Selection of cases and controls: Additional information is needed

The potential for selection bias is a concern in any case–control study. Cases and controls should be representative of the respective diseased and nondiseased persons in a well delineated target population. Even with representative cases and controls, if there is a high (or differential) nonparticipation rate among (or between) cases and controls, selection bias may be introduced. Case–control studies offer an efficient observational analytic design, but they also carry a high potential for the introduction of bias.