of the Journal about audit in gynaecology and the difficulty of data collection and the organisation of gynaecological audit meetings.I have just been attempting to audit the medical management of endometriosis and fully support his views. The most difficult part of this audit was the data collection. The regional computer centre was contacted by our audit co-ordinator and the data provided were analyscd based upon the names of women who were diagnosed to have endometriosis within the preceding year. The data suggested there were 87 patients; 25 had endometriosis hut of the rest, 25 had adenomyosis, 6 had endometritis and 9 women had pelvic pain with no pathology found.I attempted to use the drug prescriptions to obtain a list of women who had used drugs for symptomatic treatment of endometriosis. However, the pharmacy did not have a computerised record of women who were prescribed danazol, provera, or the GnRh analogues. The only way to collect the data was to go manually through all the prescriptions written for outpatients throughout the hospital for twelve months in order to collect the names of the women who were prescribed any of the above drugs. This took two doctors more than ten hours.This hospital uses a system of remote patient activity, which is located and managed by regional headquarters. At present it is only coding inpatient activity, such as gynaecological operations. Diagnostic codes are added by the coders at the regional health authority (Bunch 1992). The result is inadequate and inappropriate data retrieval for any clinical purpose such as clinical audit meetings. This audit did, however, help to uncover the insufficient and poor nature of the data storage on certain topics of gynaecology such as endometriosis and pelvic inflammatory disease.It may be possible to collect accurate clinical data without spending a lot of money provided managers can be persuaded that the primary function of a clinical information system is to collect clinical data, to improve the clinical services, and to assist with clinical audit (Smith 1992).Arrangements are now being made for improvements in the coding system for the regional computer centre and the pharmacy has now agreed to keep a computerised record of women who have been prescribed danazol or one of the GnRh analogues for endometriosis. I wonder if any one else has had a similar or better experience of audit in gynaecology on such conditions in which diagnosis, findings and treatment are so varied and in which poorly kept records are frequently encountered. Usha Mantha Immunohistochemistry in pelvic nodes of patients with cervical cancerSir, We were interested to read the short communication from Gitsch et al. (1992) in which they were unable to identify any micrometastases by immunohistochemistry in the lymph nodes of 10 patients with early stage cervical cancer who later developed recurrence. We performed a similar study on three patients with early (<1 year) recurrence after radical hysterectomy for stage 1B and 2A cervical cancer in which the node...

No abstract