I began clinical school over two years ago. It was a memorable time: finally setting out to become a doctor—there was a mixture of excitement and uncertainty. I was conscious of the image of a naive, well intentioned medical student, who becomes worn down by “the system” and ultimately ends up compromising his original ethical standpoint. For this reason I approached ethics with much interest.

One of the first principles we learnt concerned confidentiality. This appeared simple: do not tell others who are not involved in the patient’s care, information gleaned from a consultation. It also appeared reasonable and an important part of the doctor/patient relationship. During a group discussion, however, we found it more complicated. How does one hold a confidential consultation on a ward? Which discussions are appropriate with colleagues, friends, or someone on the end of a telephone?

From an early stage one area that vexed me was the issue of doctors talking to the relatives of competent patients. The General Medical Council (GMC) guidelines do not discuss sharing information with relatives of competent patients.¹ Despite this they state that protection of confidence is important, and that consent for disclosure outside the health care team should be sought. This would imply that consent should always be sought from a patient before talking to relatives when the patient is not present. This seemed difficult. Surely relatives would want and expect information?

During my clinical attachments, however, I routinely saw doctors talking to relatives without any apparent attempt to seek permission from the patient first. I, and others, found this at odds with what we were taught.

A typical scenario would be this. An elderly patient presents with chest pain. She is diagnosed with and treated for a myocardial infarction (MI). Later the ward round takes place. Surrounded by the family, granny (as she has now become) is doing well—a little tired, perhaps bemused by all the fuss. The medical firm descends. Clever words are exchanged. Chins are rubbed. The consultant asks a few pertinent questions. He then explains his diagnosis and its significance. The family rush in with their questions—they are young and sharp. Consultant and firm turn and leave. Fine.

Sometimes though they seem to drag the family with them, and then a second dialogue develops—a more earnest, serious affair, away from granny. “Just what is granny’s prognosis?”

I talked to clinicians about this. They acknowledged that it was important to involve relatives and share information with them, agreeing also that consent for disclosure should first be sought. It was, however, clear that clinicians did think carefully about exactly what was said to whom.

I was unsure what I thought. I had agreed with the guidelines. Yet I was seeing a practice that was inconsistent with the guidelines, but which functioned well. Was this poor practice or were the guidelines, the basis for our teaching, too simple? I had two concerns about this “rule bending”. First, that the patient’s autonomy may be weakened when left out of a conversation. Second, that information may be shared with relatives that the patient does not wish to be shared.

On further reflection and observation I saw how the involvement of relatives tended to act to promote exchange of information and evaluation of choices. In this way autonomy can be enhanced, particularly if one holds a view of autonomy that stresses the importance of evaluation and rational choices.^2,3

Regarding the second concern, the sharing of information with relatives that a patient wants to remain confidential: this does give rise to a few complaints. Mistakes will happen and every area of medicine has risks; our response is to balance potential benefits against risk. Here I judged the benefits of talking to relatives to be greater than the rare but significant distress caused by sharing information.

During my first year I tried to reconcile the two approaches, the one taught in the classrooms and the other learnt on the wards. I favoured the latter, not because it is theoretically better but because in practice it works well. This reflects my experience of classroom ethical teaching, which is that it is limited when based on the guidelines alone, and so needs to be grounded in clinical situations. The BMA called for more teaching of ethics.⁴ This is welcome and I hope such teaching will be integrated with practical experience to enable a fuller understanding of medical ethics.