We were interested to read the editorial by Pinner (Reference Pinner2000). In particular, the significant advocacy role we have traditionally accepted of relatives, which may conflict with the views of elderly people in general. It has become accepted good practice in clinical situations and research to seek and be influenced by the opinion of carers and relatives acting in this role.

The issue of consent among mentally incapacitated adults is a complex problem. We have been studying the views of elderly patients with mental health problems towards cardiopulmonary resuscitation (CPR). The study was confined to an acute in-patient population. Part of this enquiry required us to ask patients with a dementing illness “If your heart was to stop now, would you want us to bring you back to life?” and “If you were suffering from an incurable illness, would your answer be the same?” Other than completing a severity rating scale and depression inventory nothing else was required of the patient.

Relatives were asked for permission to approach patients. Considerable effort was made to recruit support by lengthy discussions and written material but to no avail. Eleven consecutive relatives refused, saying, in all cases, that the question would upset the patient too much.

There is evidence that relatives' proxy consent does not necessarily reflect the wishes of individuals and where divulging the diagnosis of dementia is concerned relatives wish this information to be withheld from the patient when they would expect to be told if they were affected (Reference Maguire, Kirby and CoenMaguire et al, 1996). This double standard also seems to affect psychiatrists (Hospital Doctor, 16 July, 1997). Denial has been reported as a means of coping by Alzheimer's patients (Reference Bahro, Silber and SunderlandBahro et al, 1995); is it possible that carers' decisions are influenced more by processes of denial and emotional self-protection than the needs of the patient?

Although there have been concerns that discussions about CPR with elderly patients might be distressing, the evidence indicates that elderly people are grateful for the opportunity to discuss this subject, which they consider important and upon which they wish to make their views known (Reference Morgan and KingMorgan & King, 1994). It is also clear that decisions and policies about CPR are usually absent or unclear and decisions are frequently left to junior staff in an emergency. Moreover, nursing and medical records are frequently contradictory and do not necessarily coincide with patients' expressed wishes (Reference Aarons and BeechingAarons & Beeching, 1991; Reference Morgan and KingMorgan & King, 1994).

Unless we are to make blanket decisions of policy or rely heavily on proxy consent that may not represent patients' wishes, how are these matters to be decided? At best, the present position hinders research and at worst does patients the injustice of excluding them from important decisions about their life or death.