Reply to the Letter from A Clarke

We would like to thank Angus Clarke for his insightful comments on our paper1 but disagree with some of the conclusions he draws.

In cases such as that of Jim and Mary, we agree that the standard policy is to seek access to the medical records of the affected person. Indeed, such consent was sought when the daughter was seen in the department (space constrictions in the article prevented us from describing all the details of the cases). However, despite the best efforts of the clinicians involved, no consent was ever obtained, and so the dilemma remained. Were we to leave the daughter in the dark or go against Jim's expressed wishes? The clinicians felt uneasy about telling the daughter that they could not give her specific risk information without further details about Jim, when in fact they had these details readily available. The clinicians were aware that the daughter was at a quite different risk from the one she thought she was at, but felt unable to tell her without his consent. Our conclusion is that while such cases will be rare and clinical staff can, and should, try to facilitate communication between family members, in some families this will not be successful. It is therefore important to debate how such cases ought to be dealt with when they do arise.

We also agree with Clarke that much genetic testing is currently influenced by resource allocation issues. However, contrary to what he suggests, in our department and in others, we do, in fact, offer genetic testing to unaffected individuals if a search for the particular familial mutation in affected relatives is not possible. We counsel such women that while the finding of a mutation is likely to be informative, the absence of a mutation cannot distinguish between a true negative (Mrs L has not inherited the familial inherited tendency that is present) and a false negative (Mrs L has inherited it, but the test is not yet good enough to find it). Since testing for mutations in families with a strong history of breast and/or ovarian cancer is currently at best 70% sensitive, a ‘negative’ result holds at least a 30% chance of being a false negative. This is because the BRCA genes are large and current testing techniques do not cover the genes comprehensively and because BRCA1 and 2 gene mutations do not explain all family histories such as that of Mrs L and Miss D. Even private genetic testing in the US (which is therefore free from resource allocation issues) admits to at least 15% of BRCA1 and 2 tests being falsely negative; it does not quantify the percentage of family histories such as those of Mrs L that are likely to be due to non-BRCA1/2 genes: Its overall sensitivity may, therefore, well be similar to 70% quoted above (http://www.myriadtests.com/provider/doc/tech_specs_brac.pdf accessed 20 November 2004).

In our case, Mrs L was not prepared to tolerate the uncertainty of a ‘negative’ result and wanted a prophylactic mastectomy unless we could prove that she had not inherited the familial tendency. The problem in this case is of course partly a temporal one. Had we known about the existence of Miss D and her result before Mrs L was counselled, a certain degree of fudging (with its own attendant ethical problems) might have been possible. However, since we had already explained the uncertainty of a negative genetic test result, this could not be changed without breaching Miss D's confidentiality. Furthermore, standard laboratory practice would be to use Miss D's sample as a positive control for Mrs L's predictive test, and thus her sample would have to be used without her consent.

Again, in response to Clarke, the genetics team had discussed the benefit of using Miss D's result for her wider family with her, but after an initial meeting Miss D had declined any further contact with the team.

As we have suggested, such cases represent a very small fraction of the caseload of clinical genetics departments, most families are happy to share their information, and where not, clinicians can often help by encouraging and facilitating communication. Nevertheless, the question remains, how should such cases be dealt with when they do arise? There will always be some cases in which all reasonable attempts are unsuccessful, maybe obviously as in the second case, or less obviously where we simply do not hear back from a person after we have asked them to communicate with their family. It is these cases that we should use to study the challenging issues of respect for confidentiality and right to know about relevant personal health information.