EFB 360 Exam 3

Descriptive

Epi

Explore and describe patterns of disease• Results used to generate hypotheses about causes,transmission, prevention

Analytic epi

Testing hypotheses derived from descriptive work• To understand biology,natural history by explaining the etiology (risk factors,modes of transmission, etc.

four major components of epi hypothesis

Specify subject group(s) (experimental groups)- Control/comparison group(s)- Treatment(s) or exposure(s) of interest- Outcome(s) of interest

Unit of observation

level at which data are recorded

Individual-level data

unit of observation is the person

Aggregate-level

unit of observation is a group

Analytic epidemiologic studies

-Ecologic, cross-sectional, case-control, cohort, experimental studies

-Evidence of causal associations

purpose of analytic epi studies

-test hypothesis

-identify associations between exp and disease

-measure strength of association

-test treatment and intervention

Natural Experiment

a natural event that acts as an experimental treatment in an ecosystem

Single Blinded

Patients do not know whether they are getting the drug or a placebo

Double Blinded

neither the participants or the researchers involved should know which treatment the patients are actually getting

Triple Blinded

treatment and research approaches kept a secret from the subjects and investigators, but analyses are completed with removal of investigator

What is Phase 1 of a clinical trial?

The smallest group with healthy people, evaluating the safety of a test.

What is Phase 2 of a clinical trial?

A slightly larger group that assesses tolerability, dosage, and side effects.

What is Phase 3 of a clinical trial?

A much bigger group that involves random assignment and evaluates the efficacy of a new treatment.

Ecologic comparison/correlation studies

Assess correlation between exposure and disease rates/prevalences among several(usu. >10) groups

Ecologic trend studies

Assess correlation between changes in exposure and disease levels among pops. over time

ecologic studies general make use of what type of data

secondary

advantages of ecologic studies

-may provide information about the context of health

-can be performed when individual-level measurements are not available

-can be conducted rapidly and with minimal resources

ecologic studies disadvantages

Ecological fallacy -false belief that exposure and disease are causally related based on apparent association observed at group level

ex of ecologic fallacy

-William Farr's analysis of Cholera mortality and elevation in london

-ice cream leads to shark attacks

cross sectional studies

Exposure and disease for individuals both measured at same point in time (prevalence)•

Typically descriptive in nature•

provide estimates of the magnitude of health problems

cross sectional studies advantages

efficient, can study everyone at once, less time consuming

cross sectional studies disadvantages

No cause-effect interpretation possible, (but can help identify likely causal factors)

Susceptible to bias

Can underestimate true prevalence (non-respondent bias)

Analytic Observation Studies

case control and cohorts

measures of association

risk ratio, rate ratio, odds ratio

Cohort vs. Case-Control ex lung cancer

Cohort- Investigator identifies closed population (cohort)- Classifies individuals as smokers (E+) or non-smokers (E-)- Compares incidence of lung cancer in E+ and E-

Case-control- Investigator identifies [open or closed] population- Accrues cases or samples cases (D+)- Samples non-cases from population (D-)- Compares proportion of smokers in D+ and D-

RR versus OR

RR=Cohort OR=Case control

Probability (risk)

Probability is the proportion(ie, percentage) of times an event would occur if an observation were repeated many times.

Odds

the probability of an event happening divided by the probability of the event not happening

Interpretation of Risk Difference

RD = 0 (no association)-

RD > 0 (positive association)-

RD < 0 (negative association)

the etiologic fraction (EF)

measures the proportion of disease in the exposed group that is due to the exposure

Why is EF referred to as the preventable fraction

represents the proportion/percentage of the disease among the exposed group that could be prevented if that exposure could be eliminated

PEF (population etiologic fraction)

extrapolates the EF to the larger target population

systematic error

Error that shifts all measurements in a standardized way. Decreases accuracy. Can result in bias

random error

an error that occurs when the selected sample is an imperfect representation of the overall population

Parameter

Error-free- Hypothetical- Notation: no hat- Population- Causal effect

Estimate

Error-prone• Calculated• Notation: hats• Sample• Association

clinical significance vs statistical significance

Relatively large, but not statistically significant, differences between study groups (esp. small groups) may be of clinical significance

Relatively small, but statistically significant, differences may be of no clinical significance

Sequence of Studies in Human Populations

Clinical Observations --> Analyzing Surveillance Data-->Case-control studies-->Ecologic studies-->Cross-sectional Studies-->Cohort Studies-->Randomized Trials

Types of Associations

True (real) or false (spurious) associations

Interpreting false associations

If observed association is due to bias (systematic error) in subject selection or information gathering

Interpreting real associations

If real, is it causal or due to confounding?

3 types of bias

Selection bias• Information bias• Error due to confounding

Selection Bias

Due to improper manner individuals are selected for study ex Hospital admission rate bias

Information bias

systematic errors in data collection

what is a loaded question

logical fallacy that occurs when statements or questions are designed to confuse listeners into tacitly accepting something as true

Bradford Hill's Criteria

strength of association, consistency, specificity, temporality, biological gradient, plausibility, coherence, experimentation, analogy

Streght

Relative risk as a measure of "strength" Stronger RRs carry more weight• do not dismiss small RRs• some causal relations will by nature have a weak association• e.g., smoking and cardiovascular disease

consistency

Similar findings- Using diverse methods- In different populations- Under a variety of circumstances• Consistency alone does not prove causality

Specificiy

Factor leads to a specific disease- E.g., asbestos and asbestosis; Requires knowledge at cellular level• Converse is NOT true

Temporality

Exposure must precede disease by reasonable amount of time

Biological Gradient

also known as a dose-response curve; shows a linear trend in the association between exposure and disease

Plausability

makes sense in face of known biological and and other facts

Coherence

Do the facts cohere (i.e., to stick together)?

Experimentation

Experimental evidence should support observational studies

Analolgy

Similarities among things that are otherwise different

Risk=

probability + magnitude

Phase 4

Thousands of participants, happens after FDA approval