Why costs of consultations in general practice vary.

findings indicate that the management ofpatients found to have a dyskaryotic smear does not protect completely against subsequent invasive cancer. Three of the five cases of invasive cancer were in women aged under 35. Over the past decade mortality and incidence rates for cancer of the cervix have increased in younger women, although the rates in those under 25 are still very low compared with those in older women.2' Robertson et al described 10 cases of invasive cancer of the cervix occurring in their study of 1781 women identified with a dyskaryotic smear over 20 years.12 It was not clear, however, whether those cases represented a significant excess or the number to be expected on the basis of the background rate or, indeed, a deficit as no statistical criteria were applied. Our criteria for reversion to normal were more rigorous than those reported by some workers. We checked on the subsequent outcome of these patients after a further two years. Fifty five of the 157 patients classified as having reverted to normal (35%) had a record of a subsequent smear. In 53 out of these 55 women this smear was reported as negative after a mean of 17 months (range 1-47 months). Twenty one of these 53 patients had two or more negative smears subsequently. This suggests that when several adequate repeat smears remain negative over more than 18 months regression may be real rather than apparent. Almost two thirds of women, however, had no further smears, and there may have been a selection bias in the 35% who continued to have smears despite being discharged. The poor reliability of a single negative smear to suggest apparent regression is shown by the finding that 22% of these women (83/37 1) had a follow up smear showing dyskaryosis. When the negative smear was also inadequate this figure increased to 40% (22/55). Our study clearly indicates the importance of following up women with dyskaryotic smears and especially those with a smear showing moderate dyskaryosis or two smears showing mild dyskaryosis. We do not wish to make recommendations about the exact details of how this follow up should be conducted in the current state of knowledge. There is no experimental evidence that prognosis is improved by colposcopic compared with cytological surveillance. We endorse the inter-collegiate working party's proposal for a randomised controlled trial to examine costs and benefits of these two alternatives. We …