Inpatient mental healthcare in England and Wales: patterns in NHS and independent healthcare providers

Summary

Objectives

Independent healthcare, most of it NHS-funded, provides a significant and growing proportion of inpatient mental healthcare in the UK, but information about patients in these providers is limited. This paper examines differences in the profiles of mental health inpatients in NHS and independent providers in England and Wales, and assesses whether current information systems are adequate for monitoring patient flows and care, given the plurality of service provision.

Design

Data from the national censuses of inpatients in mental health services in England and Wales in 2006 and 2007 were analysed to examine differences in demographic and other characteristics between inpatients in NHS and independent providers.

Setting

All NHS and independent providers of inpatient mental health services in England and Wales in 2006 and 2007.

Main outcome measures

Patients in independent providers were younger, 60% were on low/medium secure wards (compared with 16% in the NHS), they were 44% more likely to be detained and referrals were predominantly from NHS inpatient services. For all ethnic groups, ratios of detention on admission were higher in independent providers.

Conclusions

This analysis highlights differences between inpatients in NHS and independent providers of mental health services. We also highlight the inadequacy of current information systems for monitoring care, and the urgent need for standardized data across all NHS-commissioned mental healthcare, irrespective of whether it is publicly or privately provided. This is especially important in view of the increases in independent sector provision, and the specialist nature of their services. Such information is critical for commissioners, regulators of health and social care, and other audiences for monitoring patient flows, the quality of care provided, usage of the Mental Health Act 2007 and compliance with equality legislation.

Introduction

The independent sector provides a significant and growing proportion of inpatient mental healthcare in England and Wales, and many providers are registered to take only patients liable to be detained.¹ However, information about patients in independent providers is much more limited than that available for inpatients in NHS mental health services. This is because many of the statutory data returns that apply to the NHS, especially in relation to patient level data-sets, are not required of independent healthcare providers.

The Healthcare Commission, the Mental Health Act Commission (MHAC) and the Care Services Improvement Partnership (CSIP)/National Institute for Mental Health in England (NIMHE) undertook a national, one-day census of inpatients in NHS and independent mental health and learning disability services in England and Wales in 2006 and 2007.^2,3 A similar census, covering only mental health services, was conducted in 2005.⁴ The censuses were undertaken in support of the Department of Health's five-year action plan for improving mental health services for black and ethnic minority communities, Delivering Race Equality in Mental Health Care,⁵and included both NHS and independent (private and voluntary) providers of inpatient mental health and learning disability services. The censuses asked organizations to provide selected details about their patients, such as demographic details including ethnic group, whether or not they had been detained under the Mental Health Act 1983 and date of admission.

We analysed the 2006 and 2007 data to examine differences in the profiles of inpatients in NHS providers, and those in independent providers of mental health services in England and Wales. Although the censuses only provide a one-day snapshot of patients in these establishments, they achieved comprehensive national coverage and, in the absence of other information sources, they provide an opportunity for examining differences in the patient profiles of the two sectors. Our analyses highlight the urgent need for better information about users of mental health services.

Method

Results are presented from the national censuses of inpatients in mental health services in England and Wales conducted on 31 March 2006 and 30 March 2007 (patients in learning disability services were excluded from the analysis). The censuses covered NHS, private and voluntary providers of inpatient mental health services, and were conducted jointly by the Healthcare Commission, MHAC and CSIP/NIMHE. For brevity, the term ‘independent providers’ is used throughout to include all private and voluntary providers of mental health services. Registrations with the Healthcare Commission were used to identify the independent providers of inpatient mental health services. Further details about the census, and the protocol setting out details of the data collection, are available at http://www.healthcarecommission.org.uk/guidanceforhealthcarestaff/nhsstaff/complaintsbypatients/countmeincensus.cfm.

In 2007, all 93 NHS and 164 independent providers eligible for participation in the census returned individual records for their patients, giving complete coverage of providers. In 2006, all 108 NHS trusts and 130 of the 146 eligible independent providers participated in the census. The 16 non-NHS providers that did not participate in the 2006 census were small establishments, hence patient coverage in 2006 was also virtually complete.

We analysed the distribution of patients across NHS and non-NHS providers. For both sectors, we also analysed the characteristics of patients (age, gender, ethnicity), the type of ward they were in (level of security), their median length of stay from admission to census day, their detention status under the Mental Health Act 1983 on admission and the source of their referral. The classification of referral sources was that used in the Mental Health Minimum Data Set (MHMDS); however, some additional categories were also included in the census because feedback from providers indicated they were important (for example, prison, community mental health teams). It is not possible to confirm whether or not this measures the original referral source. For example, it is possible that some referrals via community mental teams originated from GPs. Information on diagnosis was not available from the censuses. Differences in proportions between the NHS and independent sector patient profiles were tested for statistical significance using chi-square tests.

