Introduction

It will be remembered from the previous chapter that there were two broad aims to this stream of work. These involved the investigation of patterns of variation in CMHTsOP and the examination of variations in outcomes associated with different patterns of working, reflecting degrees of integration and different consultant roles within teams. This chapter outlines the findings from a national survey, which identified the prevalence of different ways of working and provided a national sampling frame from which to select the sites for more detailed investigation.

Findings

Questionnaires were received from 376 teams, a response rate of 87.7%. At least one response was received from 67 trusts, representing just over 93% of organisations providing CMHTOP services as reported on the 2008 Mapping Framework.²²⁶ Almost three-quarters (74%) of questionnaires were completed by team managers, 15% by service or locality managers, and the remainder by team members. As not all questions were applicable to, or answered by, all respondents, figures are given as a percentage of those participants who responded to each individual question.

Team characteristics and composition

The data collected on CMHT characteristics revealed variation in what, exactly, constituted ‘a team’. Although the majority of respondents reported that their CMHT was a single and distinct team, just under 10% reported that the team was amalgamated with a memory clinic, home treatment team, a specialist outreach service, or had other broader remits than would be expected of a CMHTOP. Just over a half of teams (56%) organised their staff around geographical patches or GP practices, with only around 3% organised according to type of illness. Most teams (71%) worked within a single LA; 17% worked with two LAs; and 12% operated across three or more LA boundaries. Just under half (47%) of teams described the community they served as ‘mixed urban/rural’, with 38% being mainly urban and the remainder (15%) being rural. Most CMHTs (72%) had been in operation for more than 5 years, with just 9% being new teams (operating for under 2 years). However, this does not mean that CMHTs were stable: almost two-thirds (63%) of teams reported major changes to the structure and organisation of the team within the preceding year.

Table 42 depicts the reported location of the team base. The most common team base was a community mental health centre, followed by psychiatric or general hospital sites. The remaining teams were situated in a diverse range of locations, including GP surgeries; LA settings (ranging from social services offices to day centres); community hospitals; and other buildings such as high street offices and business parks. Just over 10% of teams had more than one office base.

TABLE 42

Location of team base

The survey collected detailed information with respect to team composition. Particular attention was paid to the ‘type’ of team membership, by distinguishing between ‘core’ and ‘sessional’ team membership.^253,254 Core team members were defined as devoting a substantial proportion of their working week to the CMHT, contrasted against sessional members who dedicated a regular but more limited input. The data revealed substantial variation in team size, with a median of 16 core members (excluding administrative staff) and a range of 1–47, with 26% of teams comprising 10 or fewer core staff members, and a similar proportion again (24%) having 20 or more.

Table 43 shows team membership by staff group from the present survey alongside comparable data from 2004.⁷³ Professional disciplines that were not formally part of the team (the first two columns) were recorded either as being accessed separately (e.g. via referral to another service – column 3), or as unavailable in the area (column 4). The table shows evidence of a greater representation of all staff groups within CMHTsOP between the two surveys, with greater ‘core’ membership, and reduced proportions being accessed outside the team or entirely absent. However, by contrast, it was noteworthy that consultants appeared less likely to be core team members in 2009. Social work and psychology continue to be the most challenging professions to integrate within teams, with around one-third and one-quarter of CMHTs reporting that they did not have social workers and psychologists respectively as team members. Just under 1 in 10 teams were unable to access psychology services at all. The data also revealed that 75% of team managers had a nursing background, 20% a social work background, with the remainder being OTs.

TABLE 43

Community mental health team membership by staff group in 2009 compared to 2004 (% reporting each category)

Referrals, assessment and outreach

Community mental health teams for older people received, on average, 36 new referrals per month, with significant variation between teams as depicted in Figure 12. High referral numbers were particularly associated with larger teams, and teams combining CMHT functions with other services (e.g. memory services). Almost all teams used a SPA and 80% of teams had formal referral criteria. Table 44 presents the source of referrals, as reported by respondents. As expected, GPs were the primary source of referrals, although psychiatrists, social workers and care homes provided a ‘large proportion’ of referrals for 20%, 15% and 13% of teams respectively. Self-referrals were accepted by about half the teams, contributing relatively small numbers in almost all cases. ‘Other’ sources of referrals included general hospitals and other health services.

FIGURE 12

Distribution of referrals received per month.

