The role of ethnicity and diagnosis in rates of adolescent psychiatric admission and compulsory detention: a longitudinal case-note study

Abstract

Objectives

To explore whether ethnic variations in psychiatric admission and detention reported for adults also apply to adolescents and to establish the influence of diagnosis.

Design

A longitudinal, case-note study over a 10-year period.

Setting

An adolescent inpatient psychiatric unit in London.

Participants

All adolescents admitted to the unit.

Main outcome measures

Rates of admission and detention under the Mental Health Act, according to catchment area population.

Results

Young Black people were nearly six times more likely than the White group to be admitted with psychosis but showed no increase in admission for non-psychotic conditions. Young people in the Other group were over three times more likely to be admitted with psychosis but showed only a modest increase in admission with non-psychotic conditions. Young Asians were over twice as likely to be admitted with psychosis but were only one-third as likely to be admitted with non-psychotic conditions. Young people with psychosis in the Black and Other groups were around three times more likely to have been detained, but there were no significant differences for non-psychotic conditions.

Conclusions

Significant ethnic variation was found in the rates of admission and detention for adolescents. However, diagnosis was also an important consideration and must be taken into account when examining for evidence of ethnic bias in the use of mental health services by young people. Further investigation is required to establish whether adolescent care pathways are providing a safe and appropriate level of inpatient care for all ethnic groups.

Introduction

There is substantial evidence of ethnic variation in psychiatric admission and use of the Mental Health Act among adults in the UK. Systematic reviews have shown that Black patients are more likely to be admitted to psychiatric hospital than White patients¹ and that the rate of compulsory detention for Black patients is increased by around four times and for Asians by around two times.^1,2 These findings have been reflected in a census of the ethnicity of psychiatric inpatients in England and Wales, showing higher than average rates of admission among some minority ethnic groups, especially Black and White/Black mixed groups and higher than average rates of detention among Black, White/Black Caribbean Mixed and Other White Groups.³ The census data did not, however, show an increased rate of detention for Asian groups.

Only one previous study has addressed ethnic variation in psychiatric admission among adolescents.⁴ Prevalence according to ethnicity was reported for all diagnostic groups, but an analysis of the risk of admission according to ethnic population and comparisons between ethnic groups in the use of detention was limited to cases of psychosis. Black adolescents with psychosis were found to be over-represented as inpatients and were more likely to have been detained on admission.

In this study, a longitudinal sample has been taken from the clinical database of an adolescent unit in South London, to investigate variations in admission and in the use of The Mental Health Act, both by diagnosis and by ethnicity.

Method

Ethical approval for this study was provided by the local Clinical Governance Committee for Child and Adolescent Mental Health Services of the South London and Maudsley NHS Foundation Trust.

Source of data

Snowsfields Adolescent Unit provides an inpatient psychiatric service, including all-hours emergency admission, for young people aged 12–17 years. The main catchment area for the unit includes the London boroughs of Lambeth, Southwark and Lewisham. The total adolescent population of this catchment area is approximately 51,500 of which 48% are White, 35% Black, 5% Asian and 12% Other. The source of the data for this study was an electronic database maintained by the first author (RC). All admissions were logged in this database, with a note of the ICD-10 diagnosis made by the team, the ethnicity of the young person using UK Census 1991 categories (derived from the self-reported ethnicity as recorded in the medical notes) and a record of whether the Mental Health Act was used at any point during admission. This study examined all completed inpatient admissions from Lambeth, Southwark and Lewisham over the 10-year period from 1 January 2001 to 31 December 2010, a total of 435 cases.

In order to obtain sufficient numbers in each group for statistical comparison, ethnic categories were collapsed into the following groups:

• White: White British, White Irish, White Other

• Black: Black British, Black Caribbean, Black African, Black Other

• Asian: Indian, Pakistani, Bangladeshi, Asian Other

• Other: other ethnic groups and mixed ethnic origins

UK Census 2001 data were used to estimate the number of young people aged 12–17 years, by ethnic group, in the catchment areas included in this study. Although UK Census 2011 data are now available, the 2001 data were judged to be the best estimate of population size during the period of this study.

ICD-10 diagnoses were simplified into two groups: psychosis and non-psychosis (predominantly affective and stress-related disorders).

Data analysis

Odds ratios (ORs) were estimated using logistic regression and used White adolescents as the reference group. Confidence intervals (CIs) were estimated at the 95% level. Differences in proportions were tested with the χ² test or, if the expected value in one or more cells within a contingency table fell below 10, with Fisher’s exact test.

Results

A total of 435 inpatients were included in the study, with a mean age of 16.3 years. Ethnic representation was as follows: White, 141 (32%); Black, 214 (49%); Asian, 13 (3%); Other, 67 (15%). There was an overall excess of female admissions at 53%. Although there was some variation in the proportion of female admissions across ethnic groups, a comparison of each group with the White group did not show any of these differences to be statistically significant.