We derived age and gender standardized ratios, and 95% confidence intervals, for detention at admission (by ethnicity and overall) for both sectors. The age and gender-specific detention rates (at admission) among the overall mental health inpatient population in England and Wales, as recorded in the census, were used as the standard for deriving the ethnic group-specific detention rates. Ethnicity coding for patients was 99% complete in both censuses; ethnicity was self-reported for 75% of patients and, for the remainder, it was reported by staff or relatives.

Results

In the 2006 census, 32,023 patient records were returned from 238 providers; 10.9% of patients were in independent providers (Table 1). In the 2007 census, 31,187 patient records were returned from 257 providers; 13.7% of patients were in independent providers. Across the two years, there was a decline in both the number of NHS providers, and the number and proportion of all inpatients in NHS providers. Conversely, there was an increase in both the number of independent providers, and the number and proportion of patients in such providers.

Table 1.

The number of NHS and independent providers of mental health services and inpatients, England and Wales

Provider	2006 census	2007 census
Providers (n)	Inpatients (n)	Inpatients (%)	Providers (n)	Inpatients (n)	Inpatients (%)
NHS	108	28,527	89.1	93	26,895	86.2
Independent	130	3496	10.9	164	4292	13.7
TOTAL	238	32,023	100	257	31,187	100

The demographic profile of patients changed little across the two years ( Table 2). In both 2006 and 2007, patients in independent providers were significantly younger than those in NHS providers (median ages 40 and 51 years, respectively), and higher proportions were men (60% compared with 55%, respectively). The proportion of all patients in NHS providers who were from black and ethnic minority groups (defined as those other than White British) increased from 20.5% in 2006 to 21.2% in 2007; however, the corresponding proportion for independent providers increased from 18.6% to 23.4%. Notable patterns by ethnicity were the significant increase between 2006 and 2007 in patients from the Other White group in independent providers, and that in both years higher proportions of patients from the White/Black Caribbean Mixed group were in independent providers compared with the NHS, while the reverse was true of patients from the Other Black group.

Table 2.

Characteristics of inpatients (%), England and Wales

	2006 census*	2007 census*
	NHS	Independent	NHS	Independent
Age (years)
<17	1.5	4.3	1.4	4.5
18–24	6.9	11.4	6.7	12.1
25–49	39.8	54.4	40.5	54.4
50+	51.8	29.9	51.4	29.0
Median age	51.0	40.0	51.0	40.0
Gender
Men	54.7	60.5	55.4	60.4
Ethnic group
White British	78.4	80.4	77.9	75.6
White Irish	1.8	2.1	1.7	1.8
Other White	4.0	2.2	3.9	9.2
White and Black Caribbean Mixed	0.8	1.4	0.8	1.4
White and Black African Mixed	0.3	0.4	0.3	0.3
White and Asian Mixed	0.3	0.5	0.3	0.3
Other mixed	0.5	0.9	0.6	0.4
Indian	1.3	1.3	1.3	1.1
Pakistani	1.1	0.7	1.0	0.8
Bangladeshi	0.5	0.3	0.5	0.1
Other Asian	0.8	0.5	0.9	0.6
Black Caribbean	3.9	4.8	4.3	4.1
Black African	2.0	2.0	2.2	1.5
Other Black	1.8	0.5	1.9	0.9
Chinese	0.3	0.2	0.3	0.1
Other	1.1	0.8	1.2	0.7
Not known	1.1	1.0	0.9	1.0
Security level of wards
General	80.3	38.7	82.4	36.9
Low secure	9.3	41.8	7.2	40.7
Medium secure	7.2	18.8	7.4	22.4
High secure	3.2	0.7	3.1	0
Median length of stay from admission to census day (days)	93	435	98	372
Detained on admission
Yes	37.4	64.4	39.8	65.5
Section of MHA under which detained
Section 2	24.9	3.7	22.8	2.9
Section 3	43.6	58.0	46.0	56.5
Section 37	3.2	9.2	3.3	7.9
Section 37/41	11.1	17.1	11.9	17.3
Other	17.1	12.0	16.0	15.4
Referral source
Self, carer, employer, education service	2.4	1.7	2.0	2.5
LA social services	2.3	7.8	2.0	7.1
GP	15.9	6.7	14.5	6.3
Criminal justice agencies**	9.3	8.2	9.2	9.2
High security	1.2	5.4	1.2	3.8
Medium security (NHS)	1.5	8.4	1.8	6.9
Other inpatient service (NHS)	15.9	39.1	16.3	40.2
Medium security (independent)	0.6	2.6	0.8	3.0
Other inpatient service (independent)	1.3	3.4	1.4	4.8
Other (non-MH) clinical specialty	14.1	2.0	10.7	3.0
Community mental health team	24.0	13.4	26.9	11.8
A&E department	5.6	0.3	6.1	0.1
Not known	5.8	1.1	7.1	1.4
Base	28,527	3496	26,895	4292

