Overview of Epidemiological Study Design

Question 1

Definition of epidemiology

Answer 1

The study of the distribution and determinants of health-related states or events in specified populations and the application of this study to control health problems. The basic science underlying much of public health and preventive medicine.

Question 2

What are the 2 opportunities to account for bias?

Answer 2

Design phase, Analysis phase

Question 3

What are the 3 methods to account for confounding in the design phase?

Answer 3

Randomization, Restriction, Matching

Question 4

What are the 4 methods to account for confounding in the analysis phase?

Answer 4

Standardization, Stratification, Multivariable adjustment, Causal inference analysis

Question 5

What are the 4 basic types of study design?

Answer 5

Cross-sectional (prevalence) study, Case-control study (Retrospective, Prospective [(nested in a prospective cohort)], Cohort study (Mostly prospective), Randomized controlled trial (Prospective)

Question 6

How are incidence and prevalence measured in prospective vs. retrospective studies?

Answer 6

Prospective design is ideal for estimating incidence(# of newly diagnosed cases in population) of disease since new cases can only be clearly identified prospectively. Prospective and retrospective can be used to estimate prevalence (proportion of individuals in population with the disease) or association of outcome with a specific candidate of predictors.

Question 7

Name 3/6 advantages of cross-sectional studies

Answer 7

Relatively easy, quick and inexpensive
Often good first step for new study issue
Good design for hypothesis generation
Can estimate overall and specific disease prevalence and sometimes rates
Can estimate exposure proportions in the population
Can study multiple exposures or multiple outcomes or diseases

Question 8

Name 3/5 disadvantages of cross-sectional studies

Answer 8

Impractical for rare exposure or diseases
Not useful for establishing temporal relation
Confounding is difficult to control
Reverse causation

Question 9

What is the most used type of study design in population studies?

Answer 9

Case-control study

Question 10

How can power be increased in matching in case-control studies?

Answer 10

Using more than one control per case (eg 4:1)

Question 11

Name 3/4 advantages of case-control studies

Answer 11

Quick, Efficient
Relatively inexpensive
Best design for rare diseases
Can study multiple exposures at the same time

Question 12

Name 3/5 disadvantages of case-control studies

Answer 12

Only assess one disease outcome
Inefficient if exposure is rare
High potential for bias (selection and recall) and confounding–Difficult to choose appropriate controls (selection bias?)–Accurate exposure information? Recall bias?
Temporal sequence problems: reverse causation
Can’t calculate incidence, population relative risk or attributable risk

Question 13

What is a nested case-control study?

Answer 13

A case-control study situated within a prospective cohort study.
Whenever a case occurs, a control can be selected from among the cohort members who do not have the disease at that point in time (the risk set).
Risk set sampling of controls allows estimation of the incidence rate ratio.

Question 14

What are 3 advantages of nested case-control studies?

Answer 14

Remove recall bias because data collected before development of disease.
Allows for the time-frame element to be included in the case-control.
Often more cost-effective than a cohort.

Question 15

What is a case-cohort study?

Answer 15

The selection of a random sample of noncases from the parent cohort and all incident cases.
No matching is performed.

Question 16

What modification needs to be performed on the hazard ratio for a case-cohort study?

Answer 16

Cox proportional hazards regression models need to be weighted.

Question 17

What are 2 advantages of case-cohort studies?

Answer 17

More efficient for several disease outcomes using the same randomly chosen subcohort.
Reduces selection bias and information bias

Question 18

What measure of risk can be measured in a case-cohort study?

Answer 18

Subcohort can be used to calculate person-time risk and disease risk (risk ratio) can be measured.

Question 19

What is the order of events in a prospective cohort study?

Answer 19

Exposure/Study beginning –> disease occurrence

Question 20

What is the order of events in a retrospective cohort study?

Answer 20

Exposure–>Disease occurrence–>Study begins

Question 21

What is the best observational design to establish a temporal relation between exposure and outcome of interest?

Answer 21

Prospective cohort study

Question 22

What are 3/6 advantages of cohort studies?

Answer 22

Estimate of risk and rate of disease occurrence over time
Studying rare exposures
Multiple outcomes of a single exposure
Establish a temporal relationship between exposure and outcome and natural history of disease
Lower potential for bias than a case-control study –avoid recall and selection bias
Results are considered more conclusive than results from case-control studies
The longer a cohort study continues, the stronger it can become

Question 23

What are 3/7 disadvantages of cohort studies?

Answer 23

Very large sample sizes, especially for rare outcomes
Expensive and time-consuming
Larger sample size than case-control
Data collection is usually very expensive
Long time commitment for follow-up
Potential threats to internal validity
–Selection bias
–Attrition problem (Loss to follow-up)
–Differences in the quality of measurement of exposure or disease (information bias)
Can not infer causal relation

Question 24

What Does the Optimal Cohort Study Look Like? Name 4/8 elements

Answer 24

Large in scale, diverse (>500K)
Address multiple diseases/risk factors
Highly efficient recruitment, data collection, sample processing
Linked personal electronic records
High sample and data quality
State-of-the-art technology
Cost effective
Data available for qualified researchers

Question 25

Name Hill’s causal criteria (5)

Answer 25

Biological plausibility
Temporality
Specificity
Consistency
Strength
–Quantitative strength
–Biologic gradient (dose-response)

Question 26

What are 2 pros of intervention studies?

Answer 26

True control of exposure (experiment)→prevent confounding

If randomized, no selection bias

Question 27

What are 3/7 disadvantages of intervention studies?

Answer 27

Feasibility
Ethical consideration
\$\$$
Time
Adherence with regimen
Volunteer “bias”
Only one exposure