To the Editor.—
The article entitled "A Model for Epidemiologic Research" by Doege and Gelfand (239:328, 1978) unfortunately tends to perpetuate a misconception about retrospective or case-control studies.The authors state, "For example, the frequently encountered case or case-control method has the important advantage of using limited but already available (retrospective) data that conserve the investigator's resources relatively well, yet carry the hazards of various known and unknown biases." This statement tends to confuse the time dimension with the strategy. Many, if not most, analytic case-control studies use data generated de novo rather than already extant. We may decide to interview patients with newly diagnosed cases of a disease as they come to a clinic for a stated period of time in the future (some may not even exist at the inception of the study). Data concerning these cases are compared with similar data from concurrently selected controls. This is
Gullen WH. Epidemiologic Research. JAMA. 1978;239(25):2655. doi:10.1001/jama.1978.03280520027003
Customize your JAMA Network experience by selecting one or more topics from the list below.