Editor,—Fether et al present a very interesting case-control study on STIs in women who have sex with women (WSW).¹ This was not a community based sample and thus prone to selection bias. In order to appreciate the results in full it would help to know how cases and controls were identified and how controls were selected. Bisexual or homosexual orientation may be difficult to disclose even in a sympathetic and non-judgmental setting. Studies using self reported sexual orientation to determine case or control status will always have a degree of differential misclassification. It is likely that WSW who volunteer this information differ not only from women who do not have sex with women but also from WSW who do not volunteer the information but admit it when prompted, and from those who do not admit it even when prompted. Without this information it is difficult to determine the importance of various prevalences quoted in the paper. All I learn from this paper at present is that women who have sex with women also take other risks.