Predictors of home discharge among patients hospitalized for behavioural and psychological symptoms of dementia

The Japanese government recently announced the ‘Five‐Year Plan for Promotion of Measures Against Dementia (Orange Plan)’ to promote people with dementia living in their communities. To achieve this, it is imperative that patients hospitalized with behavioural and psychological symptoms of dementia (BPSD) are helped to return to their own homes. The aim of the present study was to identify predictors of home discharge among patients hospitalized for BPSD.


INTRODUCTION
Japan has the highest life expectancy in the world, and this increased longevity has resulted in an increase in the number of people with dementia among the total population. [1][2][3][4] Despite efforts such as the Long-Term Care Insurance programme, the number of patients with dementia hospitalized in psychiatric hospitals continues to rise within Japan, with >50 000 patients currently hospitalized. 5 The primary reason for hospitalization of dementia patients is behavioural and psychological symptoms of dementia (BPSD), which manifest in 90% of those with demen-tia. 6 The goal of inpatient treatment is to achieve prompt remission of BPSD and return patients home as soon as possible. However, family members often disapprove of a patient's discharge, even after symptom remission, leading to prolonged hospitalization, with the national average now being 944 days. 7 Some of these patients are eventually transferred to a medical hospital, and others are reluctantly institutionalized into care facilities.
In September 2012, the Japanese government announced the 'Five-Year Plan for Promotion of Measures Against Dementia (Orange Plan)' not only to improve patients' quality of life, 8 but also to reduce medical expenditures. The policy aims to promote the independent living of demented people to enable them to remain in their own home for as long as possible, 9 with assistance provided by enhanced coordination between health-care services and social care services. 10 In this situation, it becomes more crucial to facilitate the home discharge (HD) of patients hospitalized for BPSD. As such, optimal treatment and care strategies should be developed from the early stages of hospitalization for individual patients based on their characteristics. However, the exact factors that may affect HD remain unclear.
As reviewed by Luppa et al., 11 studies in other countries have investigated predictors for institutionalization within community-dwelling populations. However, the findings from those few studies of hospitalized populations may not be directly applicable to Japan because of differences in both cultural background and the health-care insurance system. 12 Thus, the situation in Japan requires domestic research, but studies of this nature are scarce, with only one study having investigated predictors for HD in hospitalized patients with significant BPSD under Japan's current Long-Term Care Insurance system. 13 Previously, we demonstrated that patients' clinical and environmental characteristics at the time of admission are closely associated with time to discharge. 14 In the present study, we examined the frequency of HD in relation to patient profiles at the time of admission and determined factors that could be used to predict HD.

METHODS
In the present study we reviewed the medical records of patients who were hospitalized in the acute psychogeriatric ward of Ishikawa Prefectural Takamatsu Hospital. Patient anonymity was carefully preserved. This retrospective study was approved by the Ethics Committee of Ishikawa Prefectural Takamatsu Hospital.

Participants
Consecutive patients who were admitted to the acute psychogeriatric ward of Ishikawa Prefectural Takamatsu Hospital from their own home for the treatment of BPSD from April 2006 to March 2011 and completed inpatient treatment were enrolled in the study. All patients had severe BPSD such that they could not be cared for in their own home or be treated in an outpatient setting. Patients with severe physical comorbidities were judged ineligible for hospitalization in the acute psychogeriatric ward, so that medical treatment for their physical condition(s) could be prioritized. Patients were excluded from this study if they met any of the following criteria: (i) lived in care facilities prior to admission; (2) were patients of a medical hospital prior to admission; (iii) had a record of past hospitalization in the acute psychogeriatric ward of Ishikawa Prefectural Takamatsu Hospital; (iv) had behavioural symptoms prior to their cognitive decline; or (v) had psychiatric comorbidities.

Measures
Six factors (demography, reason for admission, type of dementia, BPSD, cognitive function, and functional state of daily living) were evaluated and recorded in all patients within 1 week of admission according to institutional protocols.

Demography
Information regarding a patient's age, gender, and living situation (i.e. family style, relationship with caregiver) was obtained from interviews with family members or from questionnaires administered by psychiatric social workers and completed by family members or public service staff. 'Family style' refers to the people with whom the patient was living; people were categorized as either the patient's partner or 'others', which included all other family members. Caregivers were identified as being the patient's partner, son or daughter, other family member or relative, or public service staff.

