[Discharge Dynamics and Related Factors of Long-stay Patients in Psychiatric Hospitals].

A longstanding challenge in Japan is prolonged psychiatric hospitalization and the associated difficulty of discharge, lost opportunities for patients' social participation, and stagnant reallocation of medical resources. Although the length of stay has been shortened recently on average, its distribution tends to be polarized into high-turnover and long-stay groups. To resolve these problems, we must understand the discharge dynamics of long-stay patients. Three questionnaires were sent to 733 randomly selected psychiatric hospitals (response rate: 24.3%; 178 hospitals, 2,480 patients). One questionnaire was on hospitalized patient numbers for one-year or longer stays as at the end of June 2007, recording each combination of Group (A or B), diagnosis, and hospitalization type. Group A referred to patients continuously hospitalized as at the end of June 2008; Group B referred to those discharged between July 2007 and June 2008. The second questionnaire was on hospital characteristics (founder, bed number, medical function, etc.), and the third questionnaire was on detailed patient characteristics (residential setting post-discharge, etc., for each Group B patient; a maximum of 20 patients per hospital consecutively in order of discharge). Valid data were obtained from 171 hospitals and 2,419 patients, with the latter increasing to 3,543 after weighting. The annual discharge rate (ADR; B/[A+B]) for the entire sample was 16.3%. Regarding the diagnosis, dementia showed the highest ADR (27.8%) and schizophrenia the lowest (13.5%). The ADRs for depression, bipolar disorder, and alcoholism were 23.9, 20.6, and 23.7% respectively. Regarding the hospitalization type, voluntary hospitalization (16.0%) and hospitalization for medical care and protection (16.8%) showed similar ADRs. Regarding the district, ADRs were high in Kinki (19.9%) and Kyushu (18.8%), and low in Kanto (14.1%) and Chugoku/Shikoku (14.2%). Multivariate analyses revealed that discharge within one year was significantly correlated with the diagnosis, district, hospital founder, and presence of psychiatric emergency or acute-phase treatment (acute-phase-type) wards in hospitals, but not with the hospitalization type, presence of psychiatric long-term care wards, or presence of senile dementia wards. The probability of discharge (odds ratio [95% confidence interval]) regarding the diagnosis was higher in dementia (2.47 [2.23-2.74]), alcoholism (2.09 [1.71-2.55]), depression (2.07 [1.65-2.59]), and bipolar disorder (1.70 [1.35-2.16]) than in schizophrenia (reference). Regarding the district, the probability was higher in Kinki (1.32 [1.12-1.54]) and Kyushu (1.27 [1.14-1.42]) than Kanto (reference). The probability was also lower in private hospitals (0.58 [0.51-0.66]) than in public/university hospitals (reference), and higher in hospitals with acute-phase-type wards (1.24 [1.14-1.35]) than in those without them (reference). The most common residential setting post-discharge for the total sample of weighted Group B patients was temporary hospitalization in another department prearranging psychiatric readmission (THAD, 35.8%), followed by death (18.2%), living with families/relatives (LF/R, 11.3%), a residential care facility for the aged (RCF-A, 9.5%), residential care facility for the disabled (RCF-D, 8.6%), hospitalization in another psychiatric hospital (7.4%), living alone (LA, 4.3%), permanent hospitalization in another department (PHAD, 4.3%), and others (0.7%). In dementia, death was common (31.0%) ; LF/R (1.8%) and LA (0.0%) were rare. As the age increased, the proportions of LF/R, LA, RCF-D, RCF-A, PHAD, and death changed; particularly, LA decreased and death increased markedly with age. Additionally, THAD amounted to approximately 40% in every age class of 40 years or older, contrasting with 11.4% in those under 40 years. The study's limitations include a low response rate, the elapsed time after the survey, and lack of attention paid to symptom severity. Nevertheless, it provides valuable insights into long-stay patients, including that discharge is least likely in schizophrenia and most likely via transfer or death for dementia. These results may encourage the efficient promotion of discharge and prevention of prolonged hospitalization according to patients' demographic, clinical, and social conditions.