To challenge Hilary Thomas's article (page 1, Managers & Medicine, 15 July) the evidence base on Taxol as initial treatment for ovarian cancer has been accumulating since 1993 and is given in detail in the June issue of Effective Health Care. The only peer-reviewed study published, which showed better survival with Taxol, was in early 1996.

Publication of confirmatory evidence presented in 1997 and 1998 is pending. The largest trial (ICON-3), whose initial analysis fails to show any apparent benefit for Taxol, is unlikely to contribute definitively to the evidence base before late 2000.

Meanwhile, the statement that 'findings from the group of patients comparable with the previous two randomised trials - those with advanced stage and poor prognosis - are consistent with those of the first two trials' is misleading. There was no difference in survival or progression-free survival between women receiving Taxol and those not, in any stage or prognosis- related sub-group, at a verbal presentation of the ICON-3 data in Edinburgh in July.

At what point should the National Institute for Clinical Excellence have produced Taxol guidelines? Review of the evidence base on Taxol biennially might have been appropriate. Even if NICE guidance changes to accommodate emerging evidence, what impact will this have on clinical practice? Effective Health Care notes that 'current evidence suggests chemotherapy for ovarian cancer should be paclitaxel (Taxol)/ carboplatin'.

As the evidence base looks weaker for Taxol than it did a year ago, the 80 per cent of health authorities in the BACUP survey which Hilary Thomas cites - and which did not fund Taxol in August 1998 - should be congratulated on their foresight.

Dr Maggie Harding

Locum consultant in public health medicine

Camden and Islington health authority