Bristow's questions about the application of legal responsibility to psychiatrists are timely and interesting (see pp. 412-413 this issue). I hope he will be relieved to know that I have no pat answers to these questions, although I have some observations to make about the subject matter. I find myself wanting to agree with much of what he says — I have no desire to persuade the courts to force accountability in individual cases where to do so would not only be ludicrous, but also damaging, not least to the law. I am sure, however, that if a rationale could be effectively argued for applying a legal remedy against an individual psychiatrist or health authority, then the courts would take no account whatsoever of the effect of such a remedy on the morale of current and prospective psychiatrists, as Bristow implies they should, given his remarks concerning the views of trainees.

Bristow is right to point out that with the acceleration of the hospital closure programme (in which, incidentally, The Zito Trust played a part) no thought appears to have been given to the obvious need to set out new responsibilities for all those mental health professionals working with community care patients, in particular for their ‘responsible’ medical officers. The system has changed dramatically, but the law has, unsurprisingly, remained static. An eminent consultant psychiatrist once told me that there was a time when he knew where all his patients were (they were in hospital), but nowadays he regularly has to go out and peer through letter boxes looking for them. I do not for one moment imagine that this is a particularly straightforward or, indeed, gratifying change in working practice — for psychiatrist or patient. It is, however, the way things are and it is no surprise that general adult psychiatrists in particular often find themselves on the defensive. After all, after years of institutionalisation, community care must come as something of a culture shock for everyone concerned.

They should be reassured, however, to know that in cases where patients harm someone other than themselves the courts are currently on the psychiatrists' side and they enjoy the same kind of legal protection as the police — ironically the one profession psychiatrists do not want to emulate or become. So, as in the case of Osman, which went all the way to Strasbourg, it was held that no action could be taken for public policy reasons against the police for their negligence in the investigation and suppression of crime (Osman and Another v. Ferguson and Another, 1993). This followed the case of Hill v. Chief Constable of West Yorkshire Police (1989), in which the mother of the 13th and last victim of the ‘Yorkshire Ripper’ failed in her attempt to prove negligence against the police for not arresting him earlier, given the evidence she alleged they had accumulated. Also, in the case of Clunis v. Camden & Islington Health Authority (1998), who attempted to sue Camden & Islington Health Authority, it was held by the Court of Appeal that the health authority did not owe Clunis a common law duty of care under the Mental Health Act 1983. The proper (and only) remedy, said the court, was for Clunis to write a letter to the Secretary of State under Section 124 of the Act to complain that the health authority or social services authority were not fulfilling their obligations. Notwithstanding the circular and somewhat absurd nature of this judicial reasoning, these obligations (including ‘follow-up’, in Bristow's words) are set out under Section 117 of the Act, and in Clunis' case there was no ‘follow-up’ in spite of four missed Section 117 appointments in the period leading to the ‘misadventure’, as Bristow describes it.

Bristow says that he has not heard much debate concerning the duty of physicians or surgeons to follow up non-compliant patients in the community. This is odd, as the failure to make provision for after-care can undoubtedly give rise to a breach of a doctor's duty of care in the case of physical illnesses (Reference JonesJones, 1996). Therefore, if it is considered to be normal practice to follow up the patient to check on whether he or she is (a) taking the medication and (b) whether it is effective, there could be liability for the negligent omission to do so. In cases of self-harm or injury to another patient within an institution, the psychiatrist or the employing health authority can be found liable (Reference JonesJones, 1998). There is an anomaly here — a patient negligently discharged from hospital who then suffers severe injuries as the result of a suicide attempt stands to win substantial damages in compensation, whereas if the same patient in these circumstances kills someone there is no remedy. And in the US where Bristow believes action is only successful in Tarasoff v. Regents of The University of California (1976) type cases, there has been a significant number of cases expanding the duty of care to include non-identifiable victims (Reference MackayMackay, 1990; Reference MonahanMonahan, 1993). A new form of legal action called ‘negligent release’ has now developed in some states, and this has resulted in multi-million dollar damages for plaintiffs.

