Although we agree that care in the community is perceived as a failure within the public domain and definitively as portrayed by the media, there are a number of issues around deinstitutionalisation that have not been addressed by Professor Leff (Reference Leff2001). Certainly, the apparent invisibility of community teams, the muddling of schizophrenia with personality disorder, ‘split mind’ and homicides, and the modern prevalence of homelessness are all factors. More important, perhaps, is that we do not really know why community care developed during the second half of the 20th century, and why it is now returning to what a 19th-century editor called ‘bricks and mortar humanity’ (Reference WynterWynter, 1859).

It may be that the studies following the planned resettlement of asylum populations quoted by Leff show no subsequent homelessness in the discharge populations, but this ignores the new long-stay problem. That is to say younger patients, who have never been through the asylum system, and who go in and out of in-patient units on the revolving-door circuit. Leff's experience of 20% of patients being homeless is out of date, current levels being nearer 40% or even 50% in our east London wards, for example. Another third are people readmitted from hostels, and now no longer accepted by these because of ‘risk management’, drug use or other ‘difficult’ behaviours.

This leads on to concerns about violent crime, in that there has been a decrease in the proportion of violent crimes committed by people with mental illness, but not a decrease in the numbers. This may reflect generally rising crime rates, but we remain ambivalent, in psychiatric circles, about the relationship between schizophrenia and criminality, and there seems to be a tendency to try to gloss over it. This also has an impact on what Leff calls the ‘mixed economy of care’, and a ‘complex network of interlinked facilities and professionals’. In fact this is a Gormenghast-like labyrinth, with voluntary agencies, privately run hostels, and forensic units carefully trying to ward off all difficult comers and, in the case of the latter, usually being full. The proportion of time spent on interface issues (e.g. meetings, letters, telephone conversations) compared with patient care is rising remorselessly.

Whether you call something a continuing care unit, a 24-hour nursing staffed hostel or a medium secure rehabilitation unit does not matter, since essentially you are reproducing the asylum. The fact of the matter is that we are now entering a period of reinstitutionalisation, in both the UK and other parts of the world, for reasons that we do not really understand. Deinstitutionalisation occurred in all Western industrialised countries, at a different pace and linked to very distinct national events such as the Psychiatry-Enquete in Germany, the Law 180 in Italy or Powell's ‘water tower’ speech in this country.

Now, there seems to be a similar underlying pattern across various countries. This time, it is reinstitutionalisation with a rising number of forensic beds, new-style institutions in the form of supported housing, and an increasing frequency of compulsory treatment. It is not just a matter of perceptions, but rather a notion of public safety. Thus, we see a rising tide of individualist preference over communal support (e.g. the car v. the train), a widening gap between stronger and weaker groups in society (e.g. the rich and the poor), and a medico-legal climate of blame and risk attribution. There is probably a realistic balance between what community care can do and what might benefit from old or new kinds of institutions, but such balanced acceptability needs more careful research.