Ethnic differences in risk of compulsory psychiatric admission among representative cases of psychosis in London

BMJ 1996; 312 doi: (Published 02 March 1996) Cite this as: BMJ 1996;312:533
  1. Sara Davies, senior registrara,
  2. Graham Thornicroft, director and readera,
  3. Morven Leese, statisticiana,
  4. Andrew Higgingbotham, research workera,
  5. Michael Phelan, assistant director and lecturera
  1. a PRiSM (Psychiatric Research in Service Measurement), Institute of Psychiatry, De Crespigny Park, London SE5 8AF
  1. Correspondence to: Dr Davies.
  • Accepted 23 November 1995


Objective: To compare the risk of detention under the Mental Health Act 1983 in a representative group of people with psychotic disorders from different ethnic groups.

Setting: Two defined geographical areas in south London.

Design: Annual period prevalent cases of psychosis were identified in 1993 in the study areas from hospital and community data. Standardised criteria were applied to case notes to establish diagnosis and detention under the act.

Subjects: 535 patients were identified, of whom 439 fulfilled ICD-10 criteria for psychosis.

Main outcome measures: Risk of ever having been detained under the Mental Health Act 1983, risk of detention under specific sections of the act during the study year, and risk of contact with forensic services for the different ethnic groups.

Results: 439 patients with a psychotic illness were identified. Nearly half of the white patients had been detained under the act compared with 70% and 69% of black Caribbean and black African patients, respectively. Black Caribbean and black African patients were more likely than white patients to have been involuntarily detained (adjusted odds ratio 3.67; 95% confidence interval 2.07 to 6.50 and 2.88; 1.04 to 7.95, respectively). Rates of use of sections 2, 3, and 136 in the study year were higher for black than for white patients, and black patients were more likely than white patients to have been admitted to a psychiatric intensive care facility or prison.

Conclusion: Independent of psychiatric diagnosis and sociodemographic differences, black African and black Caribbean patients with psychosis in south London were more likely than white patients to have ever been detained under the Mental Health Act 1983.

Key messages

  • Key messages

  • Black patients are significantly more likely to have been admitted to a psychiatric intensive care facility or to prison

  • The differential contact with mental health services may well set up a vicious circle

  • Purchasers and providers need to assess how accessible and responsive their mental health services are to black people


A high prevalence of severe mental illness, particularly schizophrenia, has been reported among black Caribbean people in Britain.1 2 3 4 5 6 This contrasts with lower rates of psychosis found among Irish people7 and conflicting results in Asian populations.2 8 9 Furthermore, high rates of hospital admission under the Mental Health Act 1983 have been reported for black Caribbean patients, particularly young men,10 11 12 13 both for compulsory admissions involving the police12 14 and for forensic orders.6 15 Independent of diagnosis, black Caribbean patients seem more likely to have contact with the police and forensic services,6 to be treated in intensive care facilities if detained under the act,16 and to have had a criminal conviction if they are young and male.17

Previous studies differ regarding whether the increased rate of compulsory admissions for black Caribbeans can be explained by an increased incidence of schizophrenia or other psychosis8 12 13 or whether it results from diagnostic bias14 or is an independent finding.6 13 16 Compulsory admission is more likely in patients who are living in temporary accommodation, in those not registered with a general practitioner, those attending a psychiatric outpatient department, and those with previous admissions.15 There are also major diagnostic differences between compulsory and non-compulsory admissions, with diagnoses of schizophrenia and mania overrepresented and depression underrepresented in the compulsory admissions group.13

When the increased compulsory admission rates have been shown to be an independent finding, explanations have been proposed which are either “patient” or “service” based.18 Such high rates may be attributable to different types of schizophrenia in this population,19 different perceptions of health services by black Caribbean patients, or later presentation to the psychiatric services20 21 22 or they might be because the police treat mentally ill black people differently from their white counterparts.18

With one exception,5 most previous studies have looked retrospectively at hospital admissions alone and have used clinical case note diagnoses.4 6 12 19 We now report rates of compulsory admission to hospital and contacts with forensic services as part of a wider study to evaluate the prevalence of psychosis and the provision of services among representative cases in two areas of south London.

