Mental health policy across Western European countries is focused on providing long-term psychiatric care in community-based rather than hospital-based facilities. This policy aims to rehabilitate people with severe mental illness (
1) and to enhance their chances for social inclusion. Studies show that when people with severe mental illness reside in more domestic surroundings (a family home instead of a hospital ward), this has a normalizing effect on their daily life. For instance, they show improved social functioning (
2), are less likely to be hospitalized (
3,
4), experience a better quality of life (
4–
6), and report higher satisfaction with care (
2,
4,
7).
The Netherlands distinguishes between two types of community housing programs. The first are residential facilities, which enable people with severe mental illness to live in the community within a protected environment. The other type of program is supported independent living, which offers psychosocial support to people living independently in their own home. Both programs generally differ in social context (living in a facility with other service users versus living in one’s own home), level of staff supervision (that is, daily to 24-hour supervision versus two weekly visits), and frequency of therapeutic contact. However, earlier studies in the Netherlands and in England revealed that service users in those countries are quite similar in terms of sociodemographic and clinical characteristics (
8,
9). This issue raises the question whether they differ in other respects, such as in terms of social inclusion. Social inclusion was described by Thompson and Rowe (
10) as active participation by an individual in “mainstream social, economic, educational, recreational, and cultural resources.” Previous studies have indicated that social inclusion of people with severe mental illness is far from successful. Many spend a substantial amount of their time passively—sleeping or involved with self-care activities (
11–
13). Psychiatric community housing programs are probably one of the most important long-term interventions to take up this issue (
14,
15).
With recent budget cuts, mental health service providers need to reconsider their care provision and be more cost-effective. This entails replacing the most expensive, institutionalized type of community care (residential facilities) with a less expensive alternative (supported independent living). Given that service users themselves stress the importance of independence, autonomy, and the need to be socially included (
16–
21), such development should not necessarily be viewed in negative terms.
In 2009, the Dutch Mental Health Association stated its aim for enhancing recovery and citizenship for people with severe mental illness, of which housing is an important part. The association described related aims, such as to provide more consumer choice in housing, to shorten the care trajectory from the hospital to residential facilities and supported independent living programs, and to more effectively align the consumer’s living environment to his or her needs. This approach is in line with priorities stated in 2010 by the Federation of the European Academies of Medicine for European mental health care (
22).
In this study, we aimed to gain insight into the social inclusion of service users of psychiatric community housing programs. We compared residents with service users living independently to establish whether there is a relationship between housing program and level of social inclusion. Using service user diaries, we identified three resources of social inclusion (
23): daily activities (done alone or with others), social activities in terms of visits to and from other people, and vocational participation.
Methods
Setting
This study was part of the UTOPIA study, a large-scale cross-sectional survey investigating Dutch regional institutes for residential care (RIRCs, or Dutch acronym RIBWs). In the Netherlands, 22 RIRCs provide residential care and supported independent-living programs in the community. Residential care is allocated to people with a serious psychiatric impairment who need a protective living environment and permanent supervision. Supported independent living is allocated when someone has moderate or severe impairments that affect social independence, daily living, psychological functioning, or cognitive skills or that present moderate or severe behavioral problems but without the need for a protective living environment. All key workers are trained in the psychiatric rehabilitation approach (
24).
The number of service users of Dutch RIRCs has increased rapidly in the past decade. Between 2006 and 2009, the number of available residential beds increased from 5,752 to 8,061 (40%). This increase is not proportional to the decrease of conventional psychiatric hospital beds. This trend, also referred to as “reinstitutionalization,” is taking place in other European countries as well. It additionally comprises a rising number of forensic beds, involuntary hospital admissions, and a substantial increase in the general prison population. These constituents of reinstitutionalization suggest that the modern societal approach to risk containment may prove to be an important underlying component of this process (
25). Supported independent living programs provided care to 6,935 people in 2006, compared with 9,943 people in 2009 (a 43% increase).
Instruments
Key workers (N=75) provided sociodemographic and clinical data and an assessment of service users’ level of functioning on the Health of the Nation Outcome Scales (HoNOS) (
26) (
Table 1). The HoNOS consists of 12 domains that are rated on a scale between 0, indicating no problem, and 4, indicating a very severe problem. The total mean HoNOS score is the mean sum of these 12 scores. Intraclass correlation coefficients (ICCs) for the individual items and the total score were between .74 and .88, except for the item of aggression (ICC=.61) (
26). Service users were divided into four groups that reflected their level of functioning, according to Parabiaghi and colleagues (
27): subclinical, mild, severe, and very severe problems. Key workers also filled out the Camberwell Assessment of Need Short Appraisal Scale (CANSAS) (
28), to assess their view on service users’ needs for care. Needs are rated for 22 items on a 3-point scale: 0, no need; 1, a met need; and 2, an unmet need. Phelan and colleagues (
28) established an interrater correlation of .99 and test-retest correlation of .78 of the total number of needs.
