Risk factors for readmission in schizophrenia treated with combined psychoeducation and standard therapy

Abstract Background We conducted a historical cohort study of patients with schizophrenia to identify more robust risk factors at discharge that contribute to readmission within a year. Methods and Findings The subjects underwent brief psychoeducation during hospitalization. Multivariate analysis was conducted using factors selected in the univariate analysis. Using logistic regression analysis, the number of hospital admissions (P = .01) and Schedule for Assessment of Insight Japanese version score (P = .04) were identified as risk factors for readmission, with odds ratios of 0.70 and 1.18, respectively. Conclusions These results suggest that improvement in insight and early intervention may lead to a more stable community life.

maintain a certain level of social functioning, as well as remission of psychiatric symptoms in order to maintain community life, the practice guideline published by the American Psychiatric Association (APA) recommends the implementation of psychoeducational programs as part of psychosocial interventions. 1 A systematic review reported that in addition to the usual care of patients with schizophrenia, psychoeducation reduces readmission rates. 8 We have developed and implemented a general brief psychoeducation program for inpatients with schizophrenia, which is conducted in voluntary groups in accordance with these reports. 9,10 However, even with the addition of a psychoeducational program to the standard treatment, the readmission rate within one year was 22%. Some factors, such as living environment, Global Assessment of Functioning (GAF), previous number of hospital admissions, social resources, and support, are risk factors for readmission in patients with schizophrenia who have been treated with psychotherapy and antipsychotics. 11 Because standard treatment with psychoeducational programs is expected to have different effects compared with psychotherapy and antipsychotic therapy, different risk factors may be involved regarding readmission.
We previously explored risk factors for readmission in patients with schizophrenia who underwent a psychoeducation program and reported multiple risk factors. 10 However, the study did not research factors such as comorbidities, long-acting injection (LAI), and social resources and support on discharge that may have a significant impact on readmission. Additionally, that study only investigated the individual effects of several possible risk factors and did not adjust for confounding factors. To make psychoeducation programs more effective, we need to more broadly and accurately identify risk factors for readmission in patients who have received standard treatment, including psychoeducation. If risk factors for readmission could be identified at the time of discharge from the end of the psychoeducation program, healthcare workers would be able to provide patients with additional care during hospitalization and focused care after discharge, resulting in reduction in readmission. Therefore, we conducted a historical cohort study to identify robust risk factors for readmission within one year of discharge at the time of hospital discharge.

| Subjects
The subjects of this study were patients who were diagnosed as schizophrenic by psychiatrists using the International Classification of Diseases 10 between 2010 and 2016 at the Showa University Karasuyama Hospital. During hospitalization, the subjects underwent psychoeducation programs in addition to standard therapy and were discharged within one year after completing the programs. Exclusion criteria were as follows: (1) not participating in any of the psychoeducation programs, either disease (diagnosis, treatment, and prognosis) or pharmacotherapy, (2) transferred to another medical institution after discharge, (3) discharge without a valid reason, (4) discharge with next scheduled hospitalization, and (5) more than three courses of psychoeducation programs before discharge. The program used in this study was prepared in accordance with the APA guideline. 1 Inclusion criteria for the psychoeducation program were as follows: patients who were admitted in the sub-acute phase and aimed to be discharged to the social community. Programs were held in the open space of the ward where other patients were present as spectators.
The psychoeducation program consists of eight sessions, including a lecture and group discussion of 30 mins each. 9 The psychoeducation program provides participants key information on symptoms, psychosocial interventions, medications, side effects, and relapse prevention, along with social skill training and visit to psychiatric day care service center. The psychoeducation program assessed in this study was based on a standard textbook of psychoeducation programs that have been found useful in previous reports. The quality of the psy-

| Measures
The outcome of the study was readmission for worsening of symptoms or suicide within a year after discharge. The risk factors listed in Table 1 were investigated before and after the program, and at discharge. The values of each factor were extracted from medical records. In cases where data were lacking, we directly interviewed a primary doctor. A conversion table prepared by Inada and Inagaki was used to convert doses of antipsychotics to Chlorpromazine (CPZ) equivalents. 12 Several rating scales are used to objectively evaluate the patients in this study. Psychiatric symptoms were assessed using both the Positive and Negative Syndrome Scale (PANSS) and the Brief Psychiatric Rating Scale (BPRS), both of which are common rating scales for schizophrenia. 13,14 The Schizophrenia Cognition Rating Scale Japanese version (SCoRS-J) was used to assess cognitive function. It was developed to evaluate the degree of cognitive impairment directly related to daily functioning of patients with schizophrenia. 15 The Schedule for Assessment of Insight Japanese version (SAI-J) score, an objective rating scale, was used to assess insight. 16 The SAI-J is a semi-structured interview used to measure three insight di-

