Inappropriate selection of cases and controls biased study

EDITOR—In their case-control study, Churchill et al found that teenagers who become pregnant have higher consultation rates at general practices than their age matched peers.1 However, their conclusions may be biased by inappropriate selection of cases and controls.

Churchill et al defined cases as all registered patients in the practices who had a recorded termination, delivery, or miscarriage resulting from conception. Since they ascertained cases from practice computer records, maternity books, and knowledge of practice staff, the cases included only patients who received such services locally and not from other districts. Controls were, however, selected from the ordered list for the practice closest in chronological age as the cases. Hence the control group included registered patients who were actually resident in other districts. The inclusion of such patients would seriously underestimate the consultation rates in the control group.

The median age of the subjects was 17 years. A significant proportion of the control subjects were in further education and well over a third of the over 18s were in higher education.2 Hence the selection bias is likely to be considerable and could account for the findings in the study.

Not all teenagers are sexually active

EDITOR—Churchill et al in their study dispel the myth that teenagers do not consult their general practitioners about contraception.1 They also conclude, however, that teenagers who become pregnant have higher consultation rates than their age matched peers, most of the difference owing to consultation for contraception.

We do not believe that this claim is justified, as the potential confounder of sexual activity was not controlled for. While the case group (pregnant teenagers) must be sexually active, the same cannot be said of the age matched controls. It …