Abstract
The Ten Year Mental Health Plan for Queensland (1996–2006) [1] provides a comprehensive framework for the ongoing reform of mental health services in that state. A significant component of the reform process involves the downsizing of the three stand-alone psychiatric hospitals and the development of community-based alternatives closer to where people live. Acute inpatient units have been constructed at general hospitals and extended care inpatient programs have been established across Queensland. While the decentralization of services was a primary objective of the reform policy, there was also a clear focus on community integration for those residents discharged from the three institutions. At a service level, the reforms touched the lives of a large number of patients who had to be relocated from the psychiatric hospitals to enable beds and programs to be closed. The present study describes the outcomes for 60 elderly patients who were transferred to seven extended care psychogeriatric units.
A growing literature supports the downsizing of psychiatric hospitals and the transfer of patients and resources to smaller community-based facilities [2–5]. These smaller community facilities tend to be less restrictive and offer greater opportunity for staff–patient contact, community integration and rehabilitation [4], [5]. The stimulation and interaction provided in these smaller units are likely to slow deterioration in cognitive functioning [4] and improve quality of life [6]. Trieman et al [4] demonstrated that patients who moved to communitybased facilities had better cognitive functioning and less deterioration in behaviour at three years post-move compared to a similar group who remained in hospital.
Relocation may also have negative consequences for some patients. While symptoms and aggressive behaviour have been found to increase immediately following relocation, these tend to dissipate in the days following relocation as residents become familiar with their new environment [7–10]. A more serious outcome of relocation is the increased risk of mortality [11]. Mortality appears to be associated with cognitive functioning, in that mortality rates tend to be higher in moderately competent individuals [12]. Those at either extreme of the functioning continuum are less likely to be influenced by environmental changes arising from relocation [12]. This ability to understand the relocation process is believed to assist patients in the development of coping strategies and thereby reduce the negative consequences of relocation [13]. Moreover, there is evidence to suggest that successful relocations are associated with adequate preparation of patients prior to relocation [7], [13]. Engaging patients in the relocation process and having a well-organized relocation plan are likely to reduce the negative impact of relocation.
Although a number of recent Australian studies document the effects of interagency relocation on the lives of people with mental illness, much of this work has concentrated on patients under 65 years [3], [5], [7]. It would appear that most of the patients in these studies were cognitively intact at the time of relocation and therefore in a position to assess and evaluate the experience of being transferred from one facility to another. The closure of the aged care program at a psychiatric hospital provided a rare opportunity to study the impact of relocation on elderly patients with mental health problems.
Method
Participants
All elderly patients (n = 60) at the psychiatric hospital were relocated to enable the aged care program to be closed. The patients were relocated over a two-year period as the extended care units (ECUs) were opened. Patients were selected for a given ECU on the basis that they had family living in the region or they themselves had once lived in that region. Of the 60 patients relocated, 34 (57%) were male and the group ranged in age from 63 to 95 years (mean = 75, SD = 7.87). Some 56% had five or more admissions to the facility and the duration of time spent in the facility ranged from 9 months to 43 years (mean = 13.1, SD = 8.4). The most common DSM-IV diagnosis was schizophrenia (53.3%), followed by dementia (20%), mood disorders (13.3%), Korsakov's syndrome (10%), and other psychiatric conditions (3.4%).
Setting
The patients who moved to the ECUs were selected from four separate wards at the psychiatric hospital. While every attempt was made to ensure that care was of the highest standard in the psychiatric hospital, the wards retained many institutional features such as isolation from mainstream society, large open day rooms and dormitories, and a general lack of personal space. ECUs, in comparison, range in size from eight to 22 beds and are located alongside generic aged care facilities or general hospitals. Most have single rooms with ensuite bathrooms and features that aim to improve care and maintain independence. The ECUs are staffed on a 24-hour basis by multidisciplinary clinical teams. While the ECUs do provide long-term care, there is a strong focus on preparing patients for more independent living (hostels, own home, etc.).
The activities organized to reduce the impact of relocation on the patient group are described in detail elsewhere [14]. Briefly, a ‘transition team’ comprising staff from the psychiatric hospital and the ECUs was established to develop a relocation plan for each patient. The plan contained a detailed summary of past care and current medical and treatment needs. Medical and nursing staff from the ECUs visited patients in the psychiatric hospital on a number of occasions in the three-month period prior to the move. Patients were transferred individually by a staff member familiar to the patient. Nursing staff from the psychiatric hospital were seconded to the ECUs for one week post-move to assist in the transfer of care.
