With any epidemiologic study, things can go wrong. With case control studies, the problems to look out for are: a. Is the sample large enough for the findings to be imputed to a larger population?
b. Is the control group (or control groups if more than one is used) suitable? Control groups need to be from the same sample population as the cases, and for this reason, control groups of hospital patients at major referral hospitals are suspect.
c. Are the questions asked appropriate for the study (i.e., questions that people can give reliable answers to), and are they asked in the same way for both the cases and the controls?