I am pleased that Lucassen and Parker1 have raised the serious issue of respect for confidentiality in the context of genetic information in your journal. There have been calls to reassess the medical profession's commitment to confidentiality when harm may come to those family members from whom relevant information has been withheld. When should practitioner's duty to other family members override their wish to respect the privacy of their patient?

While I agree with their conclusions, I do think that the two cases they describe could have been managed differently – perhaps with fewer resulting difficulties. In the first case (Case 1), the problem would either have been resolved or made explicit by the standard clinical practice of seeking access to the medical records of the affected individual. The genetic counselling team could have asked their client – the daughter of Jim and Mary – to obtain her father's consent (if he were still competent) for the genetics team to gain access to his medical records for the purpose of genetic counselling. Consent would either be granted or denied – in either case, the situation is much clearer and the question of a different diagnosis of Jim's dementia would be up for discussion.

In Case 2, the difficulties arise in part – at least, the problems take the shape they do – because of prior decisions made by the genetic service providers about access to molecular testing for susceptibility to breast cancer. The decision only to perform molecular testing on at-risk individuals when their family's mutation is known has been made because this ensures that those tests performed provide the most information possible. This is justified – I suspect – by an appeal to cost-benefit ratios and the efficient allocation and use of limited resources but this denies any sort of service to those whose family structure is not maximally helpful. Testing the at-risk niece Mrs L, who is seeking a prophylactic mastectomy because of her increased risk of breast cancer, would not be a maximally efficient use of resources when viewed from the public health perspective but it would be very useful in either providing strong reassurance or in justifying an important clinical decision. Performing the test might actually be more useful than Mrs L realises, but that is not a reason for not carrying out the test. Lucassen and Parker should not let the perfect (allocation of resources) be the enemy of the good (clinical practice). Furthermore, and without breaking Mrs L's confidence, the genetics team could arrange to discuss with Miss D her wish for privacy, in the hope that she might be persuaded to allow at least her mutation result to be used for the benefit of others in the family. Perhaps this could be achieved in a carefully negotiated and only semianonymous manner. Whether or not that consent could be obtained, however, the testing could go ahead on Mrs L if the genetics team agreed to less restrictive criteria for testing. The ethical issue in this case, I contend, is largely a product of clinical decisions that relate to rationing and the maximally efficient use of resources.

In both cases, I argue that a resolution of the ethical difficulties is largely in the hands of the clinicians – and if they choose not to resolve the issues then that is what requires a careful ethical analysis.