Download Final Discussion Handout - Winona State University

Final Exam STAT 601 - (Discussion Questions)
Example 1: I wish to examine the effect of a certain drug injected into laboratory rats on
such outcome variables as protein and lipid levels. How should the study be conducted
and what method of analysis would be most appropriate for analyzing the data?
Example 2: We wish to compare the survival time (in weeks) of heart patients that have
had heart catheterization performed versus those that haven’t. In an attempt to create a
fair comparison and control for potential confounding factors patients in each group are
matched according race, age, socio-economic status, smoking history, cholesterol level,
blood gas levels, etc... How could/should the resulting data be analyzed?
What is we wanted look at death as the outcome? What might analyses might we use?
Example 3: We wish to compare the percentage of mother’s having their babies
delivered at two large hospitals in the Twin Cities that have cocaine or
methamphetamines found in the meconium. How could the study be conducted and what
method of statistical analysis should be used? (see case study in appendix)
What if we wanted to identify factors related to the presence of cocaine and/or
methamphetamine in the meconium. What should we do?
Example 4: We wish to compare the readings given two medical laboratory machines
(Hitachi 717 & Hitachi 736) at the Mayo Clinic used to measure triglycerides. How
could we conduct the study and how should the data be analyzed? (Datafile: Mayo
Measuring Triglycerides)
1
Example 5: We wish estimate the percentage of alcohol dependent hospitalized patients
(non-addiction) in Rochester Methodist and St. Mary’s Hospitals. How could we conduct
the study and how could the data be analyzed?
Following up on the same example, suppose in most community hospitals the alcohol
dependent prevalence is approximately 20% and we wish to determine whether the
percentage in these two Rochester hospitals is different. What type of analysis should be
conducted?
Example 6: We have two suppliers of laboratory rats and we are trying to decide which
provides us with the best “product”. One of the primary concerns is that the rats should
all have the same weights upon arrival, thus we would like to choose the supplier that
provides us with the rats having the most consistent weight. What kind of study should
we conduct and how should we analyze the resulting data?
Example 7: An undesirable side effect of some antihistamines is drowsiness, a
consequence of the effect of the drug on the central nervous system. Researchers
compared the effect of two antihistamines and a placebo by measuring what is called the
flicker frequency in volunteers who had taken the treatment. There were 9 subjects who
took each of the 3 treatments. Flicker frequencies were recorded 6 hours after the drug
was administered. What type of design has been used? How should the analysis be
conducted?
Example 8: We wish to determine if magnetic insoles are effective in relieving
nonspecific foot pain in the workplace. How might we conduct are study and what
methods of analysis might we use?
2
Example 9: Neuropsychological Dysfunction in Children with
Chronic Low-Level Lead Absorption
The paper "Neuropsychological Dysfunction in Children with Chronic Low-Level Lead
Absorption" published in Lancet, 1, pp. 708-715, March 1975, examined the effects of
exposure to lead on the psychological and neurological well-being of children. In
summary, a group of children who lived near a lead smelter in El Paso, Texas, were
identified and their blood levels of lead were measured. An exposed group of 46 children
were identified who had blood-lead levels > 40 micrograms/milliliter in 1972 (or in few
cases 1973). A control group of 78 children was also identified with blood-lead levels <
40 micrograms/milliliter in both 1972 and 1973. All children in the study lived in close
proximity to the lead smelter.
The Data (Datafile: Lead El-Paso)
The three sets of variables used in the study could be classified as (1) IQ related measures
(e.g. verbal IQ), (2) symptomatic variables (e.g. colic) and (3) neurological function
variables (e.g. # finger taps in 10 seconds). In your analysis of these data you should try
to answer the following question.
General Research Question:
WHAT, IF ANY, EFFECT DOES LEAD EXPOSURE HAVE ON THE THREE
SETS VARIABLES USED IN THE STUDY?
There are potential confounding factors present such as socio-economic status, proximity
to plant, proximity to plant in early years of life, total years of exposure, gender, etc.
The variables in the data file are:
Demographic Variables






Area - residence of child in August 1972 (1 = 0-1 miles from smelter, 2 = 1 - 2.5
miles, 3 = 2.5 - 4.1 miles)
First 2 Years? - did the child live near the plant during first two years of life
(Yes, No)
Age - age of child in years followed by months (e.g. 1011 = 10 yrs. 11 mo.)
Agegroup - age group of child (1 = 0 - 4 yrs., 2 = 5 - 9 yrs, 3 = 10 - 14 yrs., 4 =
15 - 19 yrs.)
Gender – Male, Female
Social Index - Hollingshead Index of Social Status
Psychological Function Variables (i.e. IQs)



iqv - verbal IQ score
iqp - performance IQ score (block arrangements, pattern recognition etc.)
iqf - total IQ (this is not a direct sum or average of IQV and IQP)
Lead Exposure Variables
3





