Epidemiologic study designs
1 / 46

Epidemiologic Study Designs - PowerPoint PPT Presentation

  • Updated On :

Epidemiologic Study Designs. Nancy D. Barker, MS. Exposure Alcohol consumption Raw hamburger Smoking. Health Outcome Breast Cancer E. Coli Lung Cancer. Epidemiologic Study Design The plan of an empirical investigation to assess an E – D relationship. Type of Epidemiologic Studies.

I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
Download Presentation

PowerPoint Slideshow about 'Epidemiologic Study Designs' - mandel

An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.

- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
Epidemiologic study designs l.jpg

Epidemiologic Study Designs

Nancy D. Barker, MS

Epidemiologic study design the plan of an empirical investigation to assess an e d relationship l.jpg


Alcohol consumption

Raw hamburger


Health Outcome

Breast Cancer

E. Coli

Lung Cancer

Epidemiologic Study Design The plan of an empirical investigation to assess an E – D relationship.

Type of epidemiologic studies l.jpg
Type of Epidemiologic Studies

  • Experimental: Randomization to exposure

  • Observational: No randomization

Experimental studies l.jpg
Experimental Studies

The investigator through randomization allocates subjects to different categories of exposure.

Randomization: An allocation procedure that assigns subjects into (one of the) exposure groups being compared so that each subject has the same probability of being in one group as in any other.

Randomization l.jpg

  • Tends to make demographic, behavioral, genetic, and other characteristics of the comparison groups similar except for their exposure status.

  • If the study finds any difference in health outcome between the comparison groups, that difference can only be attributable to their difference in exposure status.

Experimental studies7 l.jpg
Experimental Studies


  • Randomization – unmeasured variables evenly distributed among exposure

  • Potential for bias is low


  • Ethical concerns

  • Cost

  • Length of study

  • Not good for rare D

Observational studies l.jpg
Observational Studies

  • Investigator observes the exposure and outcome status of each

  • Most Epidemiologic studies are observational

Observational studies9 l.jpg
Observational Studies

  • Descriptive Studies

  • Analytic Studies

Observational studies10 l.jpg
Observational Studies

Descriptive Studies

To organize and summarize data according to time, place, and person. Why?

  • Describe natural history of disease

  • Extent of public health problem

  • Identify populations at greatest risk

  • Allocation of health care resources

  • Suggest hypothesis about causation

Observational studies11 l.jpg
Observational Studies

Analytic Studies

Used to quantify the association between an exposure (E) and a health outcome (D), and to test hypotheses about causal relationships.

  • Provides a control group (baseline)

  • Test hypotheses about determinants

  • Causation

Causation l.jpg

In any research field involving the conduct of scientific investigations and the analysis of data derived from such investigations to test etiologic hypotheses, the assessment of causality is a complicated issue.

In particular, the ability to make causal inferences in the health sciences typically depends on synthesizing results from several studies, both epidemiologic and non-epidemiologic (e.g., laboratory or clinical findings).

Observational studies13 l.jpg
Observational Studies


  • Natural Setting

  • Addresses ethical concerns


  • No randomization – investigator can only account for variables measured.

Common observational study designs l.jpg
Common Observational Study Designs

  • Cohort Study

  • Case-Control Study

  • Cross-sectional Study

Design options l.jpg
Design Options

  • Directionality

  • Timing

Directionality l.jpg

  • When the exposure variable is observed relative in time to when the health outcome is observed.

Timing l.jpg

  • Timing concerns whether the health outcome of interest, and therefore all study events, has already occurred before the study actually began.

Cohort study l.jpg
Cohort Study

  • Directionality: Always forward

  • Timing: Prospective or Retrospective

Cohort study27 l.jpg
Cohort Study


  • Follow-up study

  • Longitudinal study

  • Incidence study

Cohort study example l.jpg
Cohort Study Example



Framingham Heart Study (ActivEpi)


VDTs and spontaneous abortions (ActivEpi)

The exposure variable in cohort studies l.jpg
The Exposure Variable in Cohort Studies

  • If E is common, sample from large population, then divide into E and not E.

  • If E is rare, sample E from special population, then sample not E from external comparison group.

  • Caution: Healthy worker effect.

Cohort study30 l.jpg
Cohort Study

  • Advantages

    • Forward directionality

    • Free of certain selection biases

    • Prospective cohort less prone to obtaining incorrect information

    • Can study several diseases

    • Useful for examining rare exposures

    • Retrospective cohort study is low-cost and quick, e.g., occupational studies

Cohort study31 l.jpg
Cohort Study

  • Disadvantages

    • Prospective cohort study is costly and time consuming

    • Loss of subjects from migration, lack of participation, withdrawal and death leads to bias

    • Statistically and practically inefficient for rare diseases with long latency

    • The exposed may be followed more closely than the unexposed

Example l.jpg

Measure of Effect

  • Risk Ratio (Cumulative incidence – risk)

  • Rate Ratio (Incidence density – rate)

Case control study l.jpg
Case-Control Study

  • Directionality: Always backwards

  • Timing: Always Retrospective

Case control study34 l.jpg
Case-Control Study


  • Case-control studies of aspirin and Reye's syndrome (ActivEpi)

  • Case-control study of animal food products and Creutzfeldt-Jakob disease (ActivEpi)

Case control study35 l.jpg
Case-Control Study


  • Cheaper, quicker

  • Sufficient number of cases

  • Smaller sample size

  • A variety of exposures

Case control study36 l.jpg
Case-Control Study


  • Do not allow several diseases

  • Risk of disease cannot be estimated directly

  • Selection bias

  • Information bias

  • Not for rare exposures

Case control study37 l.jpg
Case-Control Study

Selection of Controls

  • Representative of the source population from which the cases derived

  • Population-based controls

    -Cases and controls come from the same source population

  • Hospital-based controls

    -Easily accessible, tend to be cooperative, and are inexpensive

    -Not usually representative of the source population and may represent an illness caused by the exposure

Case control study38 l.jpg
Case-Control Study

Measure of Effect

  • Odds Ratio

Cross sectional study l.jpg
Cross-Sectional Study

  • Directionality: Always Non-directional

  • Timing: Always Retrospective

Cross sectional study40 l.jpg
Cross-Sectional Study


  • Convenient and inexpensive

  • Can consider several exposures and several diseases

  • Can generate hypotheses

  • Usually represents the general population

Cross sectional study41 l.jpg
Cross-Sectional Study


  • Cannot establish whether the exposure preceded disease or disease influenced exposure

  • Can identify only prevalent cases rather than incident cases

  • Possible bias since only survivors are available for study

  • May under-represent diseases with short duration

Cross sectional study42 l.jpg
Cross-Sectional Study


Cross-Sectional Study of Smoking and Peripheral Vascular Disease (ActivEpi)

Cross sectional study43 l.jpg
Cross-Sectional Study

Measure of Effect

  • Prevalence ratio (PR)

  • Prevalence odds ratio (POR)

Hybrid designs l.jpg
Hybrid Designs

Combine the elements of at least two basic designs, or extend the strategy of one basic design through repetition.

  • The Case-Cohort design

  • The Nested Case-Control Study

Incomplete designs l.jpg

Studies in which information is missing on one or more relevant factors.

  • Ecologic Studies

  • Proportional Studies

References l.jpg

  • ActivEpi Companion Textbook; Kleinbaum, Sullivan, Barker

  • Principles of Epidemiology; CDC Self-Study Course

  • Epidemiology in Medicine; Hennekens, Buring

  • Statistics in Public Health: Quantitative Approaches to Public Health Problems; Stroup, Teutsch

  • Case-Control Studies; Schlesselman