| Statistical analysis
Subjects who met the outcome were defined as the readmission group and those who did not were defined as the nonreadmission group. A univariate analysis was conducted to exclude factors that were not clearly relevant to maintain the detection sensitivity of a multivariate analysis. In the univariate analysis, P < .10 was set to ensure that related factors were not missed. The difference in each factor between the two groups was evaluated using univariate analysis (P < .10), in which t test was used for continuous variables and chi-square or Fisher's test was used for nominal variables.
Multivariate analysis was subsequently conducted using the factors selected in the univariate analysis. In the multivariate analysis, we performed logistic regression analysis (stepwise method) with readmission as the dependent variable and the factors selected in the univariate analysis as covariates. We calculated the factors and odds ratios that independently contributed to readmission (P < .05). SPSS Statistics 25 (IBM Japan) was used to perform statistical analysis.

| RE SULTS
A total of 117 subjects were included in this study. Table 1 shows the characteristics of the subjects. The mean age of the subjects was 42.4 ± 11.5 years. Twenty-eight patients were admitted for the first time. Twenty-eight subjects (23.9%) received LAI, and 98 (83.8%) were discharged from the hospital using social resources and support. Thirty-one patients (26.5%) were readmitted within one year.
The 31 patients in the readmission group and 86 patients in the nonreadmission group were analyzed. Table 2 shows the results of the univariate analysis. Significant differences were found in the number of hospital admissions (P < .01), BPRS score (P = .09), SAI-J score (P = .01), SCoRS-J score (P = .01), and LAI (P = .09). However, there was no significant difference regarding the use of social resources and support.
Since there was no multicollinearity among the factors extracted in the univariate analysis, multivariate analysis was performed including all factors. Table 3 shows the results of multivariate analysis.
Logistic regression analysis identified the number of hospital admissions (P = .01) and SAI-J score (P = .04) as risk factors for readmission, with odds ratios (95% Confidence Interval) of 0.70 (0.53-0.93) and 1.18 (1.01-1.40), respectively. This means that the risk of readmission increases 1.43 times for each additional previous hospitalization and 1.18 times for each decrease in the SAI-J score.

