Ethnicity

Ethnicity was assigned by observers according to the Office of Population Censuses and Surveys (OPCS) ethnicity categories (OPCS, 1992). This was supplemented by information on patients' and patients' parents' place of birth. The aim was to produce a group of Caribbean origin and as homogeneous a British White group as possible for comparisons.

The OPCS ‘White’ category can include people from a variety of countries. The largest minority ethnic group, the Irish, are usually subsumed in this group although they may have different mental health needs to those born in the UK of British parents (Reference Littlewood and LipsedgeLittlewood & Lipsedge, 1997). This can make explanations of any differences between groups difficult. In this study, White British=OPCS category White with mother and father born in the UK.

Patients of Caribbean origin are also a heterogeneous group. However, this group has shared histories, reasons for migration, concentrations in certain geographical areas of the UK and shared experiences of discrimination that may produce similarities in their needs and experience of services. We did not include those of African origin. Those of mixed parentage were included in the group of Caribbean origin if they had been assigned to the Black—Caribbean or Black Other OPCS group. This may reflect the experience that patients have had with the mental health system and society in general. Patients of mixed Caribbean and African parentage were excluded, as were patients with one parent from an unspecified country.

In this study Caribbean origin=OPCS Black—Caribbean or Black Other with mother or father born in the Caribbean or UK.

Follow-up

Stratified randomisation ensured that equal proportions of White British and African—Caribbean patients were allocated to each treatment arm. Patients were followed up for an average of 2 years from the time they were randomised. They were re-interviewed at 1 and 2 years. At 1 year all the instruments except the WHO Life Chart (WHO, 1992) were repeated. At 2 years all the instruments used at baseline were repeated. We present data for the 2-year follow-up.

Follow-up interviews were undertaken by independent researchers who were not involved in patient care. Patients and, when available, relatives and carers were interviewed. Each patient's case manager was interviewed. Case managers had a detailed knowledge of their patients. Throughout the study they carefully documented every visit to their patients for an assessment of care received and case management model fidelity. Other mental health professionals involved in the case were interviewed. Patients' case notes were also reviewed. The aim was to construct as accurate a picture as possible of the course of the illness, admissions and treatment.

Measurement of outcome

Outcome in psychosis is complex and variables that measure outcome are often interrelated. Van Os et al (Reference Van Os, Fahy and Jones1996) identified six dimensions of course and outcome in psychosis:

1. (a) negative symptoms/disability;

2. (b) illness severity/course;

3. (c) time living independently;

4. (d) unemployment;

5. (e) prison/vagrancy;

6. (f) depression/self-harm.

We have represented these six dimensions of outcome with 11 measures used in the UK700 project (Table 1).

Table 1

Dimensions of outcome and the variables used to represent them

Dimension of course	Variable used
Negative symptoms/disability	Mean DAS score
	Negative symptoms usually present (WHO Life Chart)
Illness severity	Illness course (WHO Life Chart)
	Usual symptom severity (WHO Life Chart)
	Mean time in hospital
Time living independently	Mean time living independently
Unemployment	Mean time unemployed
	Mean time employed
Imprisonment/vagrancy	Imprisonment over follow-up period
	Vagrancy over follow-up period
Depression/self-harm	Self-harm or completed suicide over follow-up period

Over the follow-up period, course type was rated using the WHO Life Chart (WHO, 1992) as episodic (no episode longer than 6 months), continuous (no remission longer than 6 months), neither episodic nor continuous, and not psychotic in this period. The ‘usual severity of symptoms’ rating indicated the symptomatic level of the patient during the most of the follow-up period. Ratings were severe, moderate, mild or recovered. Self-harm included all attempts at self-harm, regardless of the outcome (i.e. both parasuicide and completed suicide were included).

Treatment variables

In addition to information on course and outcome, data on five areas of treatment over the follow-up were also collected:

1. (a) whether prescribed antidepressants over the follow-up period;

2. (b) whether prescribed lithium over the follow-up period;

3. (c) time spent on antipsychotics over the follow-up period;

4. (d) whether the patient has had rehabilitation over the follow-up period;

5. (e) whether the patient has had psychotherapy over the follow-up period.

The measures used and methodology replicate those of McKenzie et al (Reference McKenzie, van Os and Fahy1995). However, in addition to the measures used above, McKenzie et al (Reference McKenzie, van Os and Fahy1995) reported involuntary admissions over the follow-up period, results from the Iager scale (Reference Iager, Kirch and WyattIager et al, 1985) to measure negative symptoms and results from the Hamilton Rating Scales for Depression (Reference HamiltonHamilton, 1960). This information was not collected by the UK700 trial.

It should be noted that the measures of outcome in this study are not the same as the main outcome measures for the UK700 study. We decided before the analysis to use the same outcome variables as we had used previously so that we could compare the results of the present study with our previous study (Reference McKenzie, van Os and FahyMcKenzie et al, 1995). We also decided, a priori, to use the same plan for statistical analysis.

Analysis

The means of continuous variables and the proportions for binary variables were compared between the two ethnic groups. Means and proportions were adjusted using regression analysis to yield odds ratios. Models included age, diagnosis, educational level and time from onset of the illness to study entry as possible explanatory variables.

Variables measuring time (e.g. time spent in hospital, time unemployed, etc.) were expressed as the proportion of the length of the follow-up period. Some outcome variables were transformed into dichotomous variables if they inclined towards two categories. For example, course of illness was transformed into ‘continuous’ and ‘non-continuous’ and usual symptom severity was transformed into ‘recovered’ and ‘non-recovered’. Similarly, for continuous variables such as time in psychotherapy and time on anti-depressants, which were severely skewed, tending towards a half-normal distribution (presence-absence), dichotomisation was also considered more appropriate. Analyses were carried out using SPSS for Windows 8.0.

