assistance

Contents lists available at ScienceDirect

Psychiatry Research

journal homepage: www.elsevier.com/locate/psychres

Timely follow-up visits after psychiatric hospitalization and readmission in schizophrenia and bipolar disorder in Japan

Yasuyuki Okumuraa,b,⁎, Naoya Sugiyamac, Toshie Nodad

a Research Department, Institute for Health Economics and Policy, Association for Health Economics Research and Social Insurance and Welfare, Tokyo, Japan bDepartment of Psychiatry and Behavioral Science, Tokyo Metropolitan Institute of Medical Science, Tokyo, Japan cNumazu Chuo Hospital, Fukkokai Foundation, Shizuoka, Japan d Atami Chuo Clinic, Fukkokai Foundation, Shizuoka, Japan

A R T I C L E I N F O

Keywords: Bipolar disorder Schizophrenia Transition care Quality of health care

A B S T R A C T

The study objective was to investigate the association between timely follow-up visits after psychiatric hospi- talization and the risk of readmission in patients with schizophrenia or bipolar disorder. A retrospective cohort study was conducted using a nationwide claims database in Japan. Between April 2014 and March 2015, all psychiatric hospitalization data were obtained and patients with a principal diagnosis of schizophrenia or bi- polar disorder were followed up from 180 days before admission to 210 days after discharge. The primary outcome of this study was psychiatric readmission during the 180-day period (between 31 and 210 days) after the index discharge. A total of 48,579 eligible patients were identified. After psychiatric hospitalization, 15% of patients received no follow-up visits to a psychiatrist within 30 days. Patients who received follow-up visits had lower readmission rates during the subsequent 180 days (21.7% vs. 37.5%; adjusted risk ratio, 0.54 [95% confidence interval, 0.52–0.57]) than those who did not. Timely follow-up visits after discharge could be helpful for reducing the readmission risk in patients.

1. Introduction

Timely follow-up visits after psychiatric hospitalization are con- sidered an important component in the clinical process for promoting further recovery and preventing relapse (Hermann et al., 2004). How- ever, it remains unclear whether timely follow-up visits after psychia- tric hospitalization are associated with a reduced risk of readmission (Beadles et al., 2015; Kurdyak et al., 2018; Lin and Lee, 2008; Marcus et al., 2017).

A cohort study of 24,934 Medicaid patients, aged 22–64 years, re- ported no association between follow-up visits within 30-days after discharge and readmission within the subsequent 6-months in a de- pression cohort and a small association in a schizophrenia cohort (Beadles et al., 2015). A recent cohort study of 71,776 commercially and Medicaid insured patients, aged 18–64 years, showed that receipt of a follow-up visit within 30 days after discharge was associated with slightly lower odds of readmission within the subsequent 90 days in schizophrenia (odds ratio [OR], 0.88) and in bipolar (OR, 0.91) cohorts (Marcus et al., 2017). A recent cohort study of 19,132 patients with schizophrenia in Canada also found small associations (hazard ratio,

0.83–0.88) between follow-up visits within 30 days after discharge and readmissions within the subsequent 180 days (Kurdyak et al., 2018). However, a cohort study of 15,607 patients with schizophrenia in Taiwan found strong associations (OR, 0.33) between follow-up visits within 60 days after discharge and readmission within the subsequent 120 days after discharge (Lin and Lee, 2008).

Thus, the strength of the association between timely follow-up visits after discharge and subsequent readmission may vary by diagnosis and country. In the present study, we aimed to investigate the association between timely follow-up visits after psychiatric hospitalization and the risk of readmission in patients with schizophrenia and bipolar disorder in Japan.

2. Methods

2.1. Design

A retrospective cohort study was conducted using the National Database of Health Insurance Claims and Specific Health Checkups of Japan (NDB). The NDB includes almost all claims in Japan (Ministry of

https://doi.org/10.1016/j.psychres.2018.10.020 Received 7 March 2018; Received in revised form 8 October 2018; Accepted 8 October 2018

⁎ Corresponding author at: Department of Psychiatry and Behavioral Science, Tokyo Metropolitan Institute of Medical Science, 2-1-6 Kamikitazawa, Setagaya-ku, Tokyo 156-8506, Japan.

E-mail address: [email protected] (Y. Okumura).

Psychiatry Research 270 (2018) 490–495

Available online 09 October 2018 0165-1781/ © 2018 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).

T

Health, Labour and Welfare, 2016; Okumura et al., 2017), with the exception of claims solely covered by public funds. The NDB includes information on patient identification number, sex, and age group, along with medical practice codes, administration dates, and diagnostic codes.

The institutional review board at the Institute for Health Economics and Policy reviewed and approved the study protocol. Acquisition of informed consent was waived because of the anonymous nature of the data.