Should we go down the same road over here? At the moment, aside from failed suicide attempts, psychiatrists are well-protected. Besides, most would-be claims are made not against individuals but against their employing health authorities and there is some justification for saying that the buck stops with the employers, as they have ultimate responsibility for hiring, supervising, monitoring and firing. Yet the independent inquiries into homicide committed by people with mental illness (over 70 have been reported to date) regularly fail to produce a response more substantial than the usual “action plan, much of which has already been implemented”.

Bristow challenges me to tell him what I think should happen, so I will have a go. As a bare minimum I think psychiatrists should listen to their patients, particularly concerning the kind of medication they are expected to take. There is not much evidence of that happening (National Schizophrenia Fellowship, 2000). I think psychiatrists should listen more to the concerns expressed by families and not dismiss them out of hand. Many a patient has been refused admission, or been discharged too early, against the advice of the family. The medical profession, particularly its more senior members, have an unfortunate reputation for treating other human beings as little more than irritants. Psychiatrists should follow the law, in their case the current Mental Health Act and its accompanying Code of Practice. They should not refuse admission to a patient just because they feel he or she is not a danger to self or others, which is a misinterpretation of the law. (If I were to run a training session on the Act, I would spend the first hour on the difference between the words ‘and’ and ‘or’.)

I would want all general adult psychiatrists to undergo training in forensic psychiatry. I would want them to ensure that proper assessments of risk were carried out for patients before they are discharged from hospital, and that the latest and most comprehensive research on risk assessment and management, including assessment tools and protocols, become an integral part of daily clinical practice (Reference Blumenthal and LavenderBlumenthal & Lavender, 2000). I would like all psychiatrists to draw up after-care plans for each patient according to the law and not have to hear a consultant from a large London teaching hospital say, “Oh we don't bother with the Care Programme Approach here, we have our own way of doing things”. I acknowledge that many of the difficulties arise from this point forward, when the patient is discharged into the multi-disciplinary care of health and social services and possibly the voluntary sector with the additional involvement of the police, from time to time. As with the many child abuse cases, the risk to the individual seems to increase exponentially the more other agencies and/or individuals become involved. I would not want to apportion responsibility to a psychiatrist where a social worker, or the police, fail in their duty, or where a voluntary sector residential facility is poorly managed, as is often the case. If, however, important information concerning, for example, non-compliance with the after-care programme, is fed back to the responsible medical officer, who then takes no further action, would it be so unreasonable for the law to say there has been a breach of the duty to care for that patient?

The question is rightly put by Bristow, how long should this responsibility (academic at the moment) continue? Six months? Six years? I do not believe it would be sensible to put an arbitrary figure on it, but I do believe that if a patient is still being described as a patient and has a responsible medical officer, and that there is a discernible active line that can be satisfactorily established by the courts, then the question answers itself. No one is suggesting that psychiatrists (or rather their health authorities) should be responsible for every ‘misadventure’ in the community. The law of negligence operates from the premise that every professional should be judged by reasonable standards commensurate with his or her profession. So that, for example, a psychiatrist will not be expected to employ the precise actuarial skills of a forensic accountant. In fact, it is difficult to deduce from current case law what exactly are the reasonable standards of care as they apply to psychiatry. There is the American case of Tarasoff v. Regents of the University of California (1976), which does not apply over here, and that is about it.

So what is the problem? It is still generally safe to discharge a patient with a history of violence into the community when it is foreseeable that injury by an unknown member of the public will be sustained. Mental health review tribunals do just this in some cases and have not been accountable for their decisions since the case of Anderson v. Gorrie (1895). The law says ‘forseeability’ is not enough, there must be ‘some further ingredient’ (Lord Keith in Hill v. Chief Constable of West Yorkshire Police, 1989), defined as a proximate relationship between the parties. And most claims that satisfy these two requirements are likely to fail on public policy grounds because otherwise it would be difficult for psychiatrists, as for the police, to do their job without doing it defensively.

Undoubtedly, there are changes ahead, particularly concerning the sharing of information about patients and the imposition of compulsory treatment via the proposed new mental health tribunal (Department of Health & Home Office, 2000). It may be that a new Mental Health Act will clarify some of the issues discussed here and in Bristow's paper, but as we do not envisage new legislation for some 4 to 5 years we will have to continue to depend on NHS guidance and any future expansion of the law by the courts.