Discussion

This study has limitations in that there was a difference in data collection periods and procedure between services where collection of one data-set commenced 11 months after the rest and involved one of the investigators rather than being left with key informant nursing staff to complete. The investigation schedule was an adaptation of an older questionnaire and not a standardised instrument. Out-of-area in-patient and low secure placements were not considered and consequently it was not possible to analyse their care pathways in relation to the local rehabilitation services.

The service user group surveyed here was characterised by significant use of in-patient services over a long period, active psychiatric symptoms, poor social functioning and unsuitability for independent living, despite interventions over a prolonged period of time. Substantial numbers had recent aggression and problems anticipated by staff if discharged to less dependent settings. The survey also demonstrated the well-documented high rate of moderate or severe physical health problems in this group of service users. ^{Reference Connolly and Kelly9} Overall, the range of in-patient rehabilitation units in this survey managed high levels of risk and disability. They also provided the opportunity for service users to move on to less supported settings but with limited availability of suitable facilities creating obstacles to this end.

Some significant differences emerged between the units. Those in community units had less extensive service histories. Those in high-dependency services were more likely to be considered appropriately placed, reflecting the high risk of challenging behaviours that made less intensive settings untenable, and less likely to have an identified community team, which would have the effect of slowing referral out of the units. The individuals in high-dependency services were also more likely to have an affective psychosis, perhaps reflecting a group who are poorly engaged or whose condition is treatment resistant and who would not comply with open settings and could only be safely managed in a closed environment. The individuals in complex needs services were more likely to have severe physical health problems.

The overall findings were similar to another recent comparable survey ^{Reference Killaspy, Rambarran and Bledin6} in terms of gender balance, ethnicity, age and comorbid substance misuse but there were fewer diagnosed with schizophrenia or schizoaffective disorder and more admitted from a community setting, possibly reflecting the higher morbidity of the inner-city London context of that service. Comparing long-term in-patient with short-term in-patient groups showed older age, longer contact with services and poorer social functioning in both studies. In our study more of the long-term in-patient group were detained but lengths of stay were lower, perhaps reflecting fewer patients who had come from long-stay mental hospital wards.

The findings support the principle that different types of rehabilitation services are required. In particular, provision is clearly warranted for those with more marked, challenging or risk behaviours, many of whom continue to be detained under the Mental Health Act 1983. The large number of service users remaining in these services due to their risk potential if discharged highlighted the need for comprehensive risk assessment to facilitate therapeutic risk-taking as part of the process of rehabilitation.

The findings indicated two areas of unmet needs in users of rehabilitation services. The first was difficulties moving on when considered no longer appropriate for their current environment. The most common reason given for this was a lack of suitable move-on accommodation, a more frequent reason than unavailable funding. This highlighted the serious lack of a range of appropriate residential settings being readily available, particularly relevant for the long-stay service user group. Similar results relating to the problems accessing appropriate move-on accommodation have been found among long-stay patients in acute admission wards. ^{Reference Commander and Rooprai10} It appears that rehabilitation and acute services are attempting to access the same limited resources.

The greatest need was for supported residential settings reflecting less intensive needs, particularly for the community-unit service users. However, there was also a substantial requirement for ongoing intensive in-patient settings in the long-term service user group. Those in the high-dependency settings were the most likely to be considered appropriately placed, probably because this setting most clearly provided for the management of their enduring risk behaviours. Additionally, nearly half had no community team involvement, highlighting the risk of disconnection from community services, particularly for those in high-dependency services. This may result in longer than necessary stays since community care coordinators have a vital role in moving on, especially when moving from in-patient to supported or independent settings.

The second area of poorly met need was the availability of appropriate services for high levels of disability. This can be seen in the fact that a quarter were service users of assertive outreach teams, indicating that even the most highly resourced and intensive of community services for severe and enduring mental illness was still insufficient to maintain a substantial minority of such service users without residential services. Some service users were simply too disabled to be supported in the community. This may be related to the decline in rehabilitation services that took place in order to create assertive outreach teams in particular, with a corresponding focus away from reducing disability to management of risk ^{Reference Holloway1,Reference Mountain, Killaspy and Holloway11} and loss of close links with residential care services.

Commissioning strategies need to take account of a ‘total system’ approach in order to facilitate service users being able to move through the various types and stages of rehabilitation necessary to gain an increased degree of independence. ^{Reference Macpherson, Shepherd and Edwards12} It has also been recommended that commissioners need to look towards developing specialised services nearer to home for service users, which could provide better care at a lower cost. ^{Reference Ryan, Hatfield, Sharma, Simpson and McIntyre13,14} Knowledge of the required types of rehabilitation in-patient services, and the corresponding size of population likely to need each type, ^{Reference Wolfson, Holloway and Killaspy2} is of critical importance to the collaborative working of commissioners and providers. Whereas only the very largest NHS trusts will be able to provide rehabilitation in-patient facilities up to the full complement of services advocated, ^{Reference Wolfson, Holloway and Killaspy2} smaller NHS trusts will need to orchestrate effective working between commissioners and providers in order to be able to access a similar range of services locally. Groups of neighbouring commissioners will increasingly need to work collaboratively in order to develop more local and specialised rehabilitation services.