1. Geographic distribution of Emerging Infections Network (EIN) members who
responded to a questionnaire regarding acceptable treatment failure rates.
Region
No. of
EIN respondents
New England
Mid Atlantic
33
61
East North-Central
West North-Central
South Atlantic
51
17
75
East South-Central
West South-Central
Mountain
15
39
25
Pacific
US territories
Canada
53
2
3
Total
374
used a “take it or leave it” approach, in which study participants
are presented with only a single bid, which they can accept or
reject. IDCs were asked the following question:
Given the standard treatment scenario described above,
do you feel that the treatment failure rate XX% listed
below would be acceptable (i.e., the rate listed below is
not too high; it is less than or equal to the percent of
patients that you would expect to fail appropriate
therapy)?
The subjects were asked to answer “yes” or “no” to the above
question, indicating that the failure rate (XX%) listed was either
acceptable or not acceptable.
Statistical analysis. For each of the 10 failure rates randomly distributed to the IDCs, the proportions of responders
who found the number acceptable and who found it unacceptable were calculated. A logistic regression analysis evaluated
the relationship between willingness to accept a specific failure
rate and the failure rate offered to the subject in the questionnaire. Median and upper and lower quartile failure rates were
estimated for the entire group of respondents. Univariable and
multivariable logistic regression was performed to identify respondent characteristics that might influence willingness to accept failure.
Statistical analyses were performed using Stata for Macintosh, version 8.0 (Stata), and Excel 2000 (Microsoft). All aspects
of this study were approved by the University of Maryland
Institutional Review Board (Baltimore).
RESULTS
Overall, 374 (54%) of 687 adult IDCs who were members of
EIN responded to the survey. This response rate is similar to
the rates from other EIN surveys [4, 6, 7]. Subjects were dis1736 • CID 2005:41 (15 December) • Perencevich et al.
tributed evenly across the entire United States (table 1). In
addition, 3 subjects were from Canada, and 2 were from a US
territory. The mean length of time (SD) that the respondents
had been in clinical practice was 15.9 8.6 years.
There were at least 26 respondents for each category of therapeutic failure rates, with a mean of 37.4 respondents in each
category (table 2). Overall, the proportion of IDCs who found
a particular failure rate to be acceptable decreased in a nonlinear
fashion as the listed failure rate increased. One hundred percent
of IDCs found a failure rate of 0.01% for the clinical vignette
to be acceptable, whereas 2 (5.1%) of 39 found a failure rate
of 30% to be acceptable (table 2). In the univariable logistic
regression model, with given failure rate as the predictor for
the binary outcome “yes” (the failure rate was acceptable) or
“no” (the failure rate was not acceptable), the odds that an
IDC would find a given failure rate to be acceptable decreased
by 0.81 for each 1% increase in a given failure rate (OR, 0.81;
95% CI, 0.77–0.85; P ! .001). In other words, if the failure rate
in the clinical vignette was increased from 1% to 2%, the odds
that IDCs would accept this increase would decrease by a factor
of 0.81; if the failure rate increased from 5% to 10%, the odds
that this increase would be acceptable would decrease by a
factor of 0.81 to the fifth power (0.815, or 0.35). A fitted logistic
curve of the expected proportion of IDCs willing to accept a
failure rate across a range of given potential failure rates is
presented in figure 1. The projected median failure rate was
6.8%. Thus, presented with a failure rate of 6.8%, one-half of
the study subjects would have been expected to agree that that
failure rate was acceptable. Seventy-five percent of subjects
would have found a failure rate of 1.6% to be acceptable, and
25% of subjects would have found a failure rate as high as
11.9% to be acceptable.
A multivariable logistic regression model was created to adjust for potential confounding of acceptable failure rates by
number of years in practice (a marker for IDC experience). No
Table 2. Acceptable rates of treatment failure, according to 374
responding infectious diseases consultants (IDCs).