^*All differences between NHS and independent providers are significant: p<0.001

^**Includes: police, courts, prison, probation services, court liaison and diversion services

In both years, about 80% of inpatients in NHS trusts were on general mental health wards, whereas in independent providers this proportion was just over one-third. In contrast, the proportion of patients on low/medium secure wards was significantly greater in independent providers (about 60%) than in NHS trusts (about 16%). About 3% of patients in NHS trusts were in high secure units.

In both years, the duration of stay in hospital from date of admission to census day was substantially longer for patients in independent compared with NHS providers (medians 435 and 93 days, respectively, in 2006, and 372 and 98 days, respectively, in 2007). The longer lengths of stay in independent providers were apparent for all ethnic groups.

The overall proportion of patients detained on admission increased from 40% in 2006 to 43% in 2007; increases in the proportions of patients detained were apparent in both NHS and independent providers. The majority (82% in 2006 and 79% in 2007) of patients detained on admission under the Mental Health Act were in NHS providers. However, the proportion of patients detained on admission was about 1.7 times higher in independent providers than in NHS organizations in both years.

There were also differences in the proportions of patients detained under different sections of the Mental Health Act 1983 (Table 2). ^a In both 2006 and 2007, the proportion of all compulsorily admitted patients who were detained under section 2 was much lower in independent providers than in NHS organizations, whereas detentions under sections 3, 37 and 37/41 were proportionately much higher. Sections 3, 37 and 37/41 accounted for about 84% of all detentions in both sectors in both years.

Although there were relatively small differences between the sectors in the proportion of referrals from criminal justice agencies, in both years there were significant differences in referrals from other sources. In NHS trusts, a higher proportion of all referrals were from GPs, community mental health teams, A&E departments and non-mental health specialties. In contrast, independent providers received higher proportions of referrals from social services, and about half of their intake was from NHS inpatient mental health services.

Age and gender standardized detention ratios for NHS and independent providers by ethnic group are presented in Table 3. In both 2006 and 2007, overall detention ratios were significantly higher in independent than in NHS providers by 44% and 39%, respectively. Detention ratios in independent providers were higher than in NHS providers for all ethnic groups (thoughsignificantly so only in the White British, Other White, White/Black Caribbean Mixed and Black Caribbean groups). In both 2006 and 2007, detention ratios in NHS providers were statistically significantly higher than the national average by up to one-third among the Black Caribbean, Black African and Other Black groups; no excess was apparent for other ethnic groups. Patients from the Black Caribbean, Black African and Other Black groups in independent providers also had higher than the average detention ratios in both years, by 55% to 77%. Additionally, patients in independent providers from the following ethnic groups had higher than average detention ratios in both 2006 and 2007: White British, Other White, White/Black Caribbean Mixed.

Table 3.

Standardized detention ratios by ethnic group for independent and NHS providers: detention on day of admission (England and Wales = 100)