Lead Group - blood lead level group (1 = blood lead level below 40 in both 1972
and 1973, 2 = greater than 40 in both 1972 and 1973 or greater than 40 in 1973
alone, 3 = greater than 40 in 1972 but less than 40 in 1973)
Lead Level - indicator of high lead in 1972 or 1973 (High = greater than 40 in
1972 and/or 1973, Low = below 40 in both 1972 and 1973)
ld72 - actual lead level in 1972
ld73 - actual lead level in 1973
totyrs - total years of life living in close proximity to the lead smelter
Symptomatic Variables





Pica - indicator of pica (Yes, No)
Colic - did the child have colic (Yes, No)
Clumsy - does the child seem abnormally clumsy (Yes, No)
Irritated - does the child seem abnormally irritable (Yes, No)
Convulsions - does the child have a history of convulsions (Yes, No)
Neurological Function (motor skills)




maxtap - maximum # of taps from left and right hand trials (removes hand
dominance)
minvis - minimum reaction time to a visual stimulus from left and right hand
trials (removes hand dominance)
minaud - minumum reaction time to an auditory stimulus from left and right hand
trials (removes hand dominance)
maxfwt - maximum # of finger to wrist taps from left and right hand trials
(removes hand dominance)
POSSIBLE QUESTIONS AND ANALYSES
1.) Formulate potential questions of interest.
2.) Discuss how we might conduct the analysis.
3.) Conduct the analysis in JMP (to the best of our ability).
4.) Discuss results.
4
EXAMPLE 10: EFFECTIVENESS OF RIGHT HEART CATHETERIZATION
IN CRITICALLY ILL PATIENTS (JAMA, 1996), Conners et al.
(DATAFILE: Right Heart Catheterization)
OBJECTIVE: To examine the association between the use of right heart catheterization (RHC) during the
first 24 hours of care in the intensive care unit (ICU) and subsequent survival, length of stay, intensity of
care, and cost of care.
DESIGN: Prospective cohort study.
SETTING: Five US teaching hospitals between 1989 and 1994.
SUBJECTS: A total of 5735 critically ill adult patients receiving care in an ICU for 1 of 9 prespecified
disease categories.
MAIN OUTCOME MEASURES: Survival time, cost of care, intensity of care, and length of stay in the
ICU and hospital, determined from the clinical record and from the National Death Index. A propensity
score for RHC was constructed using multivariable logistic regression. Case-matching and multivariable
regression modeling techniques were used to estimate the association of RHC with specific outcomes after
adjusting for treatment selection using the propensity score. Sensitivity analysis was used to estimate the
potential effect of an unidentified or missing covariate on the results.
RESULTS: By case-matching analysis, patients with RHC had an increased 30-day mortality (odds ratio,
1.24; 95% confidence interval, 1.03-1.49). The mean cost (25th, 50th, 75th percentiles) per hospital stay
was $49 300 ($17 000, $30 500, $56 600) with RHC and $35 700 ($11 300, $20 600, $39 200) without
RHC. Mean length of stay in the ICU was 14.8 (5, 9, 17) days with RHC and 13.0 (4, 7, 14) days without
RHC. These findings were all confirmed by multivariable modeling techniques. Subgroup analysis did not
reveal any patient group or site for which RHC was associated with improved outcomes. Patients with
higher baseline probability of surviving 2 months had the highest relative risk of death following RHC.
Sensitivity analysis suggested that a missing covariate would have to increase the risk of death 6-fold and
the risk of RHC 6-fold for a true beneficial effect of RHC to be misrepresented as harmful.
CONCLUSION: In this observational study of critically ill patients, after adjustment for treatment
selection bias, RHC was associated with increased mortality and increased utilization of resources. The
cause of this apparent lack of benefit is unclear. The results of this analysis should be confirmed in other
observational studies. These findings justify reconsideration of a randomized controlled trial of RHC and
may guide patient selection for such a study.
DESCRIPTION OF VARIABLES
Demographics and Disease Category
Variable name
Age
Sex
Race
Edu
Income
Ninsclas
Cat1
Variable Definition
Age
Sex
Race
Years of education
Income
Medical insurance
Primary disease category
Categories of Admission Diagnosis
Diagnosis variables are all coded as (Y or N)
Resp
Respiratory Diagnosis
Card
Cardiovascular Diagnosis
Neuro
Neurological Diagnosis
5
Gastr
Renal
Meta
Hema
Seps
Trauma
Ortho
Das2d3pc
Dnr1
Ca
Surv2md1
Aps1
Scoma1
Wtkilo1
Temp1
Meanbp1
Resp1
Hrt1
Pafi1
Paco21
Ph1
Wblc1
Hema1
Sod1
Pot1
Crea1
Bili1
Alb1
Gastrointestinal Diagnosis
Renal Diagnosis
Metabolic Diagnosis
Hematologic Diagnosis
Sepsis Diagnosis
Trauma Diagnosis
Orthopedic Diagnosis
DASI (Duke Activity Status Index)
DNR status on day1 (Yes or No)
Cancer (Yes, No, Metastatic)
Support model estimate of the prob. of surviving 2 months
APACHE score
Glasgow Coma Score
Weight
Temperature
Mean blood pressure
Respiratory rate
Heart rate
PaO2/FIO2 ratio
PaCo2
PH
WBC
Hematocrit
Sodium
Potassium
Creatinine
Bilirubin
Albumin
Categories of Comorbidities Illness
Cardiohx
Transhx
Amihx
Swang1 *
Sadmdte
Dthdte
Lstctdte
Dschdte
Death *
Acute MI, Peripheral Vascular Disease, Severe Cardiovascular Symptoms
(NYHA-Class III), Very Severe Cardiovascular Symptoms (NYHA-Class IV)
Congestive Heart Failure
Dementia, Stroke or Cerebral Infarct, Parkinson’s Disease
Psychiatric History, Active Psychosis or Severe Depression
Chronic Pulmonary Disease, Severe Pulmonary Disease, Very Severe
Pulmonary Disease
Chronic Renal Disease, Chronic Hemodialysis or Peritoneal Dialysis
Cirrhosis, Hepatic Failure
Upper GI Bleeding
Solid Tumor, Metastatic Disease, Chronic Leukemia/Myeloma, Acute
Leukemia, Lymphoma
Immunosupperssion, Organ Transplant, HIV Positivity, Diabetes Mellitus
Without End Organ Damage, Diabetes Mellitus With End Organ Damage,
Connective Tissue Disease
Transfer (> 24 Hours) from Another Hospital
Definite Myocardial Infarction
Right Heart Catheterization (RHC vs. No RHC)
Study Admission Date
Date of Death
Date of Last Contact
Hospital Discharge Date
Death at any time up to 180 Days
Ptid
Patient ID
Chfhx
Dementhx
Psychhx
Chrpulhx
Renalhx
Liverhx
Gibledhx
Malighx
Immunhx
6
POSSIBLE QUESTIONS AND ANALYSES
1.) Formulate potential questions of interest.
2.) Discuss how we might conduct the analysis.
3.) Conduct the analysis in JMP (to the best of our ability).
4.) Discuss results.
7
CASE STUDY 1: Study Finds Significant Mental Deficits in Toddlers Exposed to
Cocaine Before Birth Research Findings
Vol. 17, No. 5 (January 2003)
By Robert Mathias, NIDA NOTES Staff Writer
Since the mid-1980s, up to 1 million children born in the United States are estimated to
have been exposed to cocaine in the womb. Determining cocaine's impact on these
children's development has been difficult because there often are other possible
explanations for physical and mental problems the children may have, such as the
mother's use of other substances during pregnancy and poor prenatal care. Now, a NIDAsupported study that was able to separate the effects of cocaine from those of many other
such factors has found that children born to poor, urban women who used cocaine
throughout pregnancy were nearly twice as likely as children with similar backgrounds
but no prenatal cocaine exposure to have significant cognitive deficits during their first 2
years of life.
Mental Development Scores For Pre-natally Cocaine-Exposed and
Unexposed High-Risk Children
Tests of mental development at 6.5, 12, and 24 months showed average scores of
cocaine-exposed and unexposed children from comparable backgrounds were below the
normative score of 100 for children in the general population. At age 2, cocaine-exposed
children did significantly poorer in mental development than children in the comparison
group.
The study, led by Dr. Lynn Singer of Case Western Reserve University in Cleveland,
Ohio, is the first to show a clear association between prenatal cocaine exposure and
cognitive impairment in 2-year-olds. "Since cognitive performance at this age is
8
indicative of later performance, these children may continue to have learning difficulties
that need to be addressed when they reach school age," Dr. Singer says.
"The findings of this well-controlled study make an important contribution to a growing
body of knowledge about the effects of prenatal cocaine exposure that may help us to
identify those exposed children who are at increased risk of developmental harm," says
Dr. Vince Smeriglio of NIDA's Center on AIDS and Other Medical Consequences of
Drug Abuse. Previous findings from other NIDA-supported studies that have been
following cocaine-exposed children from birth have produced conflicting results about
cocaine's impact on developmental outcomes at this age, he notes. "Comparing and
contrasting the circumstances in this study with those found in other studies of cocaineexposed children may enable us to identify specific biological and environmental factors
that increase or reduce the developmental risk from cocaine exposure," Dr. Smeriglio
says.
The study followed a group of 415 infants born at a large urban teaching hospital from
1994 through 1996 to mothers from low socio-economic backgrounds who had been
identified by the hospital staff as being at high risk of drug abuse. Women who
participated in the study were given urine tests for drug use immediately before or after
delivery and interviewed shortly after they gave birth to produce estimates of the type,
frequency, and amounts of drugs they had used during pregnancy. Each baby's first stool,
known as meconium, also was analyzed for the presence of cocaine and its metabolites to
help establish the level of drug exposure. Of the 415 babies in the study, 218 had been
exposed to cocaine and 197 had not. Both groups of infants also had been exposed to
tobacco, alcohol, and marijuana during pregnancy.
Researchers measured the children's developmental progress at 6.5, 12, and 24 months of
age with the Bayley Scales of Infant Mental and Motor Development. Motor tests
assessed the infants' ability to control and coordinate their movements. Mental tests
assessed language, memory, and ability to solve problems at 12 and 24 months. For
example, children were asked to describe objects in pictures, remember where an object
had been hidden, and put shaped objects into the correct spaces cut out on form boards.
To isolate cocaine's effect, researchers adjusted test results for the effect of other risk
factors, such as other drugs used during pregnancy; characteristics of biological mothers
and alternative caregivers; the infants' head size, weight, length, and gestational age at
birth; and the quality of their postnatal home environments. The analysis showed that
while prenatal cocaine exposure had not affected the infants' motor development, it was
clearly linked to significant deficits in their cognitive performance at age 2. Cocaineexposed children scored 6 points lower on the Mental Development Index (MDI),
averaging 82.7 percent compared to 88.7 percent for unexposed children and an average
general population score of 100. Other findings include the following:
 From 6.5 to 24 months, MDI scores declined for both groups, but cocaineexposed children had a greater decline -- 14 points compared to a 9-point decline
for unexposed children.
 Almost 14 percent (13.7 percent) of cocaine-exposed children had scores in the
mental retardation range, below 70 on the MDI, nearly twice the 7.1-percent rate
9

found in the unexposed children and almost five times the rate (about 2.8 percent)
expected in the general population.
Nearly 38 percent (37.8 percent) of cocaine-exposed children had developmental
delays requiring remedial intervention, nearly double the 20.9 percent rate for
unexposed children.
The study found that other influences, including the mother's intelligence scores and
educational level, exposure to other substances, and the quality of the postnatal home
environment, also played significant roles in poor outcomes for cocaine-exposed
children. "However, after controlling for these factors in our analysis, we found that
cocaine still has a harmful effect on cognitive performance," Dr. Singer says. Additional
support for this conclusion comes from mothers' self-reports and biological data from
mothers and infants that established a direct link between cocaine dose and toddlers'
cognitive performance. These data showed that children of mothers who used more
cocaine and used it more frequently during pregnancy performed worse on the MDI than
children of mothers who used less of the drug.
"The only risk factor we couldn't completely control for is the effect of other drugs used
during pregnancy," Dr. Singer says, "because it is nearly impossible to find children who
have been exposed only to cocaine." The study partially adjusted for this influence by
including children who had been heavily exposed to alcohol, tobacco, and marijuana in
both groups. "Animal studies suggest there are possible synergistic effects of these drugs
in combination, and the study may not have been large enough to control for these
effects," she notes.
"We believe that cocaine exposure is a neurologic risk factor that may take a poor child
who has a lower IQ potential because of maternal and other risk factors and push him or
her over the edge to mental retardation," Dr. Singer says. For example, average IQ scores
for both cocaine-exposed and unexposed toddlers in the study were well below the
average score for the general population. "In effect, cocaine lowered the range of IQ
scores and that means more children may require early stimulation and educational
programs," she says.
"While many children in this study may require special educational services when they
enter school, it is important not to assume that the findings from a single study, with its
unique characteristics, necessarily apply to all cocaine-exposed children," cautions
NIDA's Dr. Smeriglio. Ultimately, NIDA's extensive portfolio of research on groups of
cocaine-exposed children being raised in a variety of settings should provide detailed
information about mother, child, environment, and drug-use characteristics that can be
used to develop interventions that reduce risk of harm and guide clinical care for cocaineexposed children.
Source:
Singer, L.T., et al. Cognitive and motor outcomes of cocaine-exposed infants. JAMA
287(15):1952-1960, 2002.
10