Differences were found between actual rates of admission and those expected according to catchment area population by ethnicity (Table 1). Young people in the Black group had a higher rate of admission compared to the White group (OR 2.0) as did those from the Other group (OR 1.9). Young Asians showed no significant difference in the rate of admission compared to the White group.

Table 1.

Admissions by ethnicity

Ethnicity	Catchment area population	Admissions	OR	Lower 95% CI	Upper 95% CI	P (χ2)	P (Fisher’s exact test)
White	24,508	141	1.0
Black	18,239	214	2.0	1.6	2.5	0.00
Asian	2,618	13	0.9	0.5	1.5	0.61
Other	6,124	67	1.9	1.4	2.5	0.00

The distribution of cases between diagnostic categories varied across the different ethnic groups. Only 23% of the White group were diagnosed with psychosis, in comparison to 65% for the Black group, 69% for Asians and 42% for Other. These differences in diagnostic distribution were reflected in the risk of admission according to catchment area population, when broken down by diagnosis and ethnic group (Table 2). Young people from Black, Asian and Other ethnic groups showed a statistically significant increase in the rate of admission with psychosis. This was especially pronounced for young people in the Black ethnic group (OR 5.7, p = 0.00). In contrast, the same group of Black adolescents showed no significant increase in the rate of admission with non-psychotic conditions. Those in the Other category showed a modest increase in the risk of admission (OR 1.4, p = 0.05) while Asians were less likely to be admitted for non-psychotic conditions in comparison with the White reference group (OR 0.3, p = 0.03).

Table 2.

Admissions by ethnicity and by diagnosis

Diagnosis	Ethnicity	Admissions	Psychosis within ethnic group (%)	OR	Lower 95% CI	Upper 95% CI	P (χ2)	P (Fisher’s exact test)
Psychosis	White	33	23	1.0
	Black	140	65	5.7	3.9	8.3	0.00
	Asian	9	69	2.6	1.2	5.3	0.02
	Other	28	42	3.4	2.1	5.6	0.00
Non-psychosis	White	108	1.0
	Black	74	0.9	0.7	1.2	0.58
	Asian	4	0.3	0.1	0.9	0.03
	Other	39	1.4	1.0	2.1	0.05

There was a small excess of male admissions within the psychotic group at 55%. No statistically significant variations in gender distribution were found in comparisons between the White group and each of the other ethnic groups. The non-psychotic conditions showed an overall excess for females at 62%. For the Black group, the gender split was significantly different to the white group (χ² P = 0.02), with Black females more likely than their White counterparts to be admitted with a non-psychotic condition (OR 1.2, 95% CI 0.8–1.7) and Black males less likely to be admitted (OR 0.5, 95% CI 0.3–1.0). Comparisons between the White group and the Asian and Other groups did not show any statistically significant differences in gender split.

The Mental Health Act was used at some stage during admission for 36% of all cases. Young people in the Black group were significantly more likely to have experienced formal detention than the White group (OR 2.9; 95% CI 1.8–4.7; χ² P = 0.00). The Asian group were less likely to have been detained, but this was not a statistically significant finding (OR 0.6; 95% CI 0.1–3.1; Fisher’s exact P = 0.74). Young people in the Other ethnic group showed an increase in the rate of detention similar to that of the Black group (OR 2.4; 95% CI 1.3–4.5; χ² P = 0.01).

As would be expected, use of the Mental Health Act was not evenly distributed across the diagnostic groups. Fifty percent of young people in the psychosis group had been detained at some time, whereas only 22% of the non-psychosis group had been detained. Analysis of use of the Mental Health Act by ethnicity and by diagnostic group (Table 3) showed that the risk of detention within the psychosis group was statistically significantly increased for both the Black and Other ethnic groups when compared to the White group (OR 3.0 and 3.1, respectively). Young Asians with psychosis were less likely to have been detained (OR 0.3), but this was not a statistically significant finding. In contrast, within the non-psychosis diagnostic group, no significant differences were found in the use of the Mental Health Act according to ethnicity.

Table 3.

Use of the Mental Health Act (MHA) by ethnicity and by diagnosis

Diagnosis	Ethnicity	No use of MHA	Use of MHA (%)	OR	Lower 95% CI	Upper 95% CI	P (χ2)	P (Fisher’s exact test)
Psychosis	White	23	10 (30)	1.0
	Black	61	79 (56)	3.0	1.3	6.7	0.01
	Asian	8	1 (11)	0.3	0.0	2.6	0.40
	Other	12	16 (57)	3.1	1.1	8.8	0.03
	Total	104	106 (50)
Non-psychosis	White	87	21 (19)	1.0
	Black	57	17 (23)	1.2	0.6	2.5	0.57
	Asian	3	1 (25)	1.4	0.1	14.0	1.00
	Other	28	11 (28)	1.6	0.7	3.8	0.27
	Total	175	50 (22)
All diagnoses	Total	279	156 (36)

More males than females were detained under the Mental Health Act (58%), but no statistically significant variations according to gender were found in comparisons between the White group and each of the Black, Asian and Other ethnic groups, either overall or within the two diagnostic groups.

