Abstract
Keywords
Although modern mental health services emphasize care in the community, acute inpatient treatment remains a vital element of the spectrum of psychiatric care [1–3]. Attention has been drawn to difficulties with low bed numbers in the face of underdeveloped community services [1, 4]. South Auckland Mental Health Service is a service at risk of bed shortage. It operates with a smaller number of acute adult beds (11.6 per 100 000 total population; recommended 12.8), of extended care beds (3 per 100 000; recommended 11.2), of 24 h nursestaffed residential places (31 per 100 000; recommended 37.3) and of community mental health service staff (22 per 100 000; recommended 42) than have been recommended [5–7].
Few projects have directly studied reasons for acute admission. Of those that have, common reasons include risk of self-neglect or self-harm [8, 9], need for a structured environment or nursing care, risk to others, physical comorbidity, poor compliance and relief for patient or carer [10–12]. Few studies have looked at alternatives to admission [4, 8, 13, 14], and these tend to be biased toward patients with longer stay. Our method identified two groups of patients: those newly admitted and those admitted for 5 weeks or longer. Our aim was to estimate reasons for admission, alternatives to admission, and reasons for inpatient stay in a government-funded service with a low bed-to-population ratio.
Method
Setting
South Auckland Health Mental Health Services provides care for a catchment area of 378 000 people. For 2001, 18% of the South Auckland Health population self-identified as Maori, 17% as Pacific Islanders, 8% as Asian, and 58% as Pakeha (European) and other ethnicities [15]. Within New Zealand, South Auckland has one of the greatest concentrations of people living in areas of marked socioeconomic deprivation [16].
Design and sample
(a) New admissions study
All consecutive admissions to the acute psychiatric inpatient unit were included from November 1999 to April 2000. There were no exclusions. Cases were ascertained by daily contacts with ward staff and review of the computerized Patient Information Management System.
(b) Continued stay study
This comprised all consumers who had an inpatient stay in the top quartile of total duration of stay. For 1999 this was 35 days or more. This was taken as the start-point for ‘Continued Stay’. On day 1 of the study, all those already admitted for exactly 35 days were included. New cases were added as existing admissions and new admissions came to meet the criteria. There were no exclusion criteria.
Sample size and data collection
The sample size was based on the precision with which we wanted to estimate variables of interest. For acute admissions: an alternative to admission of 30% ± 5%, with 95% confidence, the sample size was 196. For the continued stay, an alternative to remaining on the inpatient unit of 40% ± 10%, with 95% confidence, the sample size was 57.
One trained psychiatric research nurse administered the schedules.
Demographic, clinical data and reasons for admission was extracted from casenotes and by interview with the patient's primary nurse. Each patient's consultant completed the Alternatives to Admission and Alternatives to Continued Admission interviews. Auckland Ethics Committee approved the study (application 99/219).
Measures
The Reasons for Admission schedule was developed by Flannigan and colleagues [8] in their audit of admissions to two health districts in the inner London area, and has been subsequently used by others [13]. For this study, the list of ‘Reasons for Admission’ was rated using the medical and nursing admission notes, the notes of the Crisis Team on the day of admission (who gatekeep all admissions) and the problem summary of the admitting clinician. The 22-item checklist covers a range of reasons, for example: intensive observation; risk to self; nonconcordance with medication; and respite. These were graded: 0 = Not a contributory factor in the decision to admit; 1 = Minor contributory factor in the decision to admit, but not a factor in its own right; 2 = Major contributory factor to the decision to admit. More than one major factor or minor reason can be rated. A trained psychiatric research nurse made all ratings. Piloting included checking of inter rater reliability by the nurse and a research psychiatrist separately rating 20 sets of notes. Agreement on major reasons for admission was high (over 90%) and disagreement low (under 10%). This process was repeated halfway through data collection with very similar results.
The Alternatives to Admission schedule [13] was administered to the responsible consultant psychiatrist after admission. He/she was asked if an alternative care package would have allowed the admission to be diverted. A range of care components was presented that could be combined to specify an alternative package to admission. Options included residential facilities with different levels of support (e.g. paid home help or mental health nurse) and other inpatient facilities. Hours of care required to maintain the person in the option chosen were also recorded. Also, the community keyworker (mental health professional on whose caseload the patient is named and with responsibility for providing that patient's continuity of follow-up care) for each patient was asked an open question about alternatives to admission.
The Reasons for Continued Admission Beyond 35 Days schedule was an adaptation of the Reasons for Continued Admission schedule [8]. Additional reasons for continued length of stay were drawn from the literature and from a focus group with local staff. The schedule was then piloted by examining 10 individual case notes for the reasons for continued stay on the ward. Fifteen potential reasons (e.g. symptoms, need for close supervision, accommodation problems) were rated: 0 = Not a contributory factor to continued admission; 1 = minor contributory factor to the current need to remain on the ward but would not have counted as a factor in its own right; 2 = major reason for the current need to remain on the ward.
