From inter-agency to multidisciplinary work in a sector generic mental health team

The effects of establishing a full multidisciplinary community psychiatric team in central London are described. The nature of referrals, and referrers' satisfaction with the service were examined during two comparable six-month periods, before and after the changes. The number and severity of referrals increased, and referrers were more satisfied with the new service. In contrast to the inter-agency model, the new multidisciplinary team was associated with more appropriate referral of those patients with the greatest need for specialist services. Clinical skill mix in the team and referrer education are the two factors most likely to have promoted these important changes. The key role of the psychiatrist in a full multidisciplinary team who can empower multi-professional case managers in their day-to-day management of severely ill patients is highlighted.

The last forty years have seen a progressive expansion in the number and diversity of commu nity-based mental health agencies, including those provided by community psychiatric nurses, mental health social workers, and through liaison with primary care. The accessibility (Marriott et cd, 1993) and acceptability (Merson et cd, 1992) of these services may be greater than for more traditional services. However, the risk of specialist services becoming overwhelmed by the demands of mild psychiatric disorders (Donovan, 1982), while services for the severely mentally ill fragment between agencies (Ritchie et cd, 1994) are con siderable. Increasing the range of services does not mean they are necessarily more appropriate.
With these changes, professionals are looking for better ways of working together, both within and across service boundaries. Multidisciplinary working describes a model of service delivery in which health care professionals work together in a team, share a common philosophy of care and tailor their care to the needs of the individuals or population they serve (Cooper, 1990). However, little is known of the skill mix a team needs to provide the most effective care to people with severe mental health needs across community and hospital settings. Referral education as a means to promoting better use of specialist services, for instance between primary and secondary care, has been evaluated in a range of settings, and among other factors this evalua tion has focused interest on the role of referral guidelines in changing service users' behaviour (Effective Health Care, 1994).
This study examines the implications of im plementing a new model of service delivery into an existing community-based team, and considers the particular role that a psychiatrist may have in the assessment and management of a multidisciplinary team's caseload.

The study
The north Marylebone sector in central London comprises a total population of 57 400. Jarman (Jarman, 1983) indices range from 9.46 (Regent's Park) to 40.1 (Church Street) and there is a diverse mix of ethnic groups, of which 70% are white. The sector includes three hostels for homeless people, and a significant proportion of the population is not registered with a general practitioner (GP).

An inter-agency model of service [Time 1]
A team of community psychiatric nurses estab lished in 1987 developed further by appointing an occupational therapist and two social workers, eventually totalling seven case managers and one full-time administrator. The team provided an open referral system to patients over the age of 16, suffering from any mental disorder except a primary alcohol or drug problem. In the absence of the team's own psychiatrist, referrals were always under the medical care of local GPs and local hospital consultants.

The multidisciplinary model of service Â¡Time 2]
In accordance with the need to develop services, prioritising those with severe mental health problems, a full-time psychiatrist was appointed, followed by a sessional psychologist, and audit administrator.
Social workers were appointed, and a coterminous health and social service Psychiatric Bulletin (1996), 20, 345-347 Table 1. Comparison of rates of referral of severe and non-severe mental illness, and referral Global Assessment Scale (GAS) score at initial contact between Time 1 and Time 2 boundary established. The total number of case managers Increased to 10.5. Although the psy chiatrist and the psychologist had responsibility only for adult referrals, the other case managers also worked with the elderly. Referral criteria were unchanged. The team was fully trained in the use of the ICD-10 diagnostic system (World Health Organization, 1992).
The team operated a case management system, where major diagnostic and management deci sions were reached through multidisciplinary discussion at regular team meetings. Once service criteria were agreed among team members, they were disseminated to referrer groups, particularly local primary care teams. This included referral procedures and criteria, target response times, and a description of the team's new skill mix.
Data was collected retrospectively over two sixmonth periods from January to June 1992 (Time 1), and from January to June 1993 fTime 2). All referrals made to the team during these two periods were examined. Diagnoses were made in accordance with the ICD-10 diagnostic system, and general functioning at initial contact meas ured using the Global Assessment Scale (GAS; Endicott et al 1976). All referrers were sent a questionnaire asking about their satisfaction with the service (derived fromTyrer & Remington, 1979).
In evaluating the effect of changes in the service particular attention was paid to the overall demands on the service, to referrals between the ages of 16-65, and to the nature of these adult referrals. The dependent variables examined were (a) total number of referrals to the team, (b) total number of referrals per case manager, (c) severity of mental disorder of adult cases (with 'severe' defined as an organic state, manic or depressive disorder, schizophrenia, and other disorders in the schizophrenia group), (d) global functioning at initial contact, and (e) referrer satisfaction.