Reason for admission
Behavioural problems that caused distress to the caregiver and that were the primary reason for hospitalization were recorded following interviews with the caregiver. Reasons for hospitalization were classified as combative behaviour, overactivity, or apathy/ depression. Combative behaviour comprised physically or verbally aggressive behaviour, such as hitting, kicking, biting, throwing things, cursing, and screaming, during or between care provisions. Overactivity included non-aggressive behaviour that required monitoring, such as aimless wandering, trying to reach a different place, restlessness, or repetitive actions and/or mannerisms. Apathy/depression included serious apathetic or depressive behaviour, such as severe loss of appetite, refusal to eat, refusal to take medication, or suicidal tendencies. 14

Type of dementia
Diagnoses of Alzheimer's disease and vascular dementia were made by either of the two experienced geriatric psychiatrists (TK and MK) according to the Diagnostic and Statistical Manual of Mental Disorders, 4th edition. Diagnoses of dementia with Lewy bodies were made according to the consensus guidelines for the clinical and pathologic diagnosis of dementia with Lewy bodies. 15 All diagnoses were made after interviews with patients and family members or public service staff and based upon physical and neurological findings, laboratory data, and results of brain imaging studies.

Behavioural and psychiatric symptoms
The prevalence of behavioural and psychiatric symptoms was evaluated by the two geriatric psychiatrists (TK and MK) using the Behavioural Pathology in Alzheimer's Disease rating scale. 16 The presence or absence of symptoms in each of the following seven clusters comprising a subscale of Behavioural Pathology in Alzheimer's Disease was recorded: paranoid and delusional ideation; hallucinations; aggressiveness; activity disturbances; diurnal rhythm disturbances; affective disturbances; and anxieties or phobias.

Cognitive function
Cognitive function was evaluated by the geriatric psychiatrists in the acute psychogeriatric ward using the Mini-Mental State Examination (MMSE). 17

Functional state of daily living
Activities of daily living (ADL) were scored by welltrained nursing staff according to the Nishimura-style senile activities of daily living (N-ADL) scale, 18 which is one of the most used scales for the evaluation of ADL in Japan. On the N-ADL scale, five items are evaluated: walking/sitting; range of activities; dressing/ bathing, eating; and excretion. Each item is scored from 0 to 10 points, with a score of 0 indicating that that activity is not performed and a score of 10 indicating no impairment in the activity. The total (maximum score: 50) is taken as the N-ADL score. Nursing staff rated the reliability of the scale, when completed, as good. 18 Intervention Patients received pharmacological intervention as clinically indicated. All patients were treated under the supervision of the two geriatric psychiatrists based on their clinical judgement. Non-pharmacological interventions included occupational therapy and psychoeducation of patients and their families.
To aid patients settling back into their homes after discharge, we held care meetings with family members prior to discharge and identified potential problems in each patient's home care. We also visited patients' homes to confirm whether the help planned at the hospital was feasible in the specific environment and to provide advice. If necessary, we continued periodic home visits after discharge to follow up on patients' living and care status.

Statistical analysis
The primary end-point of the study was discharge home. For analysis, discharge destination was dichotomized into HD and all other destinations (not home discharge (NHD)), which included discharge to a care facility, transfer to a medical hospital, or death during hospitalization in the acute psychogeriatric ward.
Data management and statistical calculations were performed with the software package Stata version 11.0 (Statacorp, College Station, TX, USA).
Differences between groups (HD vs NHD) in terms of age, MMSE score, N-ADL score, and dose of antipsychotics used during hospitalization were evaluated by t-tests. Differences in frequency data were analyzed by χ 2 tests.
Univariate logistic regression analysis was conducted to determine the potential association of each demographic and clinical variable with HD. To determine independent predictors of HD and to create a predictive model, variables with P < 0.10 in the univariate analysis were included in a multiple logistic regression model. In this model, backward elimination was used to establish a cut-off P-value of 0.10.
For all statistical tests conducted, two-tailed P < 0.05 was considered significant. For multiple comparisons, P-values were adjusted with Bonferroni correction.

RESULTS
From April 2006 to March 2011, 600 patients were admitted to the acute psychogeriatric ward for the treatment of BPSD. Of these, 209 patients were excluded from the study because they had resided in a group home (49 patients), another care facility (92 patients), or a medical hospital (68 patients). Thus, 391 patients admitted from their own home were identified as eligible for inclusion in the study. All patients had been discharged (including those who died) when the study ended. Table 1 lists the demographic and clinical characteristics of the 391 patients at the time of admission. Of these, 163 patients (42%) were identified as having an HD after inpatient treatment. Of the 228 patients (58%) recorded as having an NHD, 52 went to a group home, 107 went to a care facility, 61 were transferred to a medical hospital, and 8 died during hospitalization. Patients in the HD group were less likely to have lived alone and were more likely to have lived with others prior to their admission compared with patients in the NHD group. Both MMSE and N-ADL scores at the time of admission were higher in the HD group, and these patients were more likely to manifest hallucinations and less likely to manifest aggressiveness in BPSD at the time of admission. The mean 1 SD length of hospital stay for the 391 patients was 132 + 211 days. The length of hospital stay was shorter in the HD than the NHD group (71 + 82 vs 176 + 259 days, respectively; P = 0.000). Table 2 details the psychotropic medications used during hospitalization by patients in the HD and NHD groups. There were no significant differences in the maximum daily dose of antipsychotics used during hospitalization or in the frequency of cholinesterase inhibitor use.
Univariate analysis revealed a significant positive correlation between HD and living with others, higher MMSE score, higher N-ADL score, and manifestation of hallucination. Negative correlations were found between HD and living alone and the manifestation of aggressiveness. Multiple logistic regression analysis identified higher MMSE and N-ADL scores as being significantly and independently associated with HD. In contrast, living alone and manifestation of aggressiveness in BPSD were independently and negatively associated with HD (Table 3).

