Safety of hormone replacement therapy after mastectomy

In reply: Eden, Pyke and Hitchins are happy with the findings of the rapid literature search that we performed for the general practitioner who wanted to know about the safety of HRT in a woman treated for breast cancer. However, they have concerns with the process. Eden is worried that we missed two Australian publications on the issue.

In fact, they misunderstand our intentions. Firstly, we did not raise the issue, nor deliver an "advisory statement". It was the general practitioner who asked the question. We were trying to provide a rapid (few days) and responsive service to provide credible information to help a doctor manage a patient. Secondly, we were not able to undertake a full systematic review (which would take a full-time researcher as long as six months and would cost accordingly).1 Instead, our best strategy was to use a cascade process of searching, looking first for the most rigorous study types that would answer the clinical question. If not available we go to the next most rigorous, and so on, stopping when we find the relevant evidence.2 For this question, the ideal study type would be a meta-analysis of randomised controlled trials (RCTs). However, no RCTs were available (as Eden, and the guidelines to which Pyke refers, note), and we had to content ourselves with observational studies. We should remember that no evidence for safety is not the same as evidence for danger. What is important is that we did not miss any trials.

Legal issues worry many doctors, even when decisions are supported by best research evidence,3 but, rather than pose a medicolegal threat, we believe that this evidence-based approach is more likely to protect doctors. Why? Failures in communication are the most common preventable cause for doctors being sued by patients.4 Yet, taking the trouble to find empirical information such as this and then discussing it with the patient is surely the most effective way of communicating the pros and cons of different treatment strategies (including, we agree, alternatives such as those mentioned by Hitchins). In the end the patient has to decide on the basis of the risks and benefits, and the choice can often be extremely difficult. It is likely to be more dangerous to assume the patient has abdicated this responsibility to the doctor without checking first. Can doctors be sued for a "safe" decision that leaves a patient exposed to unnecessary symptoms? It may be dangerous to assume that doctors can play "safe" in any one direction.

Why do experts take exception when non-experts delve in their areas for the best evidence to manage patients? After all, a cat may look at a king.5 Experts seem to welcome the attention, but seem to think they should be dispensing the information. However, until the information can be delivered more effectively, this sort of stopgap system will have to do.