Response, no. (%)
of IDCs
No. of
IDCs who
responded
Reject
Accept
0.01
0.1
1
26
44
40
0
4
9
26 (100)
40 (90.9)
31 (77.5)
2
3
5
34
37
41
10
14
20
24 (70.6)
23 (62.2)
21 (51.2)
10
15
20
36
46
31
29
42
26
7 (19.4)
4 (8.7)
5 (16.1)
30
39
37
2 (5.1)
Proposed rate
of treatment failure, %
Figure 1. Relationship between failure rate given to physicians and the percentage of physicians willing to accept that failure rate. Diamonds
represent 26–46 respondents. The line is a fitted logistic curve utilizing all of the data describing the expected percentage of physicians willing to
accept each failure rate.
change in the OR for acceptable failure rate was identified in
the adjusted model, and the number of years in practice was
not a predictor of acceptance of a particular failure rate (OR,
1.03; 95% CI, 0.998–1.06; P p .41)
DISCUSSION
The emergence of antibiotic resistance in bacterial pathogens
is directly related to antibiotic use, with the inevitable selection
of resistant bacterial pathogens [1]. Increasing rates of infections caused by antibiotic-resistant organisms have led to several types of antibiotic management programs designed to optimize prescribing in an effort to control the emergence of
antimicrobial resistance [8–12]. These strategies include restricted antibiotic formularies in hospitals and targeted education to reduce the prescription of antibiotics for nonbacterial
infectious conditions (i.e., upper respiratory infections). A large
proportion of antibiotic prescriptions are considered to be inappropriate or unnecessary [13–16]. Given the emergence of
vancomycin-resistant enterococci and vancomycin-resistant
Staphylococcus aureus [17–19], antibiotic stewardship programs,
including the CDC’s 12 Steps to Prevent Antimicrobial Resistance
among Hospitalized Adults, often aim to limit excessive vancomycin use [20, 21]. In addition, the CDC’s Hospital Infection
Control Practices Advisory Committee Guidelines for the Prevention and Control of Vancomycin Resistance advocates prudent use of vancomycin [22, 23].
Prolonged courses of antibiotic therapy may foster the development of resistance to not only the prescribed antibiotic,
but also to multiple other classes of antibiotics. Some investigators have suggested that, for certain diseases, reduction in
antibiotic treatment duration will decrease the emergence of
antimicrobial resistance [1]. Therefore, in the hospital setting,
shortening the excessive duration of appropriate vancomycin
therapy may be optimal, particularly for infections caused by
pathogens of relatively low virulence, such as coagulase-negative
staphylococci.
This current study determined an acceptable treatment failure rate for a representative case of CVC-associated bacteremia
caused by coagulase-negative staphylococci on the basis of responses from a group of clinical IDCs. CVC-associated, coagulase-negative staphylococcal bacteremia was chosen because
it represented a common scenario treated with vancomycin by
diverse groups of clinicians, such as surgeons, general internists,
and subspecialists, including pulmonary critical care and IDCs.
We chose the “take-it-or-leave it” methodology and not an
open-ended question format for our study, because responding
to open-ended questions requires more time from respondents
and decreases study participation [24]. For this reason and
others, the methodology we chose is thought to better approximate “real-life” decision-making [24]. In addition, elicitation of preferences using contingent valuation is not subject
to framing effects and anchoring biases (i.e., sensitivity of subAcceptable Treatment Failure Rates • CID 2005:41 (15 December) • 1737
ject responses to opening bids, ranges of bids provided, and
phrasing of questions) that can occur when open-ended questions are used [25]. This methodology is similar to economic
contingent valuation, which was initially utilized to generate
monetary values for such abstract quantities as wildlife preservation and water quality [5]. It has also been used to determine willingness to pay for diverse health products, including
needlestick avoidance devices, in-vitro fertilization, autologous
blood donation, and asthma therapies [26–30]. Using this
methodology, we previously determined acceptable failure rates
for standard therapy for diabetic osteomyelitis [4].
The examined case scenario (appendix 1, figure A1) involved
treatment of a patient who developed coagulase-negative staphylococcal bacteremia on day 3 after undergoing coronary artery
bypass surgery. The median acceptable failure rate for treatment
of coagulase-negative staphylococcal, CVC-related bacteremia
reported by IDCs in this analysis was 6.8%. In addition, this
methodology was able to determine that 75% of respondents
would have found a failure rate of 1.6% to be acceptable, and
25% of respondents would have found a failure rate of 11.9%
to be acceptable. This additional calculation of acceptable
ranges would not have been possible with the open-ended or
multiple-choice questionnaire formats.