DISCUSSION
From the early stages of hospitalization, we start to design a support plan for patients based on the assumption that they will live at home after leaving hospital. However, the results of the present study show that fewer than half the patients were discharged home. In addition to living alone, multivariate analyses identified lower MMSE scores, lower N-ADL scores, and the manifestation of aggressiveness in BPSD as being independently and negatively associated with HD. Findings of an association between functional state and HD were not surprising. Ono et al. demonstrated a positive correlation between N-ADL and HD in men and between revised Hasegawa's Dementia Scale (HDS-R) and HD in women. 13 However, because that study was conducted in a single institution, it was crucial to verify the generalizability of the previous findings in the present study. Although the study by Ono et al. focused on gender differences in predictors of HD, 13 our data did not show any significant effect of gender, even in univariate analysis.
It is of note that our data clearly showed a negative impact of aggressiveness on the likelihood of HD. Because we did not evaluate chronological changes in symptoms in detail, it is unclear whether the aggressiveness exhibited by our patients was less likely to respond to pharmacotherapy than other clusters of BPSD. However, some studies have indicated that antipsychotics may be more effective in treating aggression rather than hallucinations and delusions. [19][20][21][22][23] This means that the lower likelihood of HD for patients who exhibited aggressiveness at the time of admission could be due to the family's unwillingness to take the patient back into their home, even with symptom control. Because the aggressiveness exhibited by our patients could be harmful and is often directed at those close to them, the interpersonal relationships between the patients and their families may have already been disrupted prior to the patient's  hospital admission. In addition to strengthened service provisions, intervention strategies targeting family members may be required to decrease the stress associated with caregiving and to improve patient-family relationships. [24][25][26] There is growing empirical evidence that these measures can simultaneously improve the quality of life of both the patient and caregiver. 27,28 In contrast with aggressiveness, hallucinations usually manifest as disorganized speech or behaviour; they are associated with a less harmful nature towards others and are generally less disruptive. 29 A psychoeducational approach could help patients acquire an insight into their disease and enable their families to calmly deal with the problem behaviours, even if mild symptoms persist after pharmacological treatment.
The Orange Plan promotes earlier detection of and intervention for dementia. To this end, the Japanese government encourages primary care doctors to develop a better understanding of dementia and to collaborate with specialists from the earliest stages of the condition. 9 We hope that this strategy may, in turn, help prevent the development of prominent BPSD. Kunik et al. recently examined factors associated with the development of aggression and identified worsening of severe pain, caregiver burden, and declining mutuality over time as independent predictors. 30 Determining the factors precipitating BPSD could facilitate the development of preventive strategies in such cases.
Because the family is the first-line support for older people in Asian societies, stronger support systems are needed for patients living alone. 26 However, under the current Japanese Long-Term Care Insurance system, the absence of cohabitants is not taken into account when determining the level of care required. As nuclear families have become more common as the family unit, more elderly Japanese are living alone. This issue should be addressed in future changes to systems used to determine the level of care.
Several limitations of the present study must be acknowledged. The retrospective nature of the study, which was based on a review of patients' medical charts, means that the information available is limited. In addition, there could be other relevant variables that may have influenced HD. In particular, despite findings indicating its potential association with HD, we did not evaluate caregiver burden. 31 Furthermore, our routine practice does not include detailing of chronological responses to treatment. Finally, the present study was conducted in a single institution in a particular region of Japan, and the likelihood of HD largely depends on the number of social workers or resources available in the region. To confirm the generalizability of our results, future well-designed multicentre studies are warranted.
In conclusion, lower MMSE scores, lower N-ADL scores, the manifestation of aggressiveness in BPSD at the time of admission, and living alone prior to hospital admission could predict a lower likelihood of HD in patients with BPSD. These findings should be taken into consideration when managing patients with dementia to enable implementation of optimal intervention and care strategies to improve patients' quality of life.