Inappropriate referrals
Of the 338 referrals made to the service in the two six-month periods (Time 1=125, Time 2=213) 194 were not included on the following grounds: referred to the old age service (61); moved out of the area, did not attend appointment, or referred to a non-mental health agency (97); re-referrals (36).

Findings
In total, 144 referrals were appropriate, and assessed (46 in Time 1 and 98 in Time 2). The total number of referrals per case manager was 6.5 in Time 1, and 9.3 in Time 2. The 'severity' of illness of adult cases showed a statistically significant increase in Time 2 compared with Time 1. There was a decline in the GAS scores (increasing impairment of functioning), which however did not reach statistical significance.
The response to the Referrers Satisfaction Questionnaire was small (34 respondents at Time 1 and 40 at Time 2) and made interpretation of this data difficult. However, the changes in the service were generally associated with greater satisfaction. The percentage of referrers satisfied with the speed of response of the service increased from 77% to 93% (z2=2.57,d.f.=l,NS), with the quality of care from 77% to 90%, continuity of care from 65% to 75%, and discharge procedures from 64% to 81%.

Comment
Our findings suggest the changes associated with implementing a multidisciplinary model within a sector team can enhance the effectiveness of the service provided. In this example, the number of referrals per case manager increased while remaining adequate for the service to process, there was a significant change in the diagnostic distribution of referrals, and the appropriateness of referrals increased as the proportion of those suffering from a psychotic illness grew. The scale of increase in total number of referrals exceeded that of total number of case managers, before and after the service changes. Therefore, these devel opments in themselves are unlikely to explain the shift in diagnostic distribution of new referrals. Before the multidlsciplinary model was intro duced the teams' referrals showed some evidence of 'up-market drift' (Donovan, 1982), with a preponderance of cases suffering more minor psychiatric conditions. While the number of referred patients suffering from severe mental illness increased significantly over time, global functioning scores did not. This suggests that minor illness 'in crisis' may also present with poor levels of functioning, and GAS scores when used together with diagnostic groups may prove to be useful in developing referral criteria.
It is difficult to evaluate the separate contribu tions of changes in referrer behaviour outside the team, and those of skill mix and working style within the team in promoting more appro priate referrals. Systematically developed and implemented referral guidelines have been shown to contribute to referrers' use of specialist services, but the effect of less specific measures, such as disseminating referral information de scribed here, is less clear although likely to be considerably less powerful (Effective Health Care, 1994). Within the team, considerable planning and preparation was involved in im plementing a new model of working within an established team. For instance, all members were encouraged to contribute to discussions about criteria for referral, assessment and management, and to agree priorities for their routine clinical work. The multidisciplinary model includes both the opportunity to share information and skills between members, at the same time recognising discrete skills which rest closely with one or other of the clinical dis ciplines. All case managers could work with the range of mental health problems, and were empowered in working with people presenting challenging management problems through reg ular team discussions, which always included the psychiatrist, and drew on the skills of all team members. The psychiatrist was able to provide timely advice and assessment. This was particularly appreciated in the management of people with psychotic symptoms, or behaviour which endangered themselves or others. It often enabled team members, as well as referring GPs, to develop their own management plan without relinquishing clinical responsibility for the pa tients' care. Concerns have arisen about the role of the psychiatrist as leader in teams like the one we have described (Sims, 1992). Our impression is that their clinical role is one which requires similarly energetic attention.
In summary, the introduction of a multidisciplinary model was associated with increased appropriateness of referrals to this communitybased service. It is difficult to separate the relative contribution of different factors to these impor tant changes. The systematic evaluation of team skill mix, together with the effect of referrer education generally, and service guidelines more specifically, are likely to provide fertile ground for further health services research. This will be essential in informing the development of the most cost-effective mental health care across the range of treatment settings.