As part of a series of initiatives aimed at improving services for people with personality disorder the Department of Health funded a series of community-based pilot services that were set up in 2004/2005. ^{Reference Haigh1} In south-west London, one of these services, the Service User Network (SUN) project, was designed to offer community-based open access support groups for people with personality disorder. It aimed to help people develop effective ways of coping, reduce emergencies and improve access to appropriate services. A fundamental principle of the project was the utilisation of service users' own resources for support and a commitment to formally include them as an integral part of service delivery. Service users were employed in the design and early management of the project, and service users work within the team as both volunteers and as paid full team members.

We set out to record levels of attendance, examine characteristics of people utilising the service and explore the impact of the service on use of other medical and Social Services, social functioning and satisfaction with care. We hypothesised that contact with the SUN project would reduce use of in-patient mental health and emergency medical services and increase appropriate contacts with other services.

Method

Results

During the first 16 months of the service, 765 groups were held. A total of 171 people attended at least one group, completed a crisis and support plan, and became a member of the SUN project. Of these, 116 (67.8%) were women and 55 (32.2%) were men. Age was stated by 116 members and this ranged from 22 to 62 (mean 39.7, s.d. = 9.6). Information on ethnicity was provided by 132 people (77.2%) and 109 (63.7%) were White British, 13 (7.6%) were White ‘other’ and 10 (5.8%) were from Black and minority ethnic communities.

Of the 171 members, 142 (83.0%) returned to the group on at least one other occasion. The median number of groups attended was seven (IQR = 3.0-22.0). Attendance patterns were irregular, 51 people did not attend for a period of 1 month or more then returned to the project, and during periods of attendance the ratio of groups attended to groups not attended was 4 to 1. Men were as likely as women to attend on more than one occasion (32.4% compared with 31.0%, χ² = 0.20, P = 0.89) and members from Black and minority ethnic communities were as likely to return as those from British White backgrounds (8.0% of those who returned were from Black and minority ethnic communities as were 5.0% of those who did not). Mean age of the two groups was also similar (39.6 among those attending more than once and 41.1 among the remainder). At the time of the survey, 141 (82.5%) of the 171 had been in contact with the service for at least 6 months and were therefore sent a copy of the survey questionnaire. Of these, 61 (43.3%) completed the survey. Those who completed the survey were similar in terms of gender and ethnicity but were younger than those who did not complete it (mean age 37.0 years compared with 41.2, t = 5.31, P = 0.02) and had attended a greater number of groups (15 compared with 6, Z = −4.25, P<0.001).

Forty-eight (78.8%) of those who completed the survey filled out the abbreviated version of the Standardised Assessment of Personality. Their mean score was 5.3 (s.d. = 1.8), with 44 (91.7%) having a score of 3 or more, suggesting ‘probable personality disorder’. ^{Reference Moran, Leese, Lee, Walters, Thornicroft and Mann4}

Service utilisation

Levels of self-reported service utilisation were obtained from all 61 people who completed the survey. Changes in service utilisation in the 6 months prior to and following SUN membership are presented in Table 1. None of the 61 service users reported contacts with probation services during this period. Reductions in use of all services were seen following contact with the SUN project with the exception of use of telephone helplines, for which there was a slight increase. Reductions in contact with primary care, mental health and Social Services were greater for unplanned contacts than for planned ones. The ratio of planned to unplanned contacts increased from 5 to 1 in the 6 months prior to joining SUN to 10.6 to 1 in the 6 months after joining.

Table 1 Service utilisation in the 6 months before and after joining the Service User Network (SUN): total number of contacts among 61 service users

Service	6 months prior to SUN, n	6 months after SUN, n	Difference (pre – post), n	Difference in paired medians, Z a	P
Visits to general practitioner
Planned	265	239	−26	−0.19	0.85
Unplanned	40	14	−26	−2.29	0.02
Visits to mental health teams
Planned	244	287	43	−1.17	0.24
Unplanned	87	35	−52	−1.64	0.10
Visits to Social Services and housing
Planned	61	36	−25	−0.34	0.73
Unplanned	13	4	−9	−1.84	0.07
Total planned visits	570	562	−8	−0.54	0.59
Total unplanned visits	114	53	−61	−2.26	0.02
Calls to telephone helplines	123	136	13	0.41	0.68
Attendances at accident and emergency	55	36	−19	−0.94	0.35
In-patient admissions	46	15	−31	−2.33	0.02
Days in in-patient treatment	302	233	−69	−1.63	0.10

Social function and satisfaction with services

Although 48 (78.7%) respondents filled out part of the SFQ, only 42 (68.9%) fully completed it for both the pre- and post-treatment period. The mean SFQ score was 14.0 (s.d. = 4.4) in the period prior to joining SUN and 12.1 (s.d. = 6.0) in the period after. The difference in mean SFQ scores before and after having contact with SUN among these 42 people was 2.0 (95% CI 0.47-3.53, P = 0.01). Fifty-three (86.9%) of those who completed the survey responded to questions on satisfaction with the SUN project and the impact of the service. Their responses are presented in Table 2.

