Prolonged study leave for general practitioners

Abstract

of sexually transmitted diseases such as HIV infection. We also acknowledged the problems resulting from non-responders and recognised that a group of young and mobile men may have been differentially missed from our sample. In our discussion we explained that selection or underreporting might mean that the true figure for homosexual intercourse was nearer 3%. There is, however, no evidence to suggest our figures are underestimated by any more than this. Whether or not the use of female, rather than male, interviewers distorted the responses about homosexuality is impossible to assess. Our impression is that the use of trained, professional interviewers who can establish a rapport with their interviewees is a far more important factor than gender per se. When recruiting interviewers for this and other studies we have only ever had one male applicant and he did not meet our selection criteria.There are two more general points that should be emphasised. Firstly, although our respondents were selected as controls for a study of testicular cancer this did not necessarily lead to a biased sample except in so far as men not registered with a general practitioner were not sampled. In the United Kingdom, in the absence of population registers, a selection of a true random sample of the population is, in practical terms, impossible. The electoral register provides one possible means of selecting a sample but suffers from the same sort of disadvantages as using general practitioners' lists-that is, that mobile members of the population and those with no fixed address would be missed. Moreover, because dates of birth are not given the usefulness of this register is limited. The fact that only men aged 20-55 were included in our sample does not lead to bias, although it does mean that our results are generalisable to only this age group. Secondly, our response rate was over 83% and we do not know of any random survey of young and middle aged men in the United Kingdom that has achieved a higher figure. Although our study does have limitations as clearly set out in our paper it does provide baseline information that is otherwise unavailable. It is interesting that Dr Mills does not cite other comparable population based figures which he would consider to be more accurate. As far as we are aware, there are none.