| DISCUSS ION
We extensively studied risk factors for readmission in patients with schizophrenia who underwent a psychoeducation program. We found that the number of hospital admissions and SAI-J score were independently associated with readmission. The number of hospital admissions has been reported as a risk factor in several studies with different patient settings, 11 and our results support these reports.
The number of previous admissions was shown to be a robust factor associated with readmission, regardless of the psychoeducational program or current treatment.  The SAI-J score is a novel factor that was identified in this study as an independent risk factor associated with readmission. Poor insight has been found to be associated with higher risks of relapse and readmission. 19 Insight has also been reported to be associated with adherence and community functioning. 20 Lysaker et al concluded that lack of insight can lead to worsening symptoms secondary to poor treatment adherence and therapeutic alliance. This results in impaired community functioning. 21 In this study, both the DAI-10, a measure of medication adherence, and the GAF, a measure of social functioning, tended to score lower in the readmission group, but the differences were not significant. The SAI-J score includes a comprehensive and objective assessment of need and insight regarding treatment and medication, and is thought to have been extracted to sensitively indicate readmission. We previously reported that SCoRS-J after a psychoeducation program may be associated with readmission. 10 In this study, SCoRS-J was significantly different in the univariate analysis, but was not identified as a risk factor in the multivariate analysis. As a report suggested that clinical insight is associated with cognitive functions, 22 we examined the multicollinearity of both factors, but found no correlation. The results suggest that the SAI-J may be more sensitive regarding readmission than the SCoRS-J.
On the one hand, LAI of antipsychotics use was thought to be a factor that reduced readmission, but was not shown to be associated with it. LAI of antipsychotics has continued effect for 2-4 weeks with a single intramuscular injection. LAIs have been reported to be partially superior to oral medications in preventing relapse and reducing number of hospitalizations. 23 The APA guidelines suggest the use of LAIs for patients with a history of poor or uncertain adherence. 24 We previously reported a significant increase in DAI-10 scores after attending a psychoeducational program. 10 In the present study, we hypothesized that the psychoeducation program-similar to the LAI-may have improved medication adherence.
Based on the PANSS and the CPZ-equivalent dose of antipsychotic drugs, the subjects of this study are likely to be similar to the general Japanese schizophrenic population. 25,26 The one- year readmission rate in our study was 26.5%, which seems to be comparable to the approximately 30% readmission rate in Japan.
The subjects were mostly middle-aged, had less physical comorbidities, and used substances at a lower rate than the Western population. This difference may be attributed to the different approach in Japan regarding health promotion and drug regulation compared with Western countries. On the other hand, the utilization rate of social resources and support was high (83.8%).
Donisi et al reported that receiving benefits (pension, a serviceconnected disability, or other welfare benefits) influence readmission. 11 Therefore, we assessed whether lack of social resources and support can be considered as a risk factor; and our observations ruled out this possibility.
One of the limitations of this study, which is a problem associated with historical cohort studies, is that we were unable to adequately track the outcomes of patients who were transferred to other medical institutions after discharge. We attempted to follow these subjects as closely as possible, but were unable to do so com- In this study, the PANSS or BPRS, which shows the psychiatric symptoms of the subjects, adopted the total score as a factor.
However, the individual effects of the PANSS or BPRS constructs were not examined to avoid the risk of multiple analyses and multicollinearity in the statistical analysis. Further research may be needed to examine how each component affects the outcome. In this study, understanding of the psychoeducation program by the subjects was not directly assessed. In the psychoeducation program, subjects were given a questionnaire after each section, and individual support was provided to subjects who had an insufficient understanding. Following-up with the subjects helped us achieve a level of understanding of the psychoeducation program to a certain degree. Furthermore, it was our understanding that the same quality of effect was achieved. Additional studies may be needed to examine the effect of the subjects' understanding of the program on their readmission.
Since insight subsequent to the implementation of a psychoeducational program is important for its effectiveness, it may be useful to modify the content of the program to focus on insight.
Furthermore, the program could be made even more useful by incorporating individualized responses, such as implementing additional programs for patients with low SAI-J scores after psychoeducation.
In this study, we found that the number of hospital admissions and SAI-J score may be robust risk factors for readmission in patients with schizophrenia who underwent a psychoeducation program. Based on these results, improvements to the psychoeducation program and its early introduction in treatment may lead to a more stable community life for patients with schizophrenia.

ACK N OWLED G M ENTS
We would like to thank N. Watanabe and Y. Kataoka for their technical assistance. We also wish to thank the medical staff of the Karasuyama Hospital. This work was supported by the Showa University Pharmacy School. The authors would like to thank Enago (www.enago.jp) for the English language review.

CO N FLI C T O F I NTE R E S T
The authors of the present study declare that they have no conflicts of interest.

AUTH O R CO NTR I B UTI O N S
SS, TK, TT, TI, AI, and AI conceptualized the study. SS, TK, YN, HK, and TN designed the study. SS, YN, TT, HK, TN, and NF acquired the data. SS and YN analyzed the data. SS, TK, YN, TT, AI, and AI interpreted the data.

A PPROVA L O F TH E R E S E A RCH PROTO CO L BY A N I N S TITUTI O N A L R E V I E WER B OA R D
This study was approved by the Institutional Review Board of Showa University Karasuyama Hospital.

PATI E NT CO N S E NT S TATE M E NT
This study is an observational study and is posted on the institution website as a clinical study using opt-out. https://www.showa -u.ac.jp/album s/abm.php?d=505&f=abm00 010513.pdf&n=a1498 79793 1725.pdf.

R EG I S TRY A N D TH E R EG I S TR ATI O N N O. O F TH E S TU DY/ TR I A L
B-2016-042.

A N I M A L S TU D I E S
Not available.

PE R M I SS I O N TO R E PRO D U CE M ATE R I A L FRO M OTH E R S O U RCE S
We do not duplicate material from other sources.

DATA AVA I L A B I L I T Y S TAT E M E N T
The raw data belonged to the present study cannot be made publicly available, because the disclosure of personal data was not included in the research protocol of the present study.