TABLE 44

Source of referrals

Just under two-thirds (63%) of teams made their first contact with service users within 2 weeks of the date of referral (for routine cases), on average. For just under one-third (31%) this took up to a month, and for 6% even longer. For almost all teams (93%), the initial assessment was conducted in the service user’s home. Consultants and CMHNs conducted assessments in almost all teams. In addition, OTs conducted assessments in 81% of teams; social workers in 79% of teams; and psychologists in 62% of teams that had these professionals within them. All teams used some form of a key worker/case co-ordinator system, though only 60% used this for all or most clients, and for co-ordinating care between agencies. CMHNs, OTs and social workers were the most common professional disciplines acting as key workers. In 69% of teams the consultant also acted as a key worker; and in 58% of teams psychologists performed this task. Over half of respondents (58%) reported that a consultant psychiatrist would have seen ‘all or most’ patients on the team caseload, while 42% reported that they would have seen only ‘some’.

Almost all teams (97%) reported being involved in at least some liaison and outreach work. One-third of teams (34%) reported having a link worker system in care homes, and just under two-thirds (61%) reported that they provided education or training to care home workers. Just under one-quarter of teams reported that they had a link worker system in GP surgeries and general hospitals, whereas 18% conducted education and training in these settings. Other forms of outreach work, such as open clinics and case finding and screening, were rare.

Team integration and joint working

Table 45 summarises the data collected on the nine indicators of integration, introduced in Chapter 7. It shows that some aspects of integration (a SPA, colocation, single joint care plans) were features of over 80% of CMHTs. Over two-thirds of teams also used the same assessment documentation between all professional disciplines, a marked increase from a little over one-third of teams in 2004.⁷³ Sixty per cent of teams were regarded as multidisciplinary (defined as having at least a social worker and two health workers as core team members).^25,52,181 Half of CMHTs also reported that all core members were directly line managed within the team. Fewer than one-third of teams (32%) reported that their CMHTOP and social services teams shared service user records, and just 57 teams (15%) stated that health staff within the CMHTOP were able to commission social care services directly.

TABLE 45

Indicators of team integration

A composite integration score was constructed as a simple count of the number of the nine indicators (see Table 45) present in each team (following Reilly et al.;²⁵⁵ Tucker et al.¹⁸⁰). The mean score across all teams was 5.45, ranging from 1 to 9. Regression analysis explored the association between the integration score and a range of team characteristics (Table 46). Teams managed by nurses were less well integrated than teams led by managers from other professional disciplines, and teams with larger numbers of referrals per team member had lower integration scores (although this relationship weakened as referrals per team member increased, as demonstrated by the positive coefficient on the squared term). CMHTs in rural communities were found to be less well integrated than teams that served urban or ‘mixed’ populations, although this was on the very fringes of significance even at the 10% level. Further investigation showed that this was largely driven by the fact that rural teams were the least likely to have had all their staff sharing a single base, as might well be expected. Finally, the results showed that teams working with just one LA were more integrated, on average. However, a subset of these teams, those that operated in formally integrated health and social care trusts, was found to have particularly low scores.

TABLE 46

Team characteristics predicting integration

Conclusions

These findings offer a unique and timely investigation of the breadth and depth of integration in CMHTOPs, and, importantly, a high response rate (88%) gives confidence both in the representativeness of the findings and in the robustness of the statistical tests performed. However the results need to be interpreted in the context of the survey’s design and implementation. First, this study sought the views of team leaders which provides just one interpretation of the team’s working; in particular they may differ from the perspectives of consultants that have responded to previous team surveys.^73,179 Second, it is possible that some respondents may have been less circumspect than others in saying they had particular practices in place. For example, although nearly one-third of respondents reported that they could access LA service user records, contextual information provided in freetext form sometimes suggested that this was only after formal requests had been made. This contrasts with other teams that said they had direct electronic access from their desktops. It is also important to reflect on how this study conceptualised and measured ‘integration’. First, the analysis focused on a relatively narrow concept of operational integration, primarily across health and social care boundaries and professional disciplines, and at team level. As will be clear from the literature outlined above, this is distinct from broader concepts of joint working that may consider cultural aspects or integration at a ‘macro’ organisational level. Second, although the nine indicators of integration are found in key policy documentation and supported by professional consensus, there is nonetheless little evidence from the literature to validate each as critical to patient outcomes.¹⁸³ Third, the composite integration score gives equal weight to each indicator, but an alternative approach could impose differential weighting based on perceptions of each indicator’s relative importance (see, for example, Healthcare Commission⁵⁰). However without evidence linking these to service user outcomes such an approach remains highly subjective. A recent international review concluded that 24 different approaches to measuring health-care integration have been formulated, but a well-established technique has yet to emerge.²⁵⁶