The Alternatives to Continued Admission schedule [13], was administered to the responsible consultant. The consultant was asked if an alternative care package would have enabled him/her to discharge the patient once they had been admitted for 35 days. A range of options was presented that could be combined to specify an alternative package to ongoing hospitalisation. See Alternatives to Admission Schedule above.
The Health of Nations Outcomes Scale (HoNOS), Version 4, measures problems such as behaviour, impairment, symptoms and social functioning [17]. The period assessed covered 2 weeks prior to admission. We omitted items 11 (occupation) and 12 (social environment) as has been suggested for the acute inpatient situation [18]. The 10-item HoNOS was used with a total score of 40. The Global Assessment of Functioning (GAF) measured overall psychological disturbance on admission [19, 20].
Deprivation was rated using the NZDep96 small-area deprivation index [21]. The index is derived from 1996 census data and includes the proportions of nine variables, reflecting material and social deprivation. The patient's residential address prior to admission was used to classify level of deprivation.
Analysis
Descriptive analysis (means, frequencies, 95% confidence intervals, etc.) was carried out using Stata Version 6 [22]. Alternatives to admissions and to continued stay were collapsed into groups containing common components of care.
Admissions group sample: the total occupied bed days consumed by each patient were used to calculate simulated occupied bed days. Bed days were assumed to have been ‘saved’ if the responsible consultant thought that there was an alternative to acute ward care at that time. Savings were assumed to begin immediately and thus equalled the admission duration. This method presents a simple summation approach for calculating potential bed-day savings [13].
Continued stay sample: the total occupied bed days from day 35 consumed by each consumer were used to calculate simulated bed-day savings if alternatives to continued stay after 35 days were suggested. Bed-day savings were assumed to begin at 35 days, and therefore comprised the admission duration minus 35 days. (Patients whose admission could have been diverted at the point of admission would not be included in this continued-stay bed saving as it was assumed that they were never admitted.)
Results
There were 255 admissions of 226 individuals in the study period. These constitute the sample for the new admissions study. These 255 admissions occupied 8101 bed-days. In the first 4 months of the study, 60 admissions of 60 individuals met the criteria of 35 days’ stay in hospital and constitute the continued stay sample.
Admissions study
As shown in Table 1, a high proportion of the sample was single, admitted involuntarily under the Mental Health Act, and coming from areas of marked or moderate social deprivation. The most frequent principal diagnoses were: schizophrenia (45%); bipolar affective disorder (24%); depression (11%); and stress/adjustment disorder (5%). Fifteen percent had other diagnoses including personality disorder, substance abuse and organic psychosis. For 27%, this was their first admission; 19% had one previous and 54% two or more.
Characteristics of the sample (n = 226)
As shown in Table 2, major reasons contributing to the decision to admit were reinstatement of medication need for intensive observation, non-concordance with medication and risk to self involving selfneglect and suicide. Sixty-two percent had 2–3 major reasons for admission and 34% had four or more.
Major contributory reasons for admission†‡
There was an alternative to admission for 31/255 admissions (12%: 95% CI, 8–17%). As shown in Table 3, the most frequent alternative suggested was living independently, either alone or with family (76%). However, nearly all of those would have needed a mental health trained nurse visiting at least once every 24 h.
Alternatives to acute psychiatric hospital stay
Twenty-four patients who could have lived independently with intensive home visiting would have saved 325 bed-days. Three who required nonacute psychiatric inpatient care went on to consume 237 days. If all the alternatives to acute-bed use had been available, 699 total bed-days would have been saved. Assuming that these were spread evenly throughout the year, this represents a saving of 3.8 bed years in 1 year.
It was possible to interview the community key workers for 126 of the 255 admissions. There was broad agreement that at the point of admission, very few admissions could have been avoided. However 38% of admissions (48/126) were rated as avoidable if an intensive early response to deterioration had been possible.
Continued Stay Study
From 1 November 1999 to end of February 2000 (4 months), 60 admissions came to meet the criteria of 35 days’ stay in hospital.
Compared to the 226 total patients admitted, a greater proportion of those staying 35 days or more were Maori (42% v. 34%), had schizophrenia (55% v. 45%), lived in areas of marked social deprivation (46% v. 40%), had poorer global function on admission (mean GAF score 28 v. 31) and had two or more previous admissions (74% v. 54%). Only the latter difference was statistically significant (χ 2 = 7.26, df = 1, p = 0.007).
