The theory underlying medical classification is not always given sufficient regard. Here systematic methods developed in biology are summarised and compared with parallel developments in nosology. The classification of the functional psychoses is discussed as an example of a methodological problem in medicine. We consider that cladism could have a useful application in nosology, and could lead to classifications based on the concept of interrelationships of disease.

At post-mortem, Lewy bodies (LBs) were found in the brainstem and neocortex of eight out of 65 patients who had been collected during a prospective long-term study on clinically diagnosed Alzheimer’s disease. All eight patients had accompanying Alzheimer pathology which was less severe than in a sample of eight age and sex-matched patients from the same study with neuropathologically verified Alzheimer's disease. Parkinsonian features were more common in patients with LBs. There were no particular differences in duration of illness, severity of cognitive impairment, presence of hallucinations, or fluctuations in the course of illness. Frontal cerebral atrophy was more marked in patients with LBs, as was the loss of neurons in the basal nucleus of Meynert and the substantia nigra. Cognitive performance correlated with the number of pigmented neurons in the substantia nigra. We conclude that the differential diagnosis of LB dementia should be considered in patients satisfying NINCDS-ADRDA criteria for Alzheimer-type dementia who show marked Parkinsonian features and a frontal accentuation of cerebral atrophy.

Head circumference, body weight, body length, and shoulder circumference at birth were studied in 70 RDC schizophrenic patients and 70 demographically matched controls from the same delivery series, using information recorded in the medical records at the time of birth. With preterm babies removed from both samples, only head circumference among preschizophrenic infants was significantly smaller than that of controls. Preschizophrenic infants also had a disproportionately smaller head circumference in relation to body length than did controls. Among the former, a small head circumference was systematically related to an absence of family history of psychosis, but was not related to season of birth or recorded pregnancy complications. The findings were strongest for females. The results suggest there is an unidentified non-genetic factor in schizophrenia that disturbs prenatal cerebral development.

As a physician with a special interest in respiratory disease, I have been concerned with the semantic problems of my own specialty (Scadding, 1959, 1963) and of medicine generally (Scadding, 1967, 1972, 1975, 1988), and more recently have been invited to consider how far my approach may be applicable to psychiatry (Scadding, 1980, 1982, 1990). I therefore read the essay on nosology, taxonomy and classification of the functional psychoses by Parshall & Priest with interest, and venture to make the following observations.

First-degree relatives (FDRs) of 1018 schizophrenic and 812 control probands were investigated. Psychiatric morbidity was present in 34.8% of FDRs of schizophrenic probands and in 9.2% of FDRs of controls. There was significantly more psychiatric illness in the siblings and parents than in the offspring of both schizophrenic and control subjects. The morbidity risks for schizoid-schizotypal personality disorders, cannabis-use disorder and paranoid personality disorder were significantly higher in the FDRs of schizophrenic patients than in those of controls, suggesting a biological relationship.

The level of expressed emotion (EE) in 32 relationships between relatives and schizophrenic patients was assessed on three separate occasions over five years. EE was high on all three occasions in 25% of relatives, low on all three in 38%, and fluctuating in 38%; that is, in the majority of relatives (63%) the level of EE was stable over time. Three relatives who had previously shown high EE had evidence of dementia at the time of the third assessment, and showed low EE. Fourteen patients relapsed at least once over five years; patients who relapsed were evenly spread throughout those living in a home in which EE was consistently high, consistently low, or fluctuating. However, patients living in low-EE homes who did relapse did so significantly less often than those who relapsed and were living in homes in which EE was high or fluctuating. At the time of relapse, EE was not consistently high, and some patients in consistently high-EE homes did not relapse at all over five years.

A new standardised interview for the retrospective assessment of onset and early course of schizophrenia (IRAOS) was used to study the influence of age and sex on time of onset and psychopathology before first admission in 267 schizophrenic patients admitted for the first time. Mean age at onset, according to various operationalised definitions, differed by three to four years between the sexes. The age distribution at the earliest sign of mental disorder showed an early and steep increase until the age of 25 in males, and a delayed and smaller increase in females, with a second peak in women aged 45–79. Schizophrenia began with negative symptoms in 70% of cases, appearing two to six years before admission, and all positive symptoms appearing up to two years before. Both positive and negative symptoms accumulated exponentially. The early course of the disease was similar across age groups, except there was a longer period of negative symptoms before first admission in late-onset schizophrenia in women. The few significant age differences in symptoms were presumably due to general age-dependent reaction patterns like anxiety and depression or the cognitive development of personality, as indicated by an increase in fully elaborated positive symptoms, especially systematised paranoid delusions, with age.

Despite neuroleptic medication, many schizophrenic patients continue to experience residual positive psychotic symptoms. These residual symptoms cause distress and disability. We report a controlled trial of two cognitive-behavioural treatments to alleviate residual hallucinations and delusions. Forty-nine patients were recruited into the trial, of whom 27 entered the trial and completed post-treatment assessment, and 23 were reassessed at six-month follow-up. Patients were randomly allocated to either coping strategy enhancement (CSE) or problem solving (PS). Half the patients were allocated to a high-expectancy positive demand condition and half to a counterdemand condition to evaluate expectation of improvement. Patients receiving either cognitive-behavioural treatment showed significant reductions in pyschotic symptoms compared with those in the waiting period, who showed no improvement. There was some evidence, although equivocal, that patients receiving CSE improved more than those receiving PS. There was no evidence that improvements generalised to negative symptoms or social functioning, nor was there evidence that expectancy of treatment benefit contributed to the treatment effect.

