de Bock et al

Reviewer: Roisin Pill

1.This is an important subject given the interest in the new genetics and the possible implications for the GP workload and the "geneticisation" of care for patients whose family history may indicate increased risk of developing a particular condition. I am not aware of any other published study that describes the subsequent behaviour of women with a positive family history who have consulted their GPs specifically about the risk of developing breast cancer . (Previous work has focused on reactions to and action taken after counselling in genetic clinics by geneticists and /or trained counsellors.) For that reason I would like to see it published although I do have very distinct reservations about the structure and points of presentation.

2.For example, the authors define their aim as measuring the extent of patient compliance with the GP advice, a very traditional theme in medical literature. It is interesting to note that the failure of the GPs to follow the policy laid down by the genetic centres , by contrast, is described in neutral terms( e.g. the relation between GPs genetic advice and risk assessment ), and not apparently thought worthy of further investigation. In the discussion the discrepancy is then described as being in line with recently developed GP guidelines ( developed by the authors themselves.)

I would argue that the study of professional non-compliance is just as relevant here . We do not know how many Gps in the primary health care centre were involved in giving advice , whether they were responsible for passing on the risk assessment to the woman at the same consultation in which the advice was given or how many GPs were discordant and in which direction. Interviews with the latter group to explore their reasons for non-compliance might have shed light on the tensions experienced by GPs asked to apply protocols developed from population statistics to the management of the individual in front of them .

3.Indeed, the value of this paper lies in the questions it raises in the mind of the reader about the content and quality of the advice given originally and the dilemmas inevitably faced by GPS in a field where the evidence –base is changing so rapidly. This is acknowledged implicitly by the authors in their final paragraphs setting out the current evidence for effectiveness and the ambiguity about any actual benefit from these preventive strategies, and explicitly in their conclusion.

Having set out to study compliance with certain behaviours considered relevant in 1995 we are left with the message that , in this situation, it probably does not matter. It might be more helpful and give a more positive message to clinicians , teachers etc if the authors were to indicate what advice they would recommend now ,given the latest findings. This would also have the advantage of highlighting the longitudinal aspect of the study and the impact of the passage of time which is not acknowledged at all.

4.As to scientific reliability the question is clear, the overall design adequate and the participants adequately described.

However, the non-respondents are not dealt with adequately. We are told nothing about the 14 who consulted about their family history but did not consent to the study. Even more important, there were 18 still in the practice when the records who were re-analysed who were not interviewed. Their RRs and advice category can be deduced from Table 1 but there is no explanation for their omission or discussion of how that might influence the pattern of reasons.( Indeed we are given no actual figures or breakdown of the reasons given or the way that data was collected in the interview.)

An annotated figure would make it easier to follow the stages of the study and grasp the various bases being used. These are rather complicated to follow as we shift from those whose risk was assessed to all those given advice ,and then to the advice subgroups ,and, within those , to the respondents actually interviewed.

Here the abstract is particularly confusing as it is not at all clear why the bases should be different. Moreover, the conclusion about the value of the advice seems odd in the light of the stated aim to study compliance.

Some of these problems could be dealt with if the research was set in its historical context and the authors reflected on professional as well as patient compliance and distinguished their original assumptions from their current message to their GP colleagues.