- 11 Downloads
A case-control study is a study in which subjects are selected based on their outcome status, such as with disease or disease-free. Investigators select cases (subjects with the outcome of interest) and controls (subjects without the outcome of interest) and then compare the exposure (or risk factor) status in the two groups.
Case-control studies are a very common observational study design within behavioral medicine research. Because the participants are selected based on their outcome status (commonly disease status), this study design is well suited for an outcome that is rare. For diseases with long latency periods (for example, melanoma or coronary heart disease), case-control studies can also be time efficient because the outcome has already occurred at the initiation of the study. When the exposure (or risk factor) is rare, a case-control study is often not practical.
Case-control studies determine the subjects’ exposure retrospectively, commonly through historical records or self-report conducted after the exposure has occurred. Limitations of using retrospective data contribute to results from case-control studies being considered weaker than results from experimental designs that examine similar associations. Recall bias can occur when case subjects remember exposure differentially compared to controls. For example, a mother whose infant was born with a birth defect may differentially recall her use of medication during pregnancy compared to a mother of an infant without a birth defect (Rockenbauer et al. 2001). The use of retrospective data, however, may facilitate study approval by ethical review boards, particularly, when the risk factor is illegal or known to be harmful, such as illicit drug use or tobacco use.
The selection of control subjects is critical in the design of a case-control study. Subjects chosen as controls should be as similar as possible to the case subjects except, potentially, with respect to the exposure. Specifically, cases and controls should have had equal chance to be exposed to the risk factor. For this reason, cases and controls are often matched with respect to age, gender, ethnicity, and other factors.
In many case-control studies, the groups are compared by evaluating the odds ratio which is defined as the odds of exposure among the cases divided by the odds of exposure among the controls. In general, investigators cannot determine incidence rates of the disease since the subjects are selected based on disease (outcome) status. Thus, computing a relative risk directly is not possible. However, the relative risk can be approximated by the odds ratio when the outcome of interest is relatively rare.
In a typical behavioral medicine case-control example, Brent et al. (1993) investigated the association between adolescent suicide and multiple psychiatric risk factors. Sixty-seven adolescent suicide victims (cases) were matched to 67 controls with respect to age, gender, socioeconomic status, and county of residence. Investigators obtained information about the suicide victims through a “psychological autopsy protocol” in which parents, siblings, and friends were interviewed concerning the victim’s risk factors. The controls’ risk factor information was obtained from the participant and at least one parent. The study found that the suicide victims had significantly higher odds of major depression and substance abuse compared to the controls.
Usually less expensive and less time-consuming than cohort designs or experimental designs.
Often used when the outcome of interest is rare or has a long latency period.
Not practical when exposure is rare.
Sample size requirement is usually smaller than cohort or experimental designs.
Often used in initial investigations of an association due to logistical ease and relative lower cost.
May be used when exposure of participants to risk factor would be considered unethical in an experimental design.
Appropriate when studying multiple risk factors.
Usually can only address a single outcome (disease).
Susceptible to recall bias (since exposure and outcome are determined retrospectively) and selection bias (which can occur when the controls are selected in such a way that they did not have same risk of exposure as the cases).
Often considered weaker study design compared to cohort studies or randomized trials that study analogous associations.