Coagulase-negative staphylococci are the most frequently
isolated pathogens in nosocomial and catheter-related bloodstream infections, and 180% of these pathogens are resistant
to methicillin, necessitating administration of vancomycin [31,
32]. Fortunately, mortality associated with coagulase-negative
staphylococcal, catheter-related bacteremia is quite low
(∼0.7%) [31]. Importantly, the recently published guidelines
for the management of catheter-related bloodstream infections
from the IDSA, the American College of Critical Care Medicine,
and the Society for Healthcare Epidemiology of America state
that “there are no compelling data to support specific recommendations regarding the duration of therapy for devicerelated infections” [31, p. 1253]. These guidelines recommend
a choice between removal of the catheter and treatment with
systemic antibiotics for 5–7 days, or retention of the catheter
and 10–14 days of systemic antibiotics for patients with CVCrelated, coagulase-negative staphylococcal bacteremia. There
have been no randomized trials comparing these treatments
[31]. However, these guidelines do reference an observational
study by Raad et al. [33] that compared treatment failure (defined as recurrent bacteremia within 12 weeks) in cases of CVCrelated, coagulase-negative staphylococcal bacteremia among
patients who had catheters removed and among those for
whom the catheters were retained. Recurrent bacteremia occurred in 20% of patients with retained catheters, compared
with 3% of those who had their catheters removed (P ! .05).
If we interpret these failure rates using our survey data, we can
estimate that 69% of the IDCs surveyed would have found the
1738 • CID 2005:41 (15 December) • Perencevich et al.
3% recurrent bacteremia failure rate reported in the catheter
removal arm of the study to be acceptable.
Apart from pediatric trials, to our knowledge, no randomized, clinical trials have specifically reported treatment outcomes for CVC-associated, coagulase-negative staphylococcal
bacteremia [34, 35]. Thus, our methodology will be helpful in
the interpretation of future trials [36]. For example, in a hypothetical clinical trial, it is likely that failure rates with recurrent bacteremia will be lower for patients treated for 14 days
than for those treated for 3 days. We suggest that acceptable
failure rates need to be determined from physicians and patients
before any recommendations for an optimal treatment duration
are made. If the rate obtained from a study for 3 days of therapy
is deemed to be acceptable, then 11 days of excess antibiotic
treatment may be avoided. Thus, having an estimate of this
acceptable rate allows a frame of reference when evaluating
outcomes of clinical trials or retrospective cohort studies. In
addition, interpretation of the results of trials that have used
the methodology could prove useful to those who are developing clinical practice guidelines. Our recent survey, which
measured the acceptable median failure rate for treating diabetic
foot osteomyelitis, was incorporated into the recent IDSA
guidelines for the diagnosis and treatment of diabetic foot infections [4, 37].
A potential limitation of the methodology employed is that
our questionnaire used only a single case scenario and a single
definition of failure and was administered to only 1 subspecialty
of physicians. For instance, a similar scenario that involved a
patient who had undergone a recent prosthetic valve placement
and developed methicillin-resistant S. aureus bacteremia would
likely yield very different results. Thus, our measured acceptable
failure rate cannot be used in assessing trials in other clinical
conditions. However, the contingent valuation methodology
permits derivation of considerable information from a single
case and administration of the survey to a group of interest.
The development of a series of surveys with alternative scenarios is feasible, given the existence of groups such as the
IDSA EIN. The methodology could be used in the future to
assess preferences in other important groups, such as general
internists and patients, which would move us toward the goal
of determining treatment preferences from a broader perspective, such as a societal one.
Some issues not considered in this study are the acceptable
rate of treatment failure from the patients’ perspectives and the
impact of failure events on patients’ quality of life. No appropriate treatment strategy can be deployed without understanding these values. Some have advocated a shared decision-making model in antibiotic prescriptions that incorporates
individual physician and patient understanding and preferences
[38, 39]. Understanding acceptable failure rates obtained from
a large group of physicians could help to improve communi-
cation with patients and facilitate informed decision-making
from the patient’s perspective, because benchmarks allow clinicians to provide greater clarity to patients on what clinicians
regard as acceptable. In addition, our study captured considerable between-physician variability, with an order of magnitude difference in acceptable failure rates between the 25th and
75th percentiles. This finding has 2 important implications: (1)
there is no absolute in terms of what constitutes acceptable
failure rates among IDCs, and (2) communication of this difference to physicians could inform those physicians whose preferences fall far outside the median acceptable failure rate to
consider reviewing their treatment practices, relative to standard-of-care practices.
The true benefit of the methodology is that it provides plausible estimates of acceptable treatment failure rates among
expert physicians. These acceptable failure rates can be used to
assess interventions to reduce excess antibiotic treatment duration in the specific scenario of CVC-associated, coagulasenegative staphylococcal bacteremia. The potential reduction of
excess treatment duration is an important step towards the
ultimate goal of reducing exposure to antibiotics and thus slowing the emergence of antibiotic-resistant organisms.