Census categories	Census 2006 (persons)	Census 2007 (persons)
	95% confidence interval			95% confidence interval
	Standardized ratio	Lower	Upper	Observed	Standardized ratio	Lower	Upper	Observed
Independent
White British	133	126	139	1692	128	122	134	2034
White Irish	132	95	179	41	139	102	184	48
Other White	154	115	202	52	138	121	155	260
White and Black Caribbean	174	127	233	45	167	127	217	57
White and Black African	157	81	274	12	106	46	208	8
White and Asian	133	71	228	13	168	87	293	12
Other mixed	141	91	208	25	123	67	206	14
Indian	151	104	212	33	95	60	142	23
Pakistani	144	86	224	19	123	77	186	22
Bangladeshi	171	74	336	8	192	70	417	6
Other Asian	127	66	222	12	135	82	211	19
Black Caribbean	174	146	205	141	160	135	188	142
Black African	167	128	214	63	155	116	201	55
Other Black	177	101	287	16	159	111	220	36
Chinese	169	55	393	5	154	32	450	3
Other	122	72	192	18	116	71	180	20
Total	137	132	143	2195	131	127	136	2759
NHS
White British	89	87	91	7311	88	86	90	7382
White Irish	100	86	116	174	104	89	120	178
Other White	101	92	111	441	101	92	111	445
White and Black Caribbean	107	90	127	132	107	90	126	137
White and Black African	130	99	169	57	123	93	160	56
White and Asian	109	81	143	52	108	79	144	46
Other mixed	101	79	127	74	97	78	120	90
Indian	90	76	106	148	87	73	102	147
Pakistani	101	86	117	170	94	79	110	148
Bangladeshi	97	76	121	76	102	79	128	70
Other Asian	110	92	131	129	108	91	128	136
Black Caribbean	132	122	142	672	134	125	144	767
Black African	118	107	131	361	122	110	134	395
Other Black	133	120	147	366	131	118	144	377
Chinese	116	82	160	38	106	77	144	42
Other	106	90	124	155	112	96	129	188
Total	95	93	96	10,356	94	92	96	10,604
Total	100			12,551	100			13,363

Discussion

There are some limitations to the analysis presented here. First, the censuses provide a one-day snapshot and do not capture all-year inpatient activity. This could skew some of the findings towards long-stay patients and, for example, result in underestimating admissions and detentions of short duration, and overestimating the median duration of stay from admission to census day. These effects will apply to both the NHS and independent provider data, but more especially the former as the NHS has a lower proportion of long-stay patients. Second, the census does not include information on diagnosis, case-mix or socioeconomic factors, hence no adjustment to the detention ratios or other analyses was possible for these variables. Third, the reason for ethnicity being reported by staff or relatives for 25% of patients, rather than being self-assigned, is not known. It is possible these patients were not well enough to respond; the effect of this on accuracy of reporting cannot be ascertained. Finally, the detention ratios were derived using the inpatient populations as denominators rather than the general population with a mental illness, because national estimates of the latter by age, gender and ethnicity are not available. Hence the ratios measure the rate of detention among those admitted to hospital, rather than the population at risk of detention.

On the other hand, the censuses achieved comprehensive national coverage of patients in both NHS and independent providers. Although they provide only a one-day snapshot of patients in these establishments, in the absence of other information, they provide a basis for examining differences in the patient profiles of the two sectors. The findings show that patients in independent providers have a different demographic profile to patients in NHS trusts. There are also differences between the sectors in referral patterns, the proportions of patients detained on admission, the types of sections under which they are detained, the durations of stay from admission to census day and the levels of ward security. These reflect the fact that many independent providers are registered to take only patients liable to be detained, and the increase in the number of psychiatric intensive care beds (PICUs) provided by the independent sector.

A significant and growing proportion of inpatient mental healthcare is provided by independent sector establishments. This is despite the expansion of specialist community mental health services secondary to the National Service Framework for mental health. The 2007 census shows that 14% of all mental health inpatients in England and Wales on census day were in independent providers, up from 11% in 2006.^2,3 The independent sector is also a significant provider of inpatient child and adolescent mental health services.⁶ The shortfall in NHS provision of inpatient mental health services has led to sustained growth in the independent sector since the early 1990s. Mental health is the strongest growth area of independent sector hospital services, increasing by 16% in 2006 to be valued at £845 million, with NHS spending accounting for 85% of revenues.¹ Furthermore, the increase in private sector provision is largely to meet the demand for secure inpatient services, i.e. these are predominantly long-stay (often forensic) inpatients with complex mental health needs.^7–11 African-Caribbean patients are over-represented in medium secure care.⁸ The decline in NHS beds for people with mental illness over past decades may also have contributed to re-institutionalization in other forms, including within growing private sector provision of NHS-commissioned care.^9–11

Our analyses (based on a one-day count) show that almost half the intake of inpatients in independent mental health services is from NHS inpatient services. However, currently very little information is routinely available on a systematic basis for monitoring the numbers and details of patients receiving care in independent sector providers, the quality of that care and how patients move around ‘within’ the healthcare system. Further, there are concerns about the lack of collaboration between service commissioners, NHS providers and the independent sector in ensuring consistency in the quality of care and preventing patients from being lost in the commissioning system.^12,13 In a bespoke census of patients in independent sector beds in seven Strategic Health Authorities, Ryan et al. estimated the weekly cost at £2.9 million, and found that only 55% of placements were within the geographical boundaries covered by the commissioner, links with CPA care coordinators and commissioners were often weak, and service commissioners were not aware of the Mental Health Act status of 41% of patients, of diagnosis in 35% of cases and of ethnicity in 28%.¹³ One commissioner was paying for two people who had died two years earlier. Ryan et al. call for improved coordination between the independent sector, NHS provider trusts, CPA care coordinators and service commissioners, noting the inadequacy of information flows between them.