One aim of this study was to assess the extent of service integration using a set of key indicators. Although past research has highlighted access to social workers and psychologists as of particular concern, our survey suggests that some progress has been made. For example, whereas surveys undertaken in 2000 and 2004 reported that approximately half of teams contained core social workers, and about one-third had core psychologists,^73,179 our data suggest these proportions had increased to two-thirds and a half, respectively, by the time of the survey. Moreover, the proportions of teams reporting that they had no access to psychology services stood at just 10%, down from a reported 18% in 2004.⁷³ However, as the National Audit Office (NAO) note, having access to specific staff groups is not equivalent to having access in sufficient numbers.⁵² Nonetheless, it appeared that a lower proportion of consultant psychiatrists were core members of teams than had been the case previously. Whether this reflects differing degrees of engagement of consultants in CMHTsOP or in part reflects the different questionnaire respondents is unclear. Nonetheless, there is some suggestion that team processes and procedures in this study were more integrated than they were in earlier research. Eighty-eight per cent of teams in this study had a SPA, up from 60% in 2004; while 70% used the same assessment documentation, compared to a little over one-third previously.⁷³ The proportion of teams that reported using single care plans is also higher than the earlier studies discussed. However the presence of some standards of integration was more infrequent, especially those that require greater degrees of co-operation and trust between health and social services at the agency (as opposed to practitioner) level. Indeed, just 32% of CMHTsOP reported that they, and social service teams, were able to access each other’s service user records. Furthermore, health staff could directly commission LA services in only 15% of teams. This latter finding chimes with a recent study of six mental health trusts which found only one example of a CMHTOP able to commission social care.⁵⁰ This is perhaps disappointing given the repeated policy efforts in England to encourage interagency commissioning as noted above, including measures such as pooled budgets and other partnership arrangements.²⁵⁷

Although it would seem that joint working has increased since 2000, it is helpful to consider the characteristics of teams that did not score well on integration. The analysis shows that teams that worked with more than one LA tended to be less integrated, echoing the Healthcare Commission findings that it is easier to make joint working arrangements with one authority than multiple arrangements with several.⁵⁰ However, contrary to their suggestions, the length of time that the team had existed was not found to be a key determinant of closer joint working in this study. The finding that teams operating in a formally integrated care trust had lower levels of integration, after controlling for other factors, is particularly noteworthy. It may be that the initial decision to form closer structural ties across trust and LA boundaries was itself influenced by previous difficulties in joint working, as has been reported elsewhere.²⁵⁸ If so, these results would support suggestions that organisational restructuring alone is not a sufficient condition for overcoming barriers to developing integrated practice.^{179,255,257,259}

A wide range of studies on service integration have suggested that organisational change and instability can hinder joint working,^50,225,260 but this is not supported by our analysis which found that teams that had faced major organisational change in the preceding 12 months were no more or less integrated than other teams. Features that seemed to be more important included the location of the team base and the discipline of the team manager, although, interestingly, no previous work seems to have considered these. Our study found that teams based in psychiatric hospitals tended to be less well integrated than those in community mental health centres or in social service buildings. Furthermore, teams managed by nurses had lower scores than teams led by an OT or social worker. Together, these findings suggest that traditional models of CMHTs (nurse-led, hospital-based) seem less likely to incorporate joint working practices.

The negative association between the level of referrals per core team member and integration is harder to explain. It may be that such teams were understaffed (teams with high referrals also had particularly high numbers of vacancies) or were particularly busy and so had less time to devote to collaborative work, as has been suggested elsewhere.^253,260 An alternative explanation is that more integrated teams had clearer operational policies and eligibility criteria which limited referral numbers. Further work is needed to validate and better understand this finding, and also to investigate the other factors associated with effective joint working across professional disciplines and organisational boundaries; including less tangible contextual, cultural and political factors. Perhaps more importantly still, we do not yet know what particular features, or combinations of features, of integration are associated with better outcomes for older people with mental health problems.