2.2. Setting

Japan has had a universal healthcare system since 1961. Japan had 330,694 psychiatric beds in 1599 hospitals in 2014 (Ministry of Health, Labour and Welfare and National Center of Neurology and Psychiatry, 2016). Hospitals with psychiatric beds are mainly private hospitals rather than public ones. There is no obligation for the hos- pitals to follow-up patients after discharge. In general, the universal health insurance system pays for 70% of the outpatient treatment costs, the System of Medical Payment for Services and Supports for Persons with Disabilities pays for approximate 20%, and patients are re- sponsible for the remaining amount (approximately 10%).

2.3. Patient selection

We identified all patients aged<65 years who were admitted to any psychiatric unit between April 2014 and March 2015. The psy- chiatric units included in the present study are presented in Table S1. To increase traceability, we used patient identification numbers, called “ID0” (Kubo et al., 2018). Initial admissions to psychiatric units during the period were identified as index admissions. Planned admissions for electroconvulsive therapy with a hospital stay of ≤ 3 days were ex- cluded. Patients with a principal diagnosis of schizophrenia and related psychoses (International Classification of Diseases, Tenth Edition [ICD- 10]: F20–F29) or bipolar affective disorder (F30–F31) were included using the algorithm defined by the Ministry of Health, Labour and Welfare (Ministry of Health, Labour and Welfare, 2015). Patients with a secondary diagnosis of dementia (F00–F03, F05.1, and G30–G31) or intellectual disability (F70–F79) in addition to the principal diagnosis of schizophrenia or bipolar disorder were excluded. Patients hospita- lized for longer than 180 days were excluded as including these patients would have meant that some patients would not have the required follow-up period of 210 days. Patients discharged to a non-psychiatric unit or deceased were excluded. Patients who enrolled in the database at least 180 days before the index admission and 210 days after the index discharge were included. Patients admitted to any type of hos- pital unit within 30 days after the index discharge were excluded be- cause of the time window for the exposure status. All patients were followed up from 180 days before the index admission to 210 days after the index discharge.

2.4. Exposure

The exposure of interest was a timely follow-up visit to a psychia- trist. A timely follow-up visit was defined as an outpatient visit to a psychiatrist within 30 days after the index discharge (the medical practice codes for psychiatric visits are listed in Table S2). Our defini- tion of follow-up visit included passive outpatient visits as well as home-visit services by psychiatrists. In addition, the definition included psychiatric consultation for at least 5 min delivered in an individual- based format rather than a group-based format.

2.5. Outcomes

The primary outcome of this study was psychiatric readmission during the 180-day period (between 31 and 210 days) after the index

discharge. The secondary outcome was psychiatric readmission during the 90-day period (between 31 and 120 days) after the index discharge. Planned readmissions for electroconvulsive therapy were excluded from the definition of psychiatric readmission.

2.6. Other variables

As potential covariates, we extracted information on patient de- mographic characteristics (sex and age), characteristics during the 180 days prior to the index admission (Charlson index (Sundararajan et al., 2007), diagnosis of substance use disorders [ICD-10 codes: F10–F19], number of psychiatric visits, history of psychiatric admission, and

Table 1 Sample characteristics of the entire cohort.

No follow-up visit (N=7246)

Follow-up visit (N=41,333)

Characteristics n % n % Standardized difference, %

Age, years 0–19 202 2.8 1145 2.8 0.0 20–34 1609 22.2 9877 23.9 −4.0 35–49 2727 37.6 17,293 41.8 −8.6 50–64 2708 37.4 13,018 31.5 12.4

Sex, female 3625 50.0 24,088 58.3 −16.7 Number of psychiatric visits within 180 days before index admission

0 2377 32.8 4810 11.6 52.8 1–3 1693 23.4 5686 13.8 24.9 4–6 1311 18.1 8907 21.5 −8.5 7–12 1377 19.0 15,074 36.5 −39.9 13 or greater 488 6.7 6856 16.6 −31.2

History of psychiatric admission within 180 days before index admission

1160 16.0 4957 12.0 11.5

History of intensive care unit admission within 180 days before index admission

182 2.5 979 2.4 0.6

Charlson index within 180 days before admission 0 4933 68.1 28,663 69.3 −2.6 1 1501 20.7 8448 20.4 0.7

2 468 6.5 2605 6.3 0.8 3 or greater 344 4.7 1617 3.9 3.9

Diagnosis of substance use disorders within 180 days before admission

336 4.6 1838 4.4 1.0

Type of hospital at admission General hospital 823 11.4 5345 12.9 −4.6 Non-general hospital

6423 88.6 35,988 87.1 4.6

Type of unit at admission Acute care unit 3941 54.4 25,580 61.9 −15.2 Non-acute care unit 3305 45.6 15,753 38.1 15.2

Type of admission Involuntary 3136 43.3 18,579 45.0 −3.4 Voluntary 4110 56.7 22,754 55.0 3.4

Principal diagnosis (ICD- 10 codes) Schizophrenia (F2