For this study we developed a diary with the help of four service users. (A copy is available from the authors.) Each day was divided into four-hour periods: 8 a.m. to noon, noon to 4 p.m., 4 p.m. to 8 p.m., and 8 p.m. to midnight. During one week, service users were asked to record per period what they were doing, with whom (friend, relative, housemate, and so on), and where. They also assessed their satisfaction with their daily life, rating each day on a scale from 0 to 10, where 10 indicated that the day could not have been better. A mean satisfaction score for each service user was calculated. Reliability and validity data for this diary are unavailable.
With a random sample of 255 diaries, we screened for activities, visits, and vocational participation (yes or no). Activities included task-like activities and active leisure—buying groceries, being at a day center, playing sports, and so on. Visits included all visits with other people but excluding therapeutic visits. Social contacts with housemates were also scored separately. Vocational participation included paid employment, voluntary work, and sheltered or supported employment.
Sample
Sixteen RIRCs participated between October 2007 and June 2008 and provided a list of key workers. In order for us to guarantee that these workers were familiar with the organization and its methods of support, the workers had to meet the criteria of being involved in day-to-day care of service users and having employment with the RIRC for at least 24 hours per week and for more than one year (
1). We randomly selected 192 key workers, of whom 119 (about 10% of the total number of key workers and 62% of the random sample) participated. They initially approached 1,432 service users, of whom 818 service users (57%) gave written informed consent. Of the total number of participants, 677 (83%) filled out the diary. We randomly selected 255 diaries (38%) for analysis in this study. If diaries included fewer than four daily reports (N=12, 5%), we randomly replaced them. Participating service users and nonparticipating service users did not differ on gender, age, psychiatric diagnosis, the presence of substance abuse, and the presence of a personality disorder. The medical ethics committee of the University Medical Health Center in Groningen determined that the study was acceptable without an extensive formal approval procedure.
Data analysis
We used Stata/SE version 10.1 for statistical analyses. First, we conducted univariate analyses between sociodemographic, clinical, and care characteristics and the variables for social inclusion (later referred to as “social inclusion variables”), the number of reported activities, the number of reported visits (all visits made and received by service users), and vocational participation (yes or no). Second, we performed three univariate regression analyses with the social inclusion variables as the dependent variables. In these analyses, type of housing program (supported independent living versus supported housing) was the independent variable. We controlled for age; gender; education level; psychiatric diagnosis; marital status; social support from partner, family, and friends; substance use disorder; level of functioning; and unmet needs regarding social contacts. In health services research the environmental context (the RIRC) also contributes to individual outcomes (
29). We used univariate and multilevel regression analyses to explore the association of service users’ RIRC “membership” (being in the care of one of the 16 participating RIRCs) with social inclusion and to assess the influence of RIRC membership on the relationship between type of housing program and level of social inclusion.
Results
Service user characteristics
Both service user groups were similar in mean age, gender composition, civil status, and education level (
Table 1). They were comparable regarding the presence of a substance use disorder or a personality disorder, level of functioning (total mean HoNOS scores), and admissions to a psychiatric hospital in the past year. Residents were more likely to have schizophrenia, whereas service users living independently were more likely to have an anxiety disorder or depression. Concerning RIRC care, independently living service users had significantly less key worker support, and the nature of this support was more likely to be therapeutic than to be practical assistance (
Table 1). Almost nine out of ten residents were part of a shared household, with only one in 20 living with one’s family or a partner. Only a quarter of independently living service users lived with others, of whom one in two lived with family or a partner. Compared with those in independent living, residents reported having more contacts with housemates.
Social inclusion
The multilevel analysis revealed that RIRC membership was not associated with the social inclusion variables and was not a significant factor in explaining the variation of the relationship between housing program and social inclusion.
Independently living people reported more activities, alone as well as with others (
Table 2). This result remained significant in a univariate regression analysis that controlled for possible confounders (see the earlier data analysis section) and in which independent living (yes) was the independent variable; independent living explained 19% of the variance in activity level (adjusted R
2=.19, F=6.79, df=10 and 242, p<.001; β=.25, t=4.16, p<.001). Both service user groups rarely participated in sports. In regard to visiting with others, service users living independently received more visits and were also more likely to visit others. Being in a supported independent living program explained 8% of the variance in the total number of visits (adjusted R
2=.08, F=3.30, df=10 and 242, p<.001; β=.22, t=3.43, p=.001). In the area of vocational participation, we did not find significant differences between the two programs, neither in regard to the kind of vocational activity nor in hours per week spent on the activity. Of the total number of service users with vocational participation (N=87 [66%] in supported housing and N=63 [62%] in supported independent living), only about one in three had a job of 16 hours per week or more. A majority (around 40%) did not participate in vocational activities.
Both service user groups reported similar unmet needs in the areas of social inclusion, although the total number of unmet needs (in particular those reported in the area of social contacts) was significantly higher for independently living people (
Table 2). Nearly one in three service users living independently reported receiving no or insufficient support to meet their social needs, compared with slightly more than one in ten residents. Despite the relatively high need for support in the area of social contacts, more than 85% of service users reported receiving support from a partner, their family, or friends. Overall, social contacts and relationships were the areas in which unmet needs were reported the most, followed by daytime activities and paid employment.