Data collection
The patients were assessed on five occasions, 6 months and 6 weeks pre-move, and again at 6 weeks, 6 months and 18 months post-move. Clinical and background information was collected on each patient at the initial assessment. The Revised Elderly Persons Disability Scale (REPDS) [15] was used to assess functioning in eight domains: communication; physical problems; self-help problems; confusion; behaviour; sociability; psychiatric observations; and nursing care dependency. Activities in all domains are rated on a four-point scale to indicate the level of assistance required by the individual to perform a given activity (0 = ‘no assistance’ to 3 = ‘full assistance’). Thus higher scores in each domain indicate greater disability (i.e. need for more assistance). The Cohen-Mansfield Agitation Inventory [16] was used to rate levels of agitation/aggression. The instrument provides staff ratings of observable behaviours which can be collapsed to three subscales: (i) physically aggressive behaviour (score range = 10–70); (ii) physically non-aggressive behaviour (score range = 7–49); and (iii) verbally agitated behaviour (score range = 4–28). Again, higher scores represent more disturbed behaviour. The impact of relocation on orientation and cognitive functioning was assessed by a psychologist (research assistant) using the Short Portable Mental Status Questionnaire (SPMSQ) [17]. Scores range from 0–10 with a higher score indicating greater impairment.
Two scales derived from the work of Leff et al. [2] were employed to assess social contact and community involvement. The social contact scale has four items (score range = 0–24) and assesses the level of contact that patients have with relatives and friends. The community involvement scale contains 10 items (score range = 0–60) and assesses the degree of involvement that patients have in their local community. Both scales are scored on a 7-point scale (0–6) with higher scores indicating more favourable conditions (i.e. more social contact, greater community involvement). Finally, the care practices in both the old and new environments were assessed via the Hospital and Hostel Practices Profile [18]. Six domains of unit functioning are assessed: restrictiveness; possessions; provision of meals; health; rooms; and services. Favourable practices are rated 0 and negative practices (e.g. patient rooms locked during the day) are rated ‘1’. Thus, higher scores suggest greater regimentation in a given unit.
Data management/analyses
The number of patients available at each assessment time-point decreased as the project progressed due to the death/discharge of patients. Rather than impute values for missing cases we employed the mixed modelling procedure (‘proc mixed’) available in SAS/STAT (SAS, 2001) [19]. This procedure permits all available data to be used at any time-point, rather than reducing the sample size to that for data available at all time-points. To reduce the imbalance in sample sizes between epochs, we ensured that the samples included in the analysis contained a pre-location measure, and at least one post-location measure (since we were comparing pre- and post-relocation functioning). Table 1 shows the achieved sample sizes. The F-test used the Hotelling-Lawley-McKeon statistic (McKeon, 1974) [20] which performs better in small samples than the typical repeated measures F-statistic (Wright, 1994) [21].
Results
While 60 elderly patients were relocated, 13 of these (21.6%) had died prior to the final assessment at 18 months post-discharge (11 deceased by 12 months). Although those who died tended to be older than those who survived (77.30 years vs. 73.34 years), the difference was not significant at p = 0.05. The majority died from medical conditions such as: cardiac problems (5); carcinoma (3); respiratory conditions (2); and renal failure (1). The cause of death in the two remaining patients remains unclear as findings from a Coroner's inquiry are still pending. In addition, eight patients had been transferred to other accommodation options such as: nursing homes (4); hostels (3); and home (1). However, we managed to collect final assessment data on four of these patients as they were discharged from the ECUs in the 4-week period prior to the final assessment at 18 months post-discharge.
Following relocation, there were significant changes in three of the REPDS domains: communication, physical health and sociability. Scores on the communication subscale increased significantly (i.e. more problems with communication) immediately following relocation (F4,56 = 5.19, p < 0.002) but returned to pre-move levels by 6 months post-relocation (Table 1). In addition, physical health problems gradually increased (higher scores) from time 1 through to time 5, to reach significance by 18 months post-relocation (F4,56 = 6.4, p < 0.001). Finally, significant changes in sociability were observed around the time of relocation (F4,56 = 8.6, p < 0.001). While patients appear to have had more severe problems in this domain on the assessments at 6 weeks prior to and following the move, these problems improved over the following months.