Baseline characteristics

At baseline, 203 African—Caribbeans and 234 White British were identified. African—Caribbeans (AC) were younger than British White (BW) patients (AC, mean 37.3 years; BW, 40.1 years; P=0.02), they were less likely to have A-levels or a degree (AC, 8%; BW, 24%; P=0.001) and were more likely to be diagnosed as suffering from schizophrenia (AC, 50%; BW, 31%; P=0.001). On univariate analysis, African—Caribbeans had a shorter time from onset of illness to study entry than whites (AC, mean=145.8 months, s.d.=104 months; BW, mean=168.5 months, s.d.=131 months; P=0.045), but this association disappeared when baseline differences in diagnosis were taken into account. There were no significant differences in illness length, number of days in hospital, illness course or number of hospital admissions at baseline.

Follow-up

Univariate analysis

There were no significant differences in symptomatic outcome or illness course on univariate analysis but significant differences were found in the treatment received over the follow-up period (see Table 2).

Table 2

Scores for dimensions of course, treatment and outcome

Dimension of course/outcome/treatment	Variable	African-Caribbean group (s.d. or %)	British White Group (s.d. or %)
Negative symptoms/disability	Mean DAS score	1.1(0.83)	1.1 (0.76)
	Negative symptoms usually present	55 (28)	62 (27.5)
Illness severity	Continuous illness course	61 (31)	84 (37.5)
	Usual symptom severity ‘recovered’	47 (23)	50 (22)
Time living independently	Mean time in hospital (days)	84.7 (127.7)	78.8 (123.7)
	Mean time living independently (months)	15.8 (9.7)	15.1 (9.9)
Unemployment	Mean time unemployed (months)	17.1 (8.6)	17.3 (8.5)
Imprisonment/vagrancy	Whether imprisoned over follow-up period	13 (7)	8 (3.5)
	Whether a vagrant at any time over follow-up period	3 (1.5)	8 (3.5)
Depression/self-harm	Self-harm or suicide over follow-up period	17 (8.4)	33 (14.1)
Treatment variables	Prescribed antidepressants over follow-up period	25 (13)	85 (38)**
	Prescribed lithium over follow-up period	30 (15)	62 (28)***
	Months on antipsychotics	20.2 (7.3)	19.1 (8.2)
	Rehabilitation/day treatment over follow-up period	34 (17)	57 (25)*
	Psychotherapy over follow-up period	12 (6)	38 (17)**

Multivariate analysis

Significant differences between the groups for the prescription of antidepressants and psychotherapy but not day treatment/rehabilitation or lithium therapy remained after we controlled for diagnosis, age, length of illness before study entry and education. In addition, after controlling for confounders, African—Caribbeans were nearly 40% less likely to have a continuous illness course than British Whites (Table 3).

Table 3

Unadjusted and adjusted odds ratios (ORs) for significant associations in the univariate analysis

	African-Caribbean v. British White (unadjusted OR)	African-Caribbean v. British White (OR adjusted for education, age, length of illness, diagnosis)
Continuous illness course	0.76 (0.5-1.10)	0.61 (0.39-0.96)
Prescribed antidepressants over follow-up period	0.24 (0.14-0.38)	0.28 (0.16-0.48)
Prescribed lithium	0.47 (0.29-0.77)	0.68 (0.39-1.19)
Day treatment	0.63 (0.4-1.0)	0.64 (0.37-1.09)
Psychotherapy over follow-up period	0.31 (0.16-0.63)	0.33 (0.16-0.71)

Methodology

Comparison with other studies

The outcome of psychosis in people of Caribbean origin in the UK is complex. Administrative outcomes, such as readmission rates and compulsory admissions, are reported as poor but there are mixed reports of rates of recovery and the course of illness.

A recent population-based first-onset study concluded that findings of its own and other studies are consistent with a marginal effect (if any) for better outcome in African—Caribbeans (Reference Harrison, Amin and SinghHarrison et al, 1999). The authors hypothesised that inconsistent findings could be expected because of the small sample sizes of studies. We would add that inconsistent findings may be due partly to differences in definitions of the ‘African—Caribbean’ and comparison groups and differences in study design. There are problems with directly comparing this study with first-onset studies even though we have tried to minimise these.

We set out here to try to replicate the findings of our previous study in which people of Caribbean origin in the UK were found to be less likely to suffer from a continuous illness and were more likely to be recovered and also imprisoned over a follow-up period. They were less likely to be treated with antidepressants, lithium and rehabilitation and were less likely to self-harm (Reference McKenzie, van Os and FahyMcKenzie et al, 1995). We have been able to define African—Caribbean and British White groups similarly in this and the previous study. We have partly replicated the findings.

CLINICAL IMPLICATIONS

Compared with British Whites:

• ▪ People of Caribbean origin with psychosis are less likely to be treated for depression.

• ▪ People of Caribbean origin with psychosis are less likely to be treated with psychotherapy regardless of the diagnosis.

• ▪ People of Caribbean origin with psychosis are less likely to have a continuous illness course.

LIMITATIONS

• ▪ All four sites were in the inner city. This limits our ability to generalise results to the whole population or discuss rates or aetiology.

• ▪ As with most studies in this field, standard tools were used, not all of which have had thorough cross-cultural validation.

• ▪ Assessors were not blind to the ethnicity of patients, which could have introduced bias.