THE DEMOGRAPHY OF NURSES AND PATIENTS ON ACUTE PSYCHIATRIC
WARDS IN ENGLAND
Aims and objectives: To describe the ethnic and demographic composition of staff
and patients on acute psychiatric wards in England.
Background: A significant proportion of the UK population (7.6%) belong to an
ethnic minority, and there are concerns that ethnic minority patients are not well
served by psychiatry, in particular that they are subject to excessive force and
Design: Survey of a random sample of psychiatric wards close to three centres.
Methods: A survey was conducted of staff (n = 1536) and patients (n = 11,128) on
136 acute admission psychiatric wards.
Results: Ethnic minority patients were more likely to be admitted with a diagnosis of
schizophrenia, younger, more likely to be admitted for a risk of harm to others and
more likely to be legally detained. The association between ethnic minority status and
detention remains, even when risk, age, gender and diagnosis are taken into account.
Ethnic minority patients come from areas of greater social deprivation and
fragmentation. Ethnic concordance between staff and patients varies, but the greatest
difference is found in London where the proportion of minority staff is greater than
the proportion of minority patients.
Conclusions: There continues to be evidence that ethnic minority patients are subject
to an excessive amount of legal coercion in English mental health services. However
the proportion of staff belonging to an ethnic minority is greater than the proportion of
Relevance to clinical practice: Changes of recruitment strategies are required if
concordance is to be achieved.
The UK has a long history as a land of immigration, beginning with invading forces
of Romans, Saxons, Vikings and Normans from other parts of Europe up until the 11th
Century (Johnson, 1992). From the sixteenth century onwards, the first Black
immigrants arrived in small numbers following Britain’s involvement in the slave
trade. By the end of the 18th Century, at the height of the slave trade, there was a
relatively large Black population estimated between 10-20,000, concentrated in
London and the seaports (Fryer, 1984). But it is since the end of the Second World
War that the ethnic minority population in the UK has really grown, following the
passing of the 1948 British Nationality Act and the government’s encouragement of
people from the Commonwealth countries to come to the UK to work (Hansen, 2000).
The first wave of this economic migration was from the Caribbean in the 1950s and
1960s, when job opportunities were better in the UK than the Caribbean. Many of
those migrants were recruited for public sector employment, such as working on the
London buses, underground and in the health sector (Office for National Statistics,
1996). Among these migrants were nurses, particularly from the Caribbean and
Africa, recruited to work in the UK’s new National Health Service (Winkelmann-
The second wave of economic migration to the UK was from India and Pakistan,
peaking in the late 1960s and early 1970s, including the migration of Asians from
East Africa, both voluntary migration but also involuntary, with the expulsion of all
Asians from Uganda in 1972 (Robinson, 1986). However, since the 1970s
immigration from Commonwealth countries has slowed down as the UK government
began to restrict immigration with a series of legislation, culminating in the 1971
Immigration Act (Hansen, 2000). From the 1990s onwards, much of the migration to
the UK has been in the form of refugees and asylum seekers from regions affected by
war or political oppression, such as from the former Yugoslavia, Somalia, Sri Lanka
and Turkey (CVS Consultants & Migrant & Refugee Communities Forum, 1999).
Most recently, new waves of economic migrants have been arriving in the UK from
Central and Eastern Europe, with a sharp increase in numbers since 2004 when eight
new member states from that region joined the European Union (Gask, 2006).
The UK is now a multiethnic society; according to most recent national population
census in 2001 the size of the minority ethnic population was 4.5 million, representing
7.6% of the total UK population. The minority ethnicity population is highly
concentrated geographically in the large urban centres, with nearly half (48%) living
in London, comprising 29% of all London residents (Office for National Statistics,
2002). Regarding the composition of the minority ethnic population, the largest group
in 2001 were Asian or Asian British (the UK census uses self-identification ethnic
group categories), followed by Black Caribbeans, Black Africans and those of Mixed
ethnic backgrounds (Office for National Statistics, 2002). However, these broad
minority ethnic group categories obscure great diversity; for example, the Asian
population includes Indians, Pakistanis, and Bangladeshis. These categories are also
heterogeneous, for example the Indian population in the UK is composed of a variety
of regional and religious groups from the Indian continent as well as from East Africa.
There are also a variety of languages spoken such as Hindi, Gujarati, Bengali,
Marathi, Multani, Sindhi and Tamil (Commission for Racial Equality, 2007).