Need for supervision was a major contributory reason for ongoing admission for 73% (44/60). This included monitoring of mental state, support and reassurance, giving information, assisting with self-care and helping plan and visit new accommodation. Having resistant symptoms was a major reason for 62% (37/60: 95% CI, 48–74%). This included elevated mood and prominent psychotic symptoms, with ongoing changes in management to address these. Other reasons for delayed discharge included lack of suitable accommodation (27%), risk to others (12%) and medical problems (12%).
As shown in Table 3, an alternative placement was suggested to be more appropriate than the acute inpatient ward for 15 (25%) of the 60 continued stay patients. Most of these required high level supported accommodation with trained staff available 24 h. The main reason an alternative had not been used was the lack of places available in appropriate accommodation (8/15, 53%). Day-care as an adjunctive was selected for seven of the 15 for whom an alternative to admission was identified.
The 15 patients who could have been placed alternatively went on to consume an additional 870 bed-days in 4 months. (For none of these could admission have been diverted at the point of admission.) On the assumption that these bed-days could have been saved and that they would be spread evenly through the year, this represents a saving of 7.2 bed-years in 1 year. Combining this with those whose admissions could have been diverted at admission would lead to a total saving of 11.0 bed-years in 1 year.
Discussion
This is one of a small number of studies reporting reasons for admission and alternatives to admission. Each patient's inpatient consultant rated possible alternatives to admission and to continued stay. Our assumption, similar to Beck et al. [13], was that the view of the responsible medical officer would be informative. We explored the agreement between the consultant's rating of an alternative and the community key worker's rating of an alternative but made no other systematic attempt to rate whether the alternatives suggested would be viable. Also, while residential alternatives may allow someone to be placed in the community, our method did not investigate the broader approaches needed to allow wider integration and participation in that community [23].
Our setting appears similar to other government funded (public) inpatient units servicing those aged 18–64 years in poor urban areas [4]. Most patients were unemployed males, under the age of 40 years. Sixty-five percent were living in areas of moderate to marked social deprivation. Most had a functional psychosis and were admitted involuntarily. Maori, the indigenous ethnic minority, were over-represented being 34% of the inpatients compared to 17% of the South Auckland population [15]. Since the 1970s, Maori admissions have been increasing [24]. Reasons may include greater levels of deprivation for Maori, rapid urbanization, treatment mismanagement, misdiagnosis, and differential access to community care and primary care [25].
The proportion of admissions requiring intensive nursing, and of those with a high risk of self-neglect, suicide or assault indicates the severity of those admitted. Average scores of individual HoNOS items, such as those rating aggression and psychotic symptoms, were very similar to, and even higher than, average scores for admitted patients in Inner London [26]. A higher proportion of admissions was due to reinstatement of medication than was reported in a comparable study from London [8]. Medication non-concordance was a major reason for 33% of the admissions sample. This may be related to the low ratio of community staff at South Auckland Health (22 per 100 000 total population), and to our low use of cognitivebehavioural approaches to optimize concordance [27].
At the point of admission, only 12% of admissions were felt to be avoidable through an alternative placement. This is lower than reported using identical methodology in Nottingham, UK, where 22% could have been cared for using less intensive alternatives than an acute inpatient ward [13]. This may reflect the high severity of our population, the high proportion who were single or lived alone, and the high percentage at risk to themselves or others.
Most patients had multiple reasons for continued stay on the ward at 35 days. The most frequent were the need for supervision by nursing staff, ongoing symptoms that prevented independent living, and lack of suitable accommodation. Most of those who could have been discharged at 35 days required 24 h nurse-staffed accommodation, which was not available.
Reducing length of stay
The small number of admissions (31/255) that could have been diverted, and the small sample of continuedstay patients (60), limits the precision of our estimate of saved bed-days. However, our data estimates that 3.8 bed-years per year could have been saved by implementing admissions diversion, and an additional 7.4 bedyears by discharging at 5 weeks after admission. Greater availability of 24 h staffed accommodation places could lead to considerable bed savings. Such places are likely to be especially important for the new long-stay consumers [4, 28]. The cost-effectiveness of high-level residential accommodation requires testing in the context of a prospective controlled evaluation. Other possible measures to reduce length of stay include use of protocols to ensure early use of effective treatments, increased availability of assertive community treatment, day-care and rehabilitation places, and use of integrated sectorized approaches to community care [29–31]. While all of these could potentially reduce length of stay in the acute unit, it is still not surprising that severely unwell people, living in situations of social deprivation, have a high need for care. In the current situation of a low bed-topopulation ratio, only the most unwell are gaining access to inpatient care.
Footnotes
Acknowledgements
We thank South Auckland Health Mental Health services, Tuia Services, and Faleola Pacific Services for their support and the Oakley Mental Health Research Foundation and Auckland Medical Research Foundation for financial support.