Discussion

This study confirms the earlier finding that adolescents from Black ethnic groups have a much higher risk of psychiatric admission than those of White ethnicity⁴ and has important implications for the planning of adolescent inpatient services serving catchment areas with large Black populations. Our data indicate that the increased rate of admission applies to psychotic conditions only. Possible explanations for the higher rate of admission could include bias in the recognition of psychosis in Black adolescents, a greater need for admission due to poorer social support, higher levels of disturbance in young Black people with psychotic illness or a genuinely higher incidence of psychosis in the population. The available evidence supports the last of these as the most likely explanation, with the AESOP population-based incidence survey showing a higher rate of psychosis in the UK for Black ethnic groups (including young people in the 16–19-year age group) at a magnitude large enough to account for the rate of admission seen in our study.⁵

There was no overall difference in the risk of admission for non-psychotic conditions between White and Black adolescents. However, gender emerged as an important variable with Black females over-represented in comparison with White females and Black males much less likely to be admitted. Future multisite studies are required to explore this further, with large enough samples to allow for comparison between the different types of non-psychotic conditions.

Young people in the Other ethnic group (predominantly comprising young people of mixed African-Caribbean and White parentage) tended to show a similar pattern to the Black group with regard to rates of admission and detention. One important difference, however, was that the Other group did not follow the same gender variation for non-psychotic admissions. The under-representation of Black males with non-psychotic conditions was not seen in the Other group and contributed to a higher overall rate of admission for combined genders in comparison with the White group. Aside from this difference, it could be postulated that young people from a mixed Black/White ethnic background are closer in profile with regard to risk of admission and detention to young Black adolescents than to White adolescents. Further investigation with a larger sample size, to allow for the separate analysis of mixed ethic groups, would be of value in exploring this hypothesis.

Young people of Asian ethnicity did not show any overall increase in the rate of admission compared to the White group. However, of particular interest is the strong effect for diagnosis. Young Asians experienced higher rates of admission for psychosis compared with young White people (OR 2.6), but the risk of hospitalization with a non-psychotic illness was much lower (OR 0.3). The increased risk of admission for psychosis contrasts with the findings from Tolmac and Hodes,⁴ who found no difference in the rate of admission for Asian adolescents in London. A possible explanation for the low rate of admission for non-psychotic disorders among Asian adolescents could be that the true population incidence is lower in the Asian population, as has been shown in adult studies.^6,7 Similarly, data from epidemiological surveys of mental health disorder in children suggest that British Indians may have a lower incidence of the types of non-psychotic mental health problems likely to lead to hospital admission.⁸ There is, however, a lack of clear evidence for children from other ethnic groups within the Asian category.

The differing rates of compulsory detention according to ethnicity and type of diagnosis are of particular interest. Although data from adult studies in the UK have consistently shown an increased rate of detention for non-White ethnic groups, the issue of whether this risk is the same across diagnostic groups has not been examined. The only study addressing adolescent psychiatric admission collected data on all diagnostic groups but limited the analysis of the relative risk of admission and detention according to ethnicity to young people with psychosis.⁴ The findings of the present study suggest that diagnosis is an important consideration when examining for differences in the use of compulsory detention according to ethnicity.

Due to the small numbers of young people in the Asian ethnic group admitted with non-psychosis diagnoses, it is not possible to draw any confident conclusions regarding the risk of being compulsorily detained within this diagnostic group. However, for the Black and Other groups, the differing rates of detention according to whether the diagnosis is psychosis or non-psychosis is striking. These results tend to go against a hypothesis of systematic ethnic discrimination in the use of the Mental Health Act (if that were the case a similar bias would be expected for both psychosis and non-psychosis groups), but it remains a possibility that clinicians are showing bias in their interpretation of the risk associated with psychotic presentations in Black and Other ethnic groups. An alternative explanation would be that adolescents in these ethnic groups are presenting with more severe psychotic illness and/or higher-risk behaviours, such that an increased use of the Mental Health Act is warranted. Further research is needed to test these competing hypotheses.

Strengths

This is only the second study to have looked at ethnic variations in admission and use of detention among adolescents and is the only study to have included non-psychotic conditions in the analysis of risk of admission and detention according to catchment area population.

Limitations

This is a longitudinal, case note study and diagnostic or admission patterns may have changed over the course of a decade. Also, the need to merge ethnic groups for statistical purposes may have obscured some important differences within the broad ethnic categories.

DECLARATIONS

Competing interests

None declared

Funding

None declared

Ethical approval

Approval provided by the local Clinical Governance Committee for Child and Adolescent Mental Health Services of the South London and Maudsley NHS Foundation Trust

Guarantor

RC

Contributorship

RC – idea for study, design, data collection, drafting of manuscript. DB – advice on interpretation of data and background evidence, revision of manuscript

Acknowledgements

None

Reviewer

Susham Gupta