There is considerable concern about the psychiatric care and mental health status of
ethnic minorities in the UK. Some minority communities appear to have high rates of
psychiatric morbidity; for example Black Caribbean people have been found to have
higher prevalence rates of mental illness, particularly psychosis (King et al.,
2005;Bebbington et al., 2000). The recent Healthcare Commission’s ‘Count Me In’
Census (2007) of mental health inpatient in NHS and private healthcare hospitals
found that patients from a black and ethnic minority (BME) background are likely to
have a different experience and care pathway to white patients whilst in hospital.
This survey reported that 21% of all patients were from black and minority ethnic
(BME), demonstrating a huge over-representation in mental health inpatient settings
compared to the general population. Five per cent of the mental health inpatients
reported that English was not their first language; the biggest non-English speaking
groups were Bangladeshi (54%), Chinese (51%), Other (44%) and Pakistani (41%).
About 2% of inpatients said that they required an interpreter. There is general
concern that because of language barriers and cultural misunderstandings, some
people might be misdiagnosed or receive the wrong treatment. This problem has been
highlighted by organisations working with refugees and asylum seekers, with a
general lack of interpreters in the mental health system plus other issues such as
interpreters not having a good knowledge of medical and mental health terms and
how to convey non-verbal language (CVS Consultants & Migrant & Refugee
Communities Forum, 1999).
Another issue of importance is the high use of coercive measures (detention under the
Mental Health Act, seclusion, etc) with ethnic minority patients. Although whether
this actually occurs when all alternative explanations are taken into account is
disputed (Gudjonsson et al., 2000), UK psychiatry has been accused of institutional
racism (Prins, 1993;Blofeld et al., 2003).
One way to address these concerns to ensure that nursing staff match the demographic
characteristics of patients they care for (Department of Health, 2005;Department of
Health, 2003). Patients prefer staff from the same ethnic background (Napoles-
Springer et al., 2005;Chen et al., 2005;Garcia et al., 2003;Saha et al., 1999), and there
is evidence that this leads to better patient participation (Cooper-Patrick et al., 1999),
as well as improved healthcare team performance (Temkin-Greener et al., 2004).
To describe the demography of nurses and patients on acute psychiatric wards in
England, and the degree to which such matching is achieved.
A survey of the characteristics of inpatients and nursing staff on acute psychiatric
wards in three regions of England.
Sample and data collection
The data was collected as part of the City-128 study of observation and outcomes, a
multivariate cross sectional study to examine the rates of self-harm on acute
psychiatric wards, and how they were related to patient characteristics, the service
environment, physical environment, rates of conflict and containment, and staff
factors (Bowers et al., 2007). Random samples of acute psychiatric wards were taken
in three regions of England: the North West; East and West Midlands; and London.
The survey was undertaken on wards from 67 hospitals in 26 NHS Mental Health
Trusts. They served a total population of just under 19 million people, approximately
39% of the population of England. The 136 acute psychiatric wards that participated
in the study represented 25% of the estimated total of 551 wards in England.
Each ward returned data over a six month period by completion of a form at the end
of every shift. On this form new admissions were described in tick box fashion on a
limited number of variables. The postcode was also requested so that patients could be
matched to local area deprivation indicators: the Index of Multiple Deprivation, IMD
(Noble et al., 2004); and Social Fragmentation Score, SFS (Congdon, 1996).
Exclusion of all admissions with 3 or more missing data items resulted in the retention
of 11,128 admissions, 4,112 of which were accompanied by valid postcodes.
Consenting staff on the wards completed a questionnaire detailing their demographic
characteristics and experience of working in psychiatry. A total of 1536
questionnaires were returned (55% response rate).
All data was analysed using SPSS v12. Staff and patient subgroups were contrasted
using Chi Square tests, while Pearson correlation was used to explore the relationship
between staff/patient characteristics, deprivation, and other ward features. Logistic
regression was used with the patient data to identified features associated with
compulsory detention under the Mental Health Act.
Of the study patients, 94% lived in urban areas, in comparison to 80% for England as
a whole (ONS 2004), and 27% belonged to an ethnic minority group, as compared to
8% of the UK population as a whole. These statistics reflect the participation of
London, which provided one third of the sample, and is a major urban area where
48% of the total UK ethnic minority population reside (Office for National Statistics,
Staff and patient demography
Table 1 profiles the staff and Table 2 the patients, in terms of their age, gender and
ethnicity. The modal age group of staff was 30-39 years, most had been working in
their current position for between one and three years, and working in psychiatry for
more than five years. Two thirds were female, and of the nurses, the largest group was
that of staff nurse grades. Whilst the majority of nurses were of white ethnicity, the
largest minority group was African
Female nurses were significantly younger than male nurses (χ2 = 25.51, df = 5, p =
0.001), being more likely to be in the 20-29 age group, and were more likely to be of
white ethnicity (χ2 = 34.46, df = 5, p < 0.001). African nurses were less likely to have
been in their current post more than a year (χ2 = 40.24, df = 15, p < 0.001), had a
shorter duration of time working in psychiatry (χ2 = 54.38, df = 15, p < 0.001), and
were more often working in staff nurse grades (χ2 = 46.56, df = 20, p < 0.001) and
were younger in age (χ2 = 74.67, df = 25, p = 0.001).