It is ironic, in a field with many other ironies, that a number of studies (Johnstone et al, 1981; Mathai & Gopinath, 1985; Curson et al, 1992) have questioned what Mathai & Gopinath call “the hitherto inviolate (and ingenious) concept of institutionalisation”, at a time when massive deinstitutionalisation is casting into relief some of the practices to which it refers (Team for the Assessment of Psychiatric Services, 1990; Murphy, 1991; Abrahamson, 1993). However, the gap that this creates between theory and practice may be risky for the future. It would certainly be a major failure if the institutional era with which psychiatry's early professional identity was so intimately bound up were to end in unresolved contradictions. With this background, it will be argued here that re-examination of institutionalisation in the context of the longitudinal course of schizophrenic disorders both confirms the validity of the concept and sets a demanding agenda for alternative care in the community.

Compulsory treatment for anorexia nervosa was recently once again a topic for discussion following the case of J, a 16-year-old girl who unsuccessfully applied to the Court of Appeal to refuse treatment for her anorexia nervosa. In this instance legal opinion was sought in order to clarify the Children Act 1989. However, much of the media coverage focused on the controversy surrounding the compulsory treatment of anorexia nervosa, under the provisions of the Mental Health Act 1983. There is a lack of research into the compulsory treatment of anorexia nervosa, so debate has to be informed by clinical experience.

In a double-blind trial, 12 out-patients with primary anorexia nervosa received, for six weeks, either 10 mg cisapride or placebo, three times a day. Cisapride accelerated gastric emptying of a radiolabelled semisolid meal in all six patients; five gained weight and symptoms of gastric retention ameliorated in four. With placebo, three of six had emptying enhanced, four gained weight, and one's symptoms improved. For another six weeks, all patients received cisapride. In five of the patients who had received cisapride, emptying further accelerated or remained stable; in one it slowed. Of the six patients who received placebo, four had emptying accelerated, five gained weight, and symptoms improved in four. Longer administration of cisapride may, by enhancing gastric motor activity, alleviate symptoms of retention and thus help to change eating behaviour.

Lifetime criminal and psychiatric histories were examined in a consecutive series of 280 individuals of twin birth with a diagnosis of major functional psychosis who were seen and followed up at the Maudsley Hospital between 1948 and 1988. Their 210 co-twins, 35% of whom had a similar diagnosis, were ascertained and followed up over the same period. In the absence of reliable general-population estimates for lifetime conviction rates, co-twins were used as case controls. Among the 220 complete pairs, significantly more probands (25.7%) than co-twins (14.0%) were convicted, although there was no evidence for an independent genetic basis for criminal behaviour. Criminal conviction was significantly related to psychiatric diagnosis. There were specific patterns of offending, particularly among the schizophrenic men, who were also significantly more often convicted (48.6%) than the men with affective psychosis (19.4%), and more likely to receive a prison sentence. The schizophrenic patients were younger at their first conviction (mean age 22.6 years v. 30.8 years) and they had committed more violent offences than the affective group. In both diagnostic groups, ages at first psychiatric contact and first conviction were highly correlated.

“It is proposed that happiness be classified as a psychiatric disorder and be included in future editions of the major diagnostic manuals under the new name: major affective disorder, pleasant type. In a review of the relevant literature it is shown that happiness is statistically abnormal, consists of a discrete cluster of symptoms, is associated with a range of cognitive abnormalities, and probably reflects the abnormal functioning of the central nervous system. One possible objection to this proposal remains - that happiness is not negatively valued. However, this objection is dismissed as scientificaly irrelevant.”

A survey of attitudes to mental illness was conducted in a quota sample of about 2000 subjects in Malvern and Bromsgrove. Factor analysis showed three main components – benevolence, authoritarianism, and fear of the mentally ill. Residents of Bromsgrove, which is served by a traditional mental hospital, were slightly more tolerant than those living in Malvern, which has a community-based service, and has seen the closure of two mental hospitals in its vicinity during the last 10 years. The main demographic determinants of tolerance are age, education, occupation, and acquaintance with the mentally ill.

Investigation of the prevalence of eating disorders among a large sample of Japanese alcoholics admitted for hospital treatment revealed a strong association between eating disorders and alcohol abuse and alcohol dependence among young women. That association was not apparent in young men. The results of this study of the clinical courses of eating disorders and alcohol abuse and alcohol dependence suggest that women with eating disorders, especially bulimia nervosa, are at high risk of becoming alcoholic.

There is a strong prima facie case linking the abuse of psychiatry with difficulties about the concept of mental illness. However, a survey of recent Soviet literature showed that the concept of disease employed in the former USSR (where abuse was for a time widespread) was similar to its counterparts in the UK and USA in being strongly scientific in nature. A number of factors - legal, bureaucratic and professional - are important in abuse becoming widespread. These, however, fail to explain why psychiatry, rather than physical medicine, should be vulnerable to abuse. It is here that the concept of disease could be important. A scientific model of disease suggests that a significant vulnerability factor is the relatively underdeveloped status of psychiatry as a science. This leaves room for poor standards of scientific work in clinical research and practice, factors which are recognised as important in the Soviet case. In addition to the scientific element, there is an evaluative element of meaning in the concept of disease. Hence a second vulnerability factor could be the evaluatively problematic nature of judgements. of mental illness. It is concluded that a failure to recognise this factor greatly increases the vulnerability of psychiatry, not only to gross abuses, but also to inadvertent misuses of involuntary treatment in everyday practice. This conclusion, far from undermining the role of science in psychiatry, is a step towards clarifying its proper role.