Acknowledgments
Financial support. VA Health Services Research and Development Service Research Career Development Award (RCD-02026-1 to E.N.P.) and
National Institutes of Health (K23 AI01752-01A1 to A.D.H.), and the CDC
(cooperative agreement U50/CCU112346). The contents of this article are
solely the responsibility of the authors and do not necessarily represent
the official views of the CDC.
Potential conflicts of interest. All authors: no conflicts.
Acceptable Treatment Failure Rates • CID 2005:41 (15 December) • 1739
APPENDIX
Figure A1. Questionnaire distributed to the Infectious Diseases Society of America Emerging Infections Network
References
1. Levy SB. Confronting multidrug resistance: a role for each of us. JAMA
1993; 269:1840–2.
2. Fiscella K, Franks P, Zwanziger J, Mooney C, Sorbero M, Williams GC.
Risk aversion and costs: a comparison of family physicians and general
internists. J Fam Pract 2000; 49:12–7.
3. Pearson SD, Goldman L, Orav EJ, et al. Triage decisions for emergency
department patients with chest pain: do physicians’ risk attitudes make
the difference? J Gen Intern Med 1995; 10:557–64.
4. Perencevich EN, Kaye KS, Strausbaugh LJ, Fisman DN, Harris AD.
Acceptable rates of treatment failure in osteomyelitis involving the
diabetic foot: a survey of infectious diseases consultants. Clin Infect
Dis 2004; 38:476–82.
5. Arrow K, Solow R, Portney P, Learner E, Radner R, Schuman H. Report
of the NOAA Panel on Contingent Valuation. Washington, DC: National Oceanic and Atmospheric Administration, 1993. Available at:
http://www.darp.noaa.gov/library/pdf/cvblue.pdf. Accessed 20 October
2005.
6. Strausbaugh LJ, Jernigan DB, Liedtke LA. National shortages of antimicrobial agents: results of 2 surveys from the Infectious Diseases
Society of America Emerging Infections Network. Clin Infect Dis
2001; 33:1495–501.
7. Lockhart PB, Brennan MT, Fox PC, Norton HJ, Jernigan DB, Strausbaugh LJ. Decision-making on the use of antimicrobial prophylaxis for
dental procedures: a survey of infectious disease consultants and review.
Clin Infect Dis 2002; 34:1621–6.
8. Avorn J, Soumerai SB, Taylor W, Wessels MR, Janousek J, Weiner M.
Reduction of incorrect antibiotic dosing through a structured educational order form. Arch Intern Med 1988; 148:1720–4.
9. Goldmann DA, Weinstein RA, Wenzel RP, et al. Strategies to prevent
and control the emergence and spread of antimicrobial-resistant microorganisms in hospitals: a challenge to hospital leadership. JAMA
1996; 275:234–40.
10. Duncan RA. Controlling use of antimicrobial agents. Infect Control
Hosp Epidemiol 1997; 18:260–6.
11. McGowan JE Jr. Do intensive hospital antibiotic control programs
prevent the spread of antibiotic resistance? Infect Control Hosp Epidemiol 1994; 15:478–83.
12. Pestotnik SL, Classen DC, Evans RS, Burke JP. Implementing antibiotic
practice guidelines through computer-assisted decision support: clinical
and financial outcomes. Ann Intern Med 1996; 124:884–90.
13. Gonzales R, Bartlett JG, Besser RE, et al. Principles of appropriate
antibiotic use for treatment of acute respiratory tract infections in
adults: background, specific aims, and methods. Ann Intern Med
2001; 134:479–86.
14. Cooper RJ, Hoffman JR, Bartlett JG, et al. Principles of appropriate
antibiotic use for acute pharyngitis in adults: background. Ann Intern
Med 2001; 134:509–17.
15. Hickner JM, Bartlett JG, Besser RE, Gonzales R, Hoffman JR, Sande
MA. Principles of appropriate antibiotic use for acute rhinosinusitis
in adults: background. Ann Intern Med 2001; 134:498–505.
16. Gonzales R, Sande MA. Uncomplicated acute bronchitis. Ann Intern
Med 2000; 133:981–91.
17. Centers for Disease Control and Prevention. Vancomycin-resistant
Staphylococcus aureus—Pennsylvania, 2002. MMWR Morb Mortal
Wkly Rep 2002; 51:902.
18. Centers for Disease Control and Prevention. Staphylococcus aureus resistant to vancomycin—United States, 2002. MMWR Morb Mortal
Wkly Rep 2002; 51:565–7.