Patients and methods

A case identification exercise was carried out to establish the annual period prevalence of all psychotic disorders in two catchment areas with 1991 populations of 38545 and 41740. Cases were identified by combining data from a wide range of hospital and community sources: psychiatric case records; social services; general practitioners; sheltered accommodation; voluntary, private, and self help care; the clergy; services for the homeless; and prisons. Cases were included on the basis of address of residence, even if no treatment had been received in the index year, and we included both those who had and those who had not ever been in contact with mental health services.

Possible cases were patients who had a clinical diagnosis at any time in their lives of any psychotic disorder. They were rated using the operational criteria checklist version 3.2,23 a standardised procedure to produce diagnoses according to the 10th revision of the International Classification of Diseases (ICD-10).24 All affective and non-affective functional psychotic disorders were included as definite cases.

For definite cases, information was collected on sociodemographic details, past diagnosis, all contacts with mental health services, use of the Mental Health Act, physical illnesses, family history of mental illness, and contact with the police. Details of compulsory admissions were collected from the Mental Health Act offices of the two local psychiatric hospitals.

Data on ethnic group were collected from a variety of sources. Ethnic group was recorded from the case notes according to the classification system used in the Office of Population Censuses and Surveys 1991 national census.25 Information was also collected on place and country of birth. When information was not available it was sought from staff who knew the patient. In a random half of all identified cases self rated ethnic group was used to validate the category previously allocated.

Results were analysed with SPSS for Windows, version 6, and STATA (STATA Corporation, Texas, USA). Relative risks were calculated with the white population as the reference group. Confidence intervals were estimated at the 95% level. Pairwise differences in proportions were tested with Fisher's exact test. Interactions tested in addition to main effects were (a) between ethnic group and age and (b) between ethnic group and whether subjects were born inside or outside the United Kingdom. Risk factors for compulsory detention were analysed by using logistic regression models with forward selection, based on the likelihood ratio, with significance for entry of variables at P=0.05.


Five hundred and thirty five possible cases with a clinical diagnosis of psychosis were identified in the two sectors. Of these, 439 patients had an ICD-10 psychotic diagnosis as follows: 238 (44.5%) schizophrenia, 13 (2.5%) psychotic affective disorder, and 188 (35.1%) other functional psychotic disorders. These 439 patients are considered to be the representative group of people with psychotic disorders in this study. Five (1.1%) of the patients were found from the community sources alone and had had no contact with mental health services. Sociodemographic characteristics showed that most were single, living in unsupported accommodation, living alone, and not in paid employment (table 1). There were no differences in these characteristics between the sample population and those cases who were excluded because they did not have an ICD-10 diagnosis according to the operational criteria checklist.

Table 1

Social characteristics of sample

View this table:

In 93.1% of patients ethnic group was established. Of the 194 patients who were subsequently interviewed and who self rated their ethnic group, all agreed with the rating made from the case notes. The ethnic composition of the whole group is shown in table 2. The category “other” comprises three (0.7%) Indian, two (0.5%) Chinese, one (0.2) Asian and nine (2.0%) other ethnic groups; these data were combined for the analyses. The 37 (8.4%) patients whose ethnic origin was unknown were excluded from further analysis. There were no significant differences in the sociodemographic characteristics of the ethnic subgroups, except that the black patients tended to be younger. As there were no significant sociodemographic differences between the black Caribbean and black African patients, and the numbers were small, these were combined into one group of black patients for the analysis of use of specific sections of the Mental Health Act and forensic contact in the study year.

Table 2

History of compulsory treatment under Mental Health Act

View this table:

The mean number of admissions was significantly higher for both black Caribbean and black African groups (table 1). Of the whole sample, 229 (51.5%) patients had been placed on a section of the Mental Health Act at some point in their life, and this rate was significantly higher for black Caribbean and black African patients than white patients (table 2).

Differences in risk of sectioning for different age groups are shown in table 3. Logistic regression confirmed the significantly higher risk of ever having been compulsorily detained for the black groups than for the white patients, and showed a decreasing trend with age. There was no evidence that the trends differed among the different ethnic groups.