The majority of patients were over 35 years of age, and there was an almost exactly
even split between the genders. The majority were of white ethnicity, but numbers of
the different ethnic minority groups were evenly distributed, with similar numbers of
Caribbeans, Africans, Asians and other ethnicities. Only the Irish group was smaller.
Just under a third of patients were admitted compulsorily under the Mental Health
Act, and a similar number had a diagnosis of schizophrenia. The majority were
admitted for risk of harm to themselves, and a minority for risk of harm to others.
Female patients were more likely to be white and less likely to be Asian (χ2 = 16.99,
df = 5, p = 0.005), much less likely to have a diagnosis of schizophrenia (χ2 = 395.46,
df = 1, p < 0.001), more likely to be aged over 35 years (χ2 = 194.2, df = 1, p < 0.001),
less likely to be sectioned (χ2 = 38.11, df = 1, p < 0.001), less likely to be admitted for
risk of harm to others (χ2 = 250.39, df = 1, p < 0.001), but no more nor less likely to
be admitted for risk of harm to self.
All ethnic minority patients are more likely to be admitted with a diagnosis of
schizophrenia, however this association was strongest for Caribbeans, then Africans,
then Asians, and weaker but still visible for Irish patients and those of other
ethnicities (χ2 = 427.3, df = 5, p < 0.001). A similar pattern is visible with regard to
age and compulsory detention, with ethnic minority patients being younger and more
likely to be formally detained, only in the case of youth the association is most
pronounced for African patients and absent for Irish patients (χ2 = 138.54, df = 5, p <
0.001), and in the case of detention the association is equally strong for Africans and
Caribbeans, and still present for Irish patients (χ2 = 434.42, df = 5, p < 0.001). In
comparison to white patients, all ethnic minority patients with the exception of the
Irish are less likely to be admitted for risk of harm to self (χ2 = 97.09, df = 5, p <
0.001), and all minority patients including the Irish are more likely to be admitted for
risk of harm to others (χ2 = 262.3, df = 5, p < 0.001). The association between ethnic
minority status remains, even when risk, age, gender and diagnosis are taken into
account in a logistic regression equation (see Table 3).
Staff, patients and ward catchment areas
For patients, the Index of Multiple Deprivation of the area served by the ward was
associated with fewer white patients and more of most ethnic minority categories (see
Table 4). It was also associated with a high proportion of admissions suffering from
schizophrenia, detained under the mental health act, younger and admitted for risk of
harm to others. Social Fragmentation showed exactly the same pattern of
relationships, only more strongly. When admitted, ethnic minority patients were likely
to find the ward environment of a lower quality, but had a better qualified workforce
to care for them.
The Index of Multiple Deprivation was not associated with any feature of the ward
staff (see Table 5). However Social Fragmentation was greater in the areas served by
wards with higher numbers of African staff, and lower in those areas served by wards
with higher numbers of white staff. Greater numbers of white staff were associated
with larger wards with a better physical environment quality and lower vacancy rate,
but with a poorer skill mix. These same variables were reversed for higher numbers of
African staff (i.e. smaller wards, worse environment, higher vacancy and richer skill
mix) and in part reflect the geographical distribution of staff between London and
elsewhere, as London has a particularly high number of African staff and a low
number of white staff.
Demographic concordance between staff and patients
Eleven scores for demographic concordance vs. non-concordance were created. By
concordance, we mean wards where the demographic characteristics of staff and
patients are matched, for example the proportions of each gender the same amongst
staff and patients. These scores can be calculated in a number of different ways where
there are multiple categories, as in the case of ethnicity. Further exploration of the
data shows that London is significantly different in nurse staffing and patient
demographics as compared to the rest of the sample. These two categories therefore
also need to be explored in order to obtain an accurate overview of the situation.
Scores and comparisons are presented in Table 6.
Age concordance. Calculated by subtracting the proportion of patients aged 35 and
under from the proportion of staff aged 30 and under (the data source categories did
not match). Positive scores represent more older staff than patients, zero (or just
below zero) represents a good match, and negative scores represent more younger
staff than patients.