We consider that these functions should be supported and facilitated by robust, routine national information systems which are currently lacking. Currently, for example, official statistics for England do not provide the ethnic origin of detained patients.⁷ Supplementary information for mental health patients in independent sector providers currently comes from ad hoc, costly and labour-intensive one-day censuses covering these establishments, and is inadequate in terms of coverage and continuity compared with the information routinely available for patients in NHS establishments.

Data at the individual patient level, with details of age, gender, ethnicity, admission and discharge details, clinical and other information such as place of residence and commissioning primary care trust (PCT), are needed to inform service planning and commissioning and for monitoring the use and outcomes of mental healthcare. While data-sets such as Hospital Episode Statistics (HES) and the MHMDS provide patient level data for patients in NHS trusts, similar data-sets are not available for patients (NHS or private) in independent providers. Overall counts of detained patients in independent providers, with some details about patients, are available and provide a useful overview,^2–4,7,14 but they do not enable detailed analyses of the sort that patient level data can support. For example, analyses of admission and readmission rates, length of stay, place of admission and discharge, detention status and changes in it, out of area treatments, mortality by age, gender, ethnicity, provider and PCT of residence. The unique patient identifiers in such data-sets also enable patients to be ‘tracked’ as they move around the healthcare system, including both inpatient and community care.

Statutory data returns such as HES and the MHMDS, their limitations notwithstanding, are used widely and to good effect for such purposes,^15,16 but there is no contractual requirement for independent sector providers of mental healthcare to submit these data, even for their NHS-funded patients. Changes in this situation are overdue, and submission of the same data as submitted by NHS providers should be made a contractual requirement by those commissioning services for NHS-funded patients from independent providers, as recommended in the count me in census reports.^2–4 The standard contract for commissioning mental health services, currently under development by the Department of Health, provides an opportunity to make this possible.

Routine submissions of patient level data in a standardized format across all providers of inpatient mental healthcare are imperative for monitoring the placement and quality of NHS-commissioned care and patient flows on a consistent basis, irrespective of whether it is publicly or privately provided. Others have likewise called for improvements in the information that is routinely available for monitoring the care received by users of mental health services.^{9–11,15,17,18} The growth of private sector provision, the rising numbers of compulsory admissions, the Mental Health Act 2007 and equality legislation (particularly in relation to race, given the over-representation of some ethnic minority groups in inpatient mental health services), are additional reasons why information flows need to be standardized and fit-for-purpose for monitoring the impact of changing patterns in the provision of mental healthcare, particularly in relation to the acute spectrum of care. Such information is critically important for a range of audiences, in particular, for commissioners of mental health services, to support them in needs assessment and the planning, development, commissioning and monitoring of services. Such information is needed also by agencies regulating health and social care, as their remit covers both NHS and independent sector service provision.

It is therefore critically important that the Department of Health, the Information Centre for Health and Social Care, and other national agencies as appropriate, drive these changes in information, and that PCTs as commissioners enforce them robustly. Furthermore, the quality of the data flows need to be monitored routinely by the Information Centre to ensure the information is fit-for-purpose, and commissioners and providers must use the data to assure themselves of the quality of care received by their patients.

Footnotes

DECLARATIONS —

Competing interests None declared

Funding The data used for this paper are derived from two national censuses of inpatients in mental health and learning disability services funded by the Department of Health, and undertaken by the Healthcare Commission, the Mental Health Act Commission (MHAC), the National Institute for Mental Health in England (NIMHE) and the Care Services Improvement Partnership (CSIP). The views expressed are those of the authors

Ethical approval ROCR: ROCR/OR/0140/001; PIAG: PIAG 4-06(FT3)/2006;MREC 06/MRE09/51, 14 November 2007

Guarantor VSR

Contributorship VSR conceived the study, interpreted the data, drafted the paper and gave final approval of the version to be published. GMP coordinated production of the census data-set and assisted with the analyses. SB undertook the statistical analyses. SD assisted with the analyses. AD contributed to the drafting