Table 2 Satisfaction with the Service User Network (SUN): responses of 53 people using the service

Statement	Mean scorea	s.d.
I feel more supported	3.4	1.1
I feel more included	3.4	1.2
I am learning more coping strategies	3.4	1.2
I feel more empoweredb	3.2	1.1
I am having fewer crises that become emergencies	3.3	1.2
My difficulties are becoming easier to manage	3.1	1.1
I am now able to access other services more simply	3.0	1.1
The description of each group is easy to understand	3.8	0.9
I feel more able to talk/express myself	3.4	1.2
I am satisfied with the SUN project	3.2	1.1

Discussion

The SUN project was set up to deliver an innovative, deinstitutionalised approach to helping people with personality disorder develop better ways of coping with crises. These data show that nearly all SUN members who returned questionnaires had probable personality disorder and had high levels of social dysfunction equivalent to those presenting to emergency services for treatment of mental disorder. ^{Reference Tyrer, Nur, Crawford, Karlsen, McLean and Rao5} Decreased levels of contact with health and social care services in the period after joining the SUN project were reported and this was particularly marked for ‘unplanned’ contacts with services and use of in-patient treatment. Higher levels of social functioning in the 6 months following contact with the service than in the prior 6 months were reported, the reduction in SFQ score being both clinically significant and larger than that reported in previous studies of out-patient psychological treatment for people with personality disorder. ^{Reference Davidson, Norrie, Tyrer, Gumley, Tata and Murray7,Reference Huband, McMurran, Evans and Dugan8}

Care needs to be taken interpreting these findings, however, because of a number of important methodological limitations. First, all the data we collected were self-reported and it is possible that response bias resulted in some or all of the differences we found. Self-report data may lead to an overdiagnosis of personality disorder when compared with structured interviewing. ^{Reference Verheul, Hartgers, van den Brink and Koeter9} Although data on service utilisation could have been cross-checked against routine records we did not have ethical committee approval to obtain these data and had to rely on self-report. The other main limitation is the response rate. Nevertheless, 43% is better than response rates for postal surveys among people with both personality disorder ^{Reference Blum, Pfohl, John, Monahan and Black10} and depression. ^{Reference Graham, Banrjee and Gill11} It may have been possible to reduce the amount of missing data by integrating the completion of questionnaires into group sessions. However, we took the view that such an approach may have undermined the group aims of trying to strengthen peer support and placing less emphasis on the role of service providers. The net result is that our data cannot be extrapolated with certainty to those not returning questionnaires.

One of the methodological strengths of the project was that we used measures that have been widely used in other studies of people with personality disorder including a national evaluation of ten other community-based services in England. Thus these findings can be compared with those from other services that delivered more traditional approaches to helping people with personality disorder. ^{Reference Crawford, Price, Gordon, Josson, Taylor and Bateman12} Age and gender of SUN members is similar to that of other people using community-based services for people with personality disorder. For instance, women made up two-thirds of all SUN members, and although men are as likely to have personality disorder as women ^{Reference Coid, Yang, Tyrer, Roberts and Ullrich13} the preponderance of women among SUN members is in keeping with those of other community-based services for people with personality disorder. ^{Reference Crawford, Price, Gordon, Josson, Taylor and Bateman12} The higher proportion of people from Black and minority ethnic communities using the SUN project reflects the multi-ethic population in the local area. Over 90% of SUN members who returned questionnaires had probable personality disorder. Although this is also true of people with personality disorder accessing other new personality disorder services, the high levels of personality disturbance and social dysfunction we found are interesting given that all those who accessed the service were self-referrals. Providers of other personality disorder services have expressed concerns that opening them up to self-referral may mean that people with low levels of need access the service, ^{Reference Crawford, Rutter, Price, Weaver, Josson and Tyrer6} but we found no evidence to support this notion in this survey.

Providers of mental health services have been told that it is unacceptable to withhold services from people on the basis of their having a personality disorder. Given that many service providers do not currently offer specialist personality disorder services, and may not have access to new funding, the development of specialist personality disorder services is a considerable challenge. General population surveys suggest that approximately 5% of people may have a personality disorder, ^{Reference Coid, Yang, Tyrer, Roberts and Ullrich13} but most interventions for people with personality disorder developed to date offer intensive services over relatively long periods of time and therefore tend to be offered to only a small minority of those that may potentially benefit from them. ^{Reference Rutter and Tyrer14} The SUN project delivers a different kind of service aimed at enhancing opportunities for peer support and helping people with personality disorder make better use of existing services. Providing open access allows people to utilise the service at a time of their own choice and broadens the scope for support to those who are only able to attend in an irregular manner. Such a service has the potential to be offered to many more people than traditional psychotherapy services.

We believe that one of the strengths of the service was that service users were involved during the development phase of the project and also contributed to the design of the patient survey. Service users were instrumental in essential aspects of the project such as publicity design, setting group rules for acceptable behaviour and liaising with other services. This generated a respectful collaboration that fostered a sense of inclusion and empowerment leading to a positive reappraisal of the resources for support that service users themselves possessed.

The exploratory nature of this study, together with the limitations in the study design mean that we cannot be certain what the true costs and benefits of this service are. However, at a cost of only £280 000 per year during the early years of its development the service is relatively inexpensive and the reductions in the use of other services by people using the SUN project during this period suggest it may be a cost-effective means of helping people with personality disorder. Future research should aim to collect longitudinal data from service users and to combine these with data from other health and social care services on actual levels of service contact. A randomised trial of a new service run elsewhere, along similar lines to the SUN project, could generate high-quality data on the effects and cost-effectiveness of support groups for people with personality disorder.