Finally, there were significant improvements in both social contact (F4,51 = 2.9, p = 0.05) and community access (F4,51 = 15.2, p < 0.001) for patients following relocation. This was supported by data from the Hospital and Hostel Practices Profile which indicated that the practices adopted by the new units were significantly less restrictive and controlling than the psychiatric hospital wards (χ2 = 5.15, p = 0.025). The hospital wards from which patients had been discharged had a mean of 28 (range = 17–34) restrictive practices out of a possible 48, whereas the new ECUs had a mean of 15 (range = 11−19).
Discussion
The closure of the aged care program at a large psychiatric hospital presented a rare opportunity to study the impact of relocation on a group of elderly patients who were moved to smaller community-based facilities. Although the lack of a comparison/control group presents a limitation, the collection of data on two occasions prior to relocation provides a solid base from which to assess change in the group post-relocation. While attrition due to death and discharges reduced the sample size over time, the use of the mixed modelling procedure for data analysis enabled us to include to the maximum extent possible, information collected at each assessment. Moreover, our analysis indicated that those who were lost to follow-up were not significantly different from those who remained in the study on measures of age, functioning and time in hospital.
Any study comparing hospital and community care options demands that the community options be substantially different from the hospital environment. It is clear that the hospital wards in this study were isolated from community amenities and retained many institutional features. Our data reveals that the new ECUs had significantly fewer restrictive practices when compared to the old hospital wards (11 vs. 28). The significant increase in community access following relocation is likely to be a direct consequence of the greater independence afforded patients in the new ECU environment. The location of ECUs close to shops and activities facilitated this increase in community-based activities. Notwithstanding this, the overall scores for both community access and social contact remain at the lower end of the scale, which suggests that there is room for improvement in these domains.
The significant decline in communication observed immediately following relocation returned to pre-move levels by 6 months. While a decline in functioning soon after relocation has been identified in previous research [8], it is likely that the decline in communication identified here is related to staff unfamiliarity with the patients rather than actual deterioration in communication. Given that immediately following relocation most of the staff in the ECUs were unfamiliar with the patients and their strategies for communication, staff may have given the patients a lower rating on the communication subscale.
It is clear that the move was a source of stress for some patients. The significant increase in problems with social functioning around the time of relocation (6 weeks preand post-relocation) supports this observation. However, in contrast to previous studies [9], [22] this appears to be have been short-lived and social functioning actually improved in the months following relocation. The preparatory planning for relocation and the support provided to staff and patients following relocation are likely to have offset the negative consequences of relocation [7].
While the physical health of the group was in decline prior to relocation, this continued over the study period to reach significance by the final assessment at 18 months post-discharge. In effect, four patients were further relocated to nursing homes and 13 others had died in the 18 months post-move. It is likely that the ongoing decline in physical health was an artefact of advanced age in the patients rather than the stress of relocation. Although the death of 13 patients (18.3% by 12 months) may appear high, it is similar to the mortality rate for patients in the wards from which the patients were discharged. In the five-year period prior to the project the mortality rate in these wards was 18.5% per annum. Comparison of mortality and morbidity across similar studies is difficult as participants differ in health status, age, psychological state and level of functioning. However, a recent review of 78 studies of relocated elderly patients identified mortality rates ranging from 3.7% to 48.6% at 12 months post-relocation [6].
Changes in measures pre- and post-relocation (means, standard deviations, test statistics)
Conclusion
The re-provision of services for the elderly patients described in this study did not appear to have a negative impact on the social, physical and cognitive functioning of the group. Although 21.6% of the patients were deceased by 18 months, this does not appear to be outside the expected mortality range for similar cohorts. The success of relocation for any group of individuals is likely to result from adequate preparation prior to relocation and the provision of appropriate accommodation following relocation. While the decentralization of services away from the institutions has been achieved, the findings from this study suggest that the integration of people into the broader community will continue to pose a challenge for service providers.
Footnotes
Acknowledgements
We thank David Chant (Queensland Centre for Mental Health Research) for assistance with data analysis and Queensland Health for funding of the evaluation.