Gender concordance. Calculated by subtracting the proportion of patients male from
the proportion of staff male. Positive scores represent more male staff than patients,
zero represents a good match, and negative scores represent more female staff than
Ethnic concordance 1 (simple absolute concordance). Calculated by taking the
absolute difference (i.e. disregarding the sign) between the proportion of staff white
and the proportion of patients white. Low scores represent a good match, high scores
a bad match, but composition of the minority nurses and patients may still differ
without being reflected in this score.
Ethnic concordance 2 (complex absolute concordance). Calculated by summing the
absolute differences for each ethnic category. Low scores represent a good match,
high scores bad match. Perhaps overall the most accurate measure, but does not
identify within which categories any mismatch occurs.
Ethnic concordance 3 (simple directional concordance). Calculated by subtracting the
proportion of patients white from the proportion of staff white. Positive scores
represent more majority staff than patients, zero represents a good match, and
negative scores represent more minority staff than patients.
White concordance, Irish concordance, African concordance, Caribbean concordance,
Asian concordance, Other ethnicity concordance. Each of these scores was calculated
by subtracting the proportion of patients within the category from the proportion of
staff within the same category. Positive scores therefore represent more within
category staff than patients, zero represents a good match, and negative scores
represent more outside category staff than patients. The white concordance score is
the same as 'Ethnic concordance 3'.
Because they are based upon proportions, in all cases, scores represent percentage
point differences in group composition (i.e. 0.01 represents a one percentage point
Overall, nursing staff are older than patients, however this non-equivalence is higher
in London, and arises because the staff in London are older than the staff elsewhere (t
= 2.26, df = 134, p = 0.025), and the patients in London are younger than the patients
elsewhere (t = 3.92, df = 134, p < 0.001). Overall, patients are more likely to be male
than nursing staff, but in this case London nurses have a greater concordance with
patients. While the patient gender mix does not differ between London and elsewhere,
there are a greater proportion of male staff in London's workforce (t = 3.81, df = 134,
p < 0.001), leading to a better match in this respect.
On all three overall scores of ethnic concordance, staff do not match patients.
Although London has high numbers of ethnic minority staff and patients, there is
significantly less ethnic concordance on London's psychiatric wards. The cultural
concordance 3 score demonstrates that this arises because there is an excess of ethnic
minority staff over patients in London. Overall there is a 9 percentage point excess of
ethnic minority staff, but while outside London there is concordance on this score, in
London there is a 29 percentage point excess of ethnic minority nurses over patients.
Further examination of the table shows that this ethnic non-equivalence arises because
there is a high number of African staff, and low numbers of white (and to a lesser
degree Asian) staff. These features are magnified by the differences between London
and elsewhere. In London the proportion of African staff is very high, with White and
Caribbean staff under represented; while outside London there is a strong under
representation of Asian staff.
The strong association between ethnic minority status and compulsory admission has
been reported before (Bhui K. et al., 2003). It has been suggested that this association
is due to raised rates of schizophrenia in minority populations and/or to different
demographic profiles of particular communities (e.g. more young people). However
the logistic regression undertaken demonstrates that both Africans and Caribbeans are
nearly three times more likely to be compulsorily admitted, even when age, gender,
diagnosis and risk are taken into account. The AESOP study (Morgan et al., 2006) has
demonstrated that some, but not all, of this differential arises through different routes
of referral and access to psychiatric care, with minority patients less likely to consult
their GP, and more likely to access care via the police. That still leaves open the
question why the Police are involved in more minority admissions. These stark
figures (53% of Africans vs 22% of whites are detained on admission) are a strong
pointer that some form of discrimination is taking place for all ethnic minority groups,
even though this has not been evidenced in lab studies of risk assessment (Lewis et
The profile of female patients was very different from that of male patients, although
the balance of genders was even. That balance represents an overall change, as female
admissions outnumbered males prior to 1981 (Prior and Hayes, 2001). Female
patients in this study were more likely to be white and less likely to suffer from
schizophrenia or be compulsorily detained.
The balance between male and female staff in psychiatric nursing appears to be
remarkably unchanging. An early study showed that 32% of psychiatric nurses were
male (John, 1961), compared to the 34% in this study. This figure matches that in a
recent report (Ferguson K et al., 2004), which also shows that among new recruits to
mental health nursing, the proportion of men is lower (27%). In this study younger
age was associated with female gender, also suggesting that a larger proportion of
female nurses are being trained. This could alter the future gender balance of the
workforce, but only if attrition rates are the same for both genders. If so, mental health
nursing is set to become female to a greater degree, and gender concordance will