Table 3

Effect of age on differences in rates of ever having been compulsorily detained under Mental Health Act. Values are proportions (percentages) of patients

View this table:

Sociodemographic characteristics were considered possible risk factors for sectioning, along with ICD-10 diagnosis and number of admissions, for the logistic regression. The risk factors identified are shown in table 4 with their adjusted odds ratios (black Caribbean and black African group compared with white group, and living alone compared with living with others). Backwards selection resulted in the same selection of risk factors. When the logistic regression model was refitted, controlling for age, the results did not change. The effect of being black Caribbean or black African, taking account of the other significant risk factors for sectioning under the Mental Health Act, was to raise the odds by 3.67 (2.07 to 6.50) and 2.88 (1.04 to 7.95), respectively.

Table 4

Independent effects on risk of having been detained under Mental Health Act 1983

View this table:

The sections of the Mental Health Act used to detain patients in the study year show that, even though the numbers are very small, the rates of detention of black patients under sections 136, 2, and 3, were significantly higher than those for white patients (table 5). This is most pronounced for section 136. For the other sections, larger proportions of black patients than white patients were detained, but the numbers were small. Results of ever having contact with forensic mental health services (table 6) show that black patients were significantly more likely than white patients to have been admitted at some time in the past to a psychiatric intensive care facility, or to prison.

Table 5

Numbers of patients compulsorily admitted under sections of Mental Health Act during study year

View this table:
Table 6

Numbers (percentages) of patients ever in contact with forensic services

View this table:


The results indicate that black African and black Caribbean patients are more likely than white patients to have been compulsorily detained in a psychiatric hospital at some time in their lives. This finding is independent of psychiatric diagnosis, total number of admissions a year, age, sex, marital status, employment, living setup, or type of accommodation. They are also more likely than white patients to have been admitted to a psychiatric intensive care unit or to prison.

As far as is known, this is the first study to measure the rates of ever having been compulsorily detained under the Mental Health Act 1983 for a representative group of people with psychosis in a whole catchment area population.

Ethnic groups, as reported in published findings, may fail to show the heterogeneity of such groups,26 27 but we have justified the categorisations we used in the analysis where the numbers were small. The number of black African patients was small and so their results need to be interpreted with caution. We principally report here ethnic differences in lifetime ever risk of sectioning for different ethnic groups; other published studies use rates of sectioning under the Mental Health Act 1983 for a specific number of admissions.8 10 12 13 15 20 The black Caribbean and black African patients described here are younger, and have had contact with services at a younger age than white patients and so they do not have an increased “exposure time.” Our study has the further advantage of only counting each individual once so that we avoid the problem of overrepresenting a few individuals who have repeated admissions.

Our findings show a 50% overall risk of ever having been compulsorily admitted to hospital among patients with psychosis. A previous study has shown that for a similar area in south London over a similar time period, 26% of all annual psychiatric inpatient admissions were compulsory, with no difference between ethnic groups, independent of diagnosis.28

In a wider context the most recent unpublished figures from the Mental Health Act Commission show that in 1989-90, 7.2% of all admissions to psychiatric hospitals in the United Kingdom were under sections of the Mental Health Act, and of these 4.0%, 1.3%, and 0.8% were under sections 2, 3, and 4 of the Mental Health Act, respectively. This compares in our study with 8.8%, 8.5%, and 1.2%, respectively, in 1992-3, and as our data count individuals, they will underestimate compulsory admission rates. These data do show rates of sectioning in south London that are far higher than the United Kingdom average of a few years ago, and this is confirmed by other recent reports,29 30 although the reductions in the numbers of available beds nationally means that these figures need to be interpreted cautiously.

This study indicates that independent of diagnosis, black people come into contact with mental health services differently from other groups. Black African and black Caribbean patients are more socially isolated, have greater contact with the police and forensic services, and are more likely to receive involuntary treatment.6 29 30

Black patients may see mental health services as inaccessible or inappropriate to their requirements.12 Other recent work suggests that the outcome for black Caribbean and black African patients may be more favourable in terms of risk of self harm and duration of illness despite more involuntary admissions and more imprisonments.31 This suggests a complex picture.

Whatever the reasons for these higher compulsory admission rates among black patients, this differential experience of contact with services may well establish a vicious circle in which black patients may see services as untherapeutic, may delay seeking help, and will have an increased likelihood of compulsory admission.28 Purchasers and providers need to address the issue of how accessible and responsive their mental health services are to black people.

We thank Professor Robin Murray and Dr Peter Jones for their helpful comments on an early draft of the manuscript.


  • Funding Department of Health.

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
View Abstract