Johnson provides an overview of many of the limitations of guidelines and how they may be affected detrimentally by poor quality or absence of evidence, the wishes of the guideline authors and politics1.
Some of the examples given deal specifically with ovarian cancer and reflect our own recent experiences. The Scottish Intercollegiate Guidelines Network is currently developing guidelines for the management of epithelial ovarian cancer. Some questions about surgical management of this disease cannot be answered due to a lack of good quality evidence; certainly, there is only one randomised controlled trial that shows surgery to be of benefit in ovarian cancer2.
When evidence does exist, however, it should be included if it fulfils pre-defined criteria, in our case, the SIGN 50 document3. For example, evidence that shows a 25% reduction in the risk of death after surgery for Stage III ovarian cancer by a gynaecological oncologist compared to a general gynaecologist or a general surgeon, should be included in the guideline4. There is, however, potential for the evidence to be flawed by confounding, as is usual in many non-randomised trials.
Limitations of guidelines are not because of evidence alone. Diagnostic criteria can be problematic and Johnson points to the potential problems that surgico-pathological FIGO staging can cause in developing guidelines in ovarian cancer surgery. We have found most discussions revolve not around the necessity of removing the cervix in staging, but rather, the removal of retroperitoneal lymph nodes to stage adequately disease apparently confined to the ovary. FIGO staging is internationally recognised, and guidelines should follow this system. Further therapy depends on the stage of the disease, and positive node status affects this. Does this mean all women with apparent Stage I disease should have full retroperitoneal lymphadenectomy? We will find no evidence to answer this question.
We hope that despite these possible limitations, a finished guideline will be a comprehensive body of evidence, in the main reflecting clinicians' current working practices. This will benefit all clinicians dealing with the disease, not only those who need guidance.