According to their diaries, service users rated their daily life with a mean±SD score of 6.8±1.2 out of 10 (“couldn’t be better”). Service users living independently were less satisfied with their daily life than people in supported housing.
To gain more insight into the differences in social inclusion between both housing programs, which seemed to serve a similar group of users in respect to sociodemographic and clinical characteristics, we conducted the same regression analyses for service users who experienced very severe problems in functioning (
27). However, the results remained unaltered.
Discussion
In this explorative study, social inclusion was associated with the type of community housing program. We found independently living service users to be more likely than residents to be socially included, in terms of activities and visits. This finding is consistent with previous findings (
30).
A finding that seems counterintuitive is that independent living was associated with more unmet needs in the area of social inclusion and with lower satisfaction with daily life. Perhaps independently living service users, who are more likely to interact with people without mental illness in typical community surroundings, have a different frame of reference than residents and are therefore more critical about their social inclusion and daily life. In addition, participating key workers mentioned that people living independently sometimes experience independent living itself as a daily task or even as “survival” instead of “living.” Consequently, these service users may feel unable to participate in social and vocational activities. Previous studies have shown that independently living service users indeed feel more isolated and lonelier than residents (
2,
31). These service users should be identified and receive additional support for domestic activities as well as for social and vocational activities. Also, independent living does not seem to be associated with higher levels of participation in more competitive activities, such as paid employment. In both programs, only 8% of service users had a regular paid job and roughly 40% of service users had no form of vocational work. Although these numbers themselves might seem discouraging, from an international perspective Dutch service users seem to represent a relatively high level of vocational participation (
8). Yet, there are several possible, and perhaps also complementary, explanations for this finding.
First, the focus of RIRC care, concerning service users’ daily life, may still be too concentrated on aspects related to housing and “having something to do,” instead of on supporting and promoting service users to participate in the regular job market. For example, of the 105 people without any vocational participation, only 9% were in a supported employment trajectory. Second, the broader social and economic context is important. People in residential facilities in the Netherlands are required by the Exceptional Medical Expenses Act (Dutch acronym: AWBZ) to pay an income-dependent statutory contribution (ranging from €146 to €2,097 in 2011) for the care they receive. This contribution is taken from their public assistance, employee benefits, or salary. Under current rules, unmarried residents are left with a minimum of €455 in spending money per month. Service users living independently also have to pay a statutory contribution. The maximum contribution is less defined, because the level and nature of this support—and therefore its costs per individual—vary to a large extent. For both service user groups, an increase in income would result in a higher statutory contribution. This increase could mean that someone starting in paid employment must cope with the downsides of having a job (anxiety and stress, for example), without the financial gain. This policy is understandable but not vocationally stimulating. Third, the dependence on a small income is in itself a barrier for social inclusion, given the expenses that result from undertaking social activities.
Another important social issue is the fact that young people with severe mental illness are often not able to graduate from postsecondary and tertiary education as a result of their illness. Not having the right certificates or diplomas to gain access to paid employment creates an extra setback for these already vulnerable people. An intervention such as supported education (
32–
34), which supports people with severe mental illness to study at a school within the standard education system, may prove to be important in improving the educational tenure of service users. Still, research into the effects of these programs on service user outcomes is largely absent. This should be addressed in future studies.
Finally, the analyses concerning social inclusion for the subset of service users who experience very severe problems in functioning rendered the same results as for the total sample. This result is somewhat remarkable; one could argue that for this particular service user group, supported housing was probably better able to enhance social inclusion, bearing in mind the higher levels of staff supervision and professional support. Given the preference of service users to be as independent as possible and their need to be socially included, the decision to allocate people to these residential programs should therefore be very carefully considered.
A strength of our study is that we used a large, representative sample of people residing in psychiatric community housing programs in the Netherlands and were able to compare residents with those living independently. In addition, service users were involved in completing the diary to register their use of time and social inclusion. The main limitation of our study lies in its cross-sectional design. We did not conduct a follow-up to establish whether people were better off over time in one program or the other. Also, the service users were not randomly assigned to either program. We ascertained that both service user groups were quite similar, however. Another limitation concerns the diary, in that we did not have information on how and under what circumstances each service user filled out the diary during the week.
Conclusions
Compared with residential programs, supported independent living programs seemed to positively influence the level of social inclusion among service users, in terms of their being active and receiving and making visits with others. Apart from this finding, we feel that the future challenge for mental health services and associated providers lies in improving the level of vocational participation of service users in both types of program. Additional resources, such as supported education and supported employment programs, need to be made more readily available to service users to improve their chances of social inclusion.
Acknowledgments and disclosures
An unconditional grant was received from the Alliance of Dutch Regional Institutes for Residential Care.
The authors report no competing interests.