19. Centers for Disease Control and Prevention. Vancomycin-resistant
Staphylococcus aureus—New York, 2004. MMWR Morb Mortal Wkly
Rep 2004; 53:322–3.
20. Centers for Disease Control and Prevention. 12 Steps to prevent an-
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
timicrobial resistance among hospitalized adults: campaign to prevent
antimicrobial resistance in healthcare settings. Atlanta, GA: US Centers
for Disease Control and Prevention (CDC), 2003. Available at: http:/
/www.cdc.gov/drugresistance/healthcare/ha/12steps_HA.htm.Accessed
3 November 2004.
Sunenshine RH, Liedtke LA, Jernigan DB, Strausbaugh LJ. Role of
infectious diseases consultants in management of antimicrobial use in
hospitals. Clin Infect Dis 2004; 38:934–8.
Centers for Disease Control and Prevention. Interim guidelines for
prevention and control of staphylococcal infection associated with reduced susceptibility to vancomycin. JAMA 1997; 278:461–2.
Hospital Infection Control Practices Advisory Committee (HICPAC).
Recommendations for preventing the spread of vancomycin resistance.
Infect Control Hosp Epidemiol 1995; 16:105–13. [erratum: Infect Control Hosp Epidemiol 1995; 16:498].
Klose T. The contingent valuation method in health care. Health Policy
1999; 47:97–123.
Hammond JS, Keeney RL, Raiffa H. The hidden traps in decision
making. Clin Lab Manage Rev 1999; 13:39–47.
Fisman DN, Mittleman MA, Sorock GS, Harris AD. Willingness to pay
to avoid sharps-related injuries: a study in injured health care workers.
Am J Infect Control 2002; 30:283–7.
Neumann PJ, Johannesson M. The willingness to pay for in vitro
fertilization: a pilot study using contingent valuation. Med Care
1994; 32:686–99.
Lee SJ, Neumann PJ, Churchill WH, Cannon ME, Weinstein MC,
Johannesson M. Patients’ willingness to pay for autologous blood donation. Health Policy 1997; 40:1–12.
Lee SJ, Liljas B, Neumann PJ, Weinstein MC, Johannesson M. The
impact of risk information on patients’ willingness to pay for autologous blood donation. Med Care 1998; 36:1162–73.
Zillich AJ, Blumenschein K, Johannesson M, Freeman P. Assessment
of the relationship between measures of disease severity, quality of life,
and willingness to pay in asthma. Pharmacoeconomics 2002; 20:
257–65.
Mermel LA, Farr BM, Sherertz RJ, et al. Guidelines for the management
of intravascular catheter–related infections. Clin Infect Dis 2001; 32:
1249–72.
Edmond MB, Wallace SE, McClish DK, Pfaller MA, Jones RN, Wenzel
RP. Nosocomial bloodstream infections in United States hospitals: a
three-year analysis. Clin Infect Dis 1999; 29:239–44.
Raad I, Davis S, Khan A, Tarrand J, Elting L, Bodey GP. Impact of
central venous catheter removal on the recurrence of catheter-related
coagulase-negative staphylococcal bacteremia. Infect Control Hosp Epidemiol 1992; 13:215–21.
Jantausch BA, Deville J, Adler S, et al. Linezolid for the treatment of
children with bacteremia or nosocomial pneumonia caused by resistant
gram-positive bacterial pathogens. Pediatr Infect Dis J 2003; 22:
S164–71.
Kaplan SL, Deville JG, Yogev R, et al. Linezolid versus vancomycin for
treatment of resistant Gram-positive infections in children. Pediatr
Infect Dis J 2003; 22:677–86.
Raad I, Darouiche R, Vazquez J, et al. Efficacy and safety of weekly
dalbavancin therapy for catheter-related bloodstream infection caused
by gram-positive pathogens. Clin Infect Dis 2005; 40:374–80.
Lipsky BA, Berendt AR, Deery HG, et al. Diagnosis and treatment of
diabetic foot infections. Clin Infect Dis 2004; 39:885–910.
Butler CC, Kinnersley P, Prout H, Rollnick S, Edwards A, Elwyn G.
Antibiotics and shared decision-making in primary care. J Antimicrob
Chemother 2001; 48:435–40.
Elwyn G, Gwyn R, Edwards A, Grol R. Is ’shared decision-making’
feasible in consultations for upper respiratory tract infections? Assessing the influence of antibiotic expectations using discourse analysis.
Health Expect 1999; 2:105–17.
Acceptable Treatment Failure Rates • CID 2005:41 (15 December) • 1741