Reporting participation in case-control studies.

Abstract

Although low participation in case-control studies may threaten their validity, it is difficult to evaluate the potential problem of nonresponse across studies because the level of participation among subjects is not reported in a consistent manner. This commentary presents a proposal for standard reporting of the process by which control subjects are recruited to a study and the outcomes of recruitment attempts. We encourage authors to provide a description of the sampling frame, the method of selecting units from the frame, and the number of units selected. For each unit selected, an outcome should be classified into four broad categories: ineligible sampling unit, unable to determine eligibility, respondent not eligible, and respondent screened and eligible. Details should be provided in each category; these details would differ according to the method used to contact potential respondents (by random-digit dialing, by telephone from a list, in person, or by mail). The proportion who participated can be reported as the investigator sees fit, as long as reporting of the components is complete. The sampling frame and outcomes should be reported separately for various parts of the study, such as controls from different sources or those who gave biological specimens. Standardized reporting will allow an evaluation of the success of various methods in recruiting potential participants, and will contribute important information for assessing the validity of individual studies.