Osu Introduction to Epidemiology 2410 Exam 2 – Flashcards
Unlock all answers in this set
Unlock answersquestion
Koch-Henle Postulates Infectious Disease Etiologoy
answer
Microorganism must be present in every case Microoganism must be isolated and grown in pure culture Innoculation of microorganism into susceptible host must reproduce disease Microorganism must be observed and recovered
question
Consistency of independent investigations Criteria of Judgment Sir Bradford Hill (1964)
answer
Strength of association (Dose Response) Specificity of association Temporal relationship between causal factor and disease occurrence Coherence & Biological Plausibility
question
Consistency of independent investigations Criteria of Judgment Surgeon General's Report, 1964 Sir Bradford Hill, 1965
answer
Similar findings from independent students utilizing diverse methods of study in different populations under a variety of circumstances.
question
Meta-Analysis of Colon Cancer ; NSAIDs
answer
Combined RR = 0.57, 95% CI = 0.50-0.64
question
Criteria of Judgment Surgeon General's Report, 1964 Sir Bradford Hill, 1965
answer
Strength of association (Dose Response) Strong associations (high relative risks) and the presence of a dose response provide firmer evidence of causality.
question
Criteria of Judgment Surgeon General's Report, 1964 Sir Bradford Hill, 1965
answer
Temporal relationship between causal factor and disease occurrence Exposure to the causal factor must precede the onset of disease.
question
Temporality
answer
Cause----->Effect Exposure to the causal factor must precede the onset of disease.
question
Specificity of association Surgeon General's Report, 1964 Sir Bradford Hill, 1965 Harris, 2005
answer
The causal factor should be linked to one or more specific steps in the development of the target disease through molecular evidence
question
The adipose cells in obese individuals
answer
both greater in number and size than in lean individuals and the tissue contains a much higher percentage of macrophages
question
What happens to the stressed adipose cells as a result of storing excessive amounts of fat?
answer
they release inflammation including factors and undergo apoptosis
question
What do both outcomes of stressed adipose cells do?
answer
activate macrophages in a traditional M1 inflammatory state in which they release the tumor necrosis factor which recruits and activates additional immune cells to the site
question
What causes tissue to become resistant to insulin?
answer
low level sustained inflammation
question
What happens when the tissue become resistant to insulin?
answer
First step of developing diabetes
question
Tobacco Smoke Carcinogens
answer
Heterocyclic Amines Nitrosamines Polycyclic Aromatic Hydrocarbons
question
Coherence & Biological Plausibility
answer
The association should be plausible with known biologic facts about the pathophysiology of the disease, and the available evidence concerning the natural history, biology, and epidemiology of the disease should form a cohesive body
question
Experimental
answer
Randomized Clinical Trial Field & Community Trials
question
Observational
answer
Prospective Cohort Study Retrospective Cohort Study Retrospective Case Control Study Cross Sectional Study Ecologic Study Case Report
question
Experimental Study
answer
Investigator introduces or removes an exposure in order to observe its influence on a health outcome. This is the basis of a 'clinical trial".
question
Observational Study
answer
Investigator attempts to delineate associations between exposures and diseases without intervention
question
Phase III RCT Clinical Trials Human Experiments
answer
A randomized prospective study to determine the effectiveness of a preventive intervention on disease development or a therapeutic treatment on disease survival and prognosis
question
Clinical Trials: Phase 1
answer
Determine Toxicity of Dose
question
Clinical Trials: Phase 2
answer
Determine preliminary estimates of response, quickly quickly reject ineffective treatments
question
Clinical Trials: Phase 3
answer
Determine effect of new treatment relative to standard treatment
question
True or False: All clinical trials are experiments
answer
True
question
Flow Chart of Clinical Trials
answer
1) Potential new treatment 2) Determine maximum tolerable dose 3) Eliminate in effective treatments 4) Preliminary estimate of effect 5) Test effect against Gold Standard
question
Clinical Trial of Scurvy: James Lind, 1847
answer
Total of 12 patients treated, 2 Recovered, 10 Did Not
question
Double Blind
answer
Both investigator and subject are blinded to treatment
question
Single Blind
answer
Subjects are blinded to treatment
question
Considerations in Phase III Clinical Trials
answer
Duration of study Comparability (homogeneity) of patients Treatment allocation Controls Patient management and compliance Ethical considerations
question
Phase III Clinical Trial
answer
A comparative planned experiment involving randomization of human subjects to evaluate the relative merits of two or more treatments.
question
Example of Field Trial:Poliomyelitis Vaccine, 1954
answer
Approximately 750,000 children from 44 states, Canada, and Finland were registered for study and randomly assigned to receive either Salk Polio Vaccine or placebo Children were monitored during June through December, 1954 The vaccinated group had a polio rate of 28 per 100,000 The placebo group had a polio rate of 69 per 100,000 (Relative Risk = 2.46) A group refusing to participate had a polio rate of 46 per 100,000 (Relative Risk = 1.64)
question
Thailand Trial: HIV Vaccine
answer
16,396 individuals randomized to vaccine versus placebo arms Observed Rates after 3 years of followup Vaccine Group: p1 = 51 / 8,197 Placebo Group: p2 = 74 / 8,198 Relative Risk = p1 / p2 = 0.688 1 - RR = 1 - 0.688 = 0.312 = 31.2% (in the headlines!)
question
Case Report
answer
Report of unusual case (or a series of cases) documenting potentially diagnostic symptoms, time course, and treatment response. The patient's medical history, clinical signs and symptoms, laboratory results, and treatment response are typically included. Case reports may be helpful in the identification of new diseases and syndromes. Since case reports do not include comparisons with control subjects, their application in determination of disease etiology is limited
question
Case Reports of HIV/AIDS
answer
1980: First case of Kaposi Sarcoma (KS) reported in SF 1981: KS and Pneumocystis Pneumonia (PCP) cases reported in gay men in SF and NYC 1982: KS and PCP cases reported in Haitian Immigrants 1986: HIV isolated in Paris & Bethesda
question
Cross Sectional Survey (Prevalence Study)
answer
Study period Representative population sample Quantify population health characteristics Estimate prevalence Examples: National Household Surveys on Drug Abuse National Health Interview Surveys Behavioral Risk Factor Surveillance Survey
question
Ecologic Study
answer
Correlation between exposure rates and disease rates among different populations Exposure status of cases and healthy subjects (non-cases) is not known Correlation does not prove causation Hypothesis-generating study
question
Ecological Fallacy
answer
drawing cause and effect inferences from correlation analysis of observations of group data Since outcomes and exposures are measured on groups (populations) rather than individuals, conclusions regarding cause and effect may be spurious
question
Cohort Studies
answer
Investigations designed to estimate the relative risk
question
Case Control Studies
answer
Investigations designed to estimate the odds ratio
question
Cohort Study: Setup
answer
1. Large sample of healthy people 2. Conduct baseline interviews 3. Determine the risk factors 4. Longitudinal follow-up 5. Determine outcomes 6. Estimate the Relative Risk 7. Term coined by Wade H. Frost
question
Risk Factor Questionnaire
answer
Demography Medical History Physical Characteristics (BMI) Family History (Genetics) Nutrition Life Style (smoking, drinking, exercise) Reproduction History Occupation Environmental Exposures
question
Prospective Cohort Study: Features
answer
Direct estimates of exposure-specific incidence rates and relative risk (RR) Requires large samples Relatively expensive, time consuming, and labor intensive
question
Case Control Study: Setup
answer
1. Small Sample Size 2. Interview cases and controls 3. Determine the risk factors 4. Estimate the Odds Ratio
question
Retrospective Case Control Study
answer
Less expensive Smaller sample size Shorter duration No direct estimation of incidence rates or Relative Risk (RR) The RR is approximated by the Odds Ratio (OR) under certain conditions
question
Dose Response:Lung Cancer & Smoking
answer
Odds Ratios for the dose response are always calculated with numbers in the reference group
question
OR
answer
odds of risk factor for the cases/ odds of risk factor for the controls
question
Random Error
answer
Multi Directional error in measurement arising due to natural or unexplained and unpredictable forces
question
Systematic Error (Bias)
answer
Unidirectional error in measurement arising due to bias introduced by the investigator or the subjects under investigation.
question
Error in Epidemiologic Research
answer
"If you shut your door to all errors, truth will be shut out." Rabindranath Tagore, 1916
question
Confidence Intervals and P Values
answer
Traditional methods to address random error in research
question
Parameter
answer
error free numerical constant that describes a characteristic of a population, vis a vis., the "true value"
question
Estimate
answer
researcher's best guess of the value of a parameter based upon a sample of data. Estimates are subject to random variation (random error)
question
Point Estimate
answer
investigator's best guess of the value of a parameter based upon sample
question
Confidence Interval
answer
a range that expresses the margin of error about the point estimate
question
Normal Distribution
answer
u = mean and o = standard deviation u ± 1 = 68% u ± 2 = 95% ** u ± 3 = 99%
question
Symmetric 95% Confidence Interval
answer
Interpretation: There is a 95% probability that the calculated interval contains the true parameter.
question
Null Hypothesis ( HO )
answer
Unique hypothesis in which all effects are zero and there are no differences between treatments
question
Alternative Hypothesis ( HA )
answer
Hypothesis in which there are nonzero effects and there are differences between treatments
question
hypothesis testing
answer
we decide which hypothesis to reject by calculating the probability of the observed sample statistic under HO (called the P Value)
question
P Value
answer
the probability of observing the sample value (estimate) of the parameter of interest or a more extreme value under the null hypothesis. we pretend the null hypothesis is true, and calculate the probability of observing the sample value. If the P Value is small (say <5%), we reject the null hypothesis.
question
The Z Distribution
answer
u = 0 and =1
question
How do we get the P Value?
answer
estimated by calculating a sample Z value from the data and utilizing the standard normal distribution to obtain the P Value (Gerstman: Appendix 2) Z = sample estimate - value under HO standard error of estimate
question
Decision and P Value
answer
The P Value is the probability of a larger Z if the null hypothesis is true. If P < 0.05, reject the null hypothesis and conclude that the alternative hypothesis is more likely. Note that for P < 0.05, the sample Z ? 1.96.
question
Case Control Study: Setup
answer
1. Small Sample Size 2. Interview cases and controls 3. Determine the risk factors 4. Estimate the Odds Ratio
question
95 % CI for the RR (or OR)
answer
Convert RR estimate to natural log, ln (RR) Obtain SE of ln (RR) Use the standard formulas: 95 % CI = ln (RR) ± 1.96 x SE of ln RR Z = ln (RR) / SE of ln RR
question
Confidence Intervals & Hypothesis Testing
answer
95 % CI = estimate ± 1.96 x SE of estimate Interpretation: Probability is 95% that interval contains true parameter Z = (estimate - null value) / (SE of estimate) Interpretation: If sample Z exceeds 1.96, then P value < 5% and we reject the null hypothesis and conclude that the estimate is significantly different than the null value
question
Reproducibility
answer
Repeatability or Consistency of the test
question
Validity
answer
Precision or Accuracy of the test
question
Clinical Screening Tests
answer
Simple Rapid Inexpensive Safe Acceptable to the Public Accurate
question
Sensitivity
answer
proportion of subjects with disease who have a positive test result
question
Specificity
answer
proportion of subjects without disease who have a negative test
question
Positive Predictive Value
answer
Proportion with a positive test who have disease
question
Negative Predictive Value
answer
Proportion with a negative test who do not have disease.
question
Two Stage Screening Tests
answer
often used for early detection of disease
question
Stage I
answer
High Sensitivity to include all cases
question
Stage II
answer
High Specificity to exclude normals
question
As prevalence increases
answer
PVP increases and PVN decreases
question
PVP & PVN
answer
inversely related
question
Prevalence
answer
does not affect sensitivity & specificity
question
As specificity increases
answer
PVP increases and PVN decreases (for a second stage of testing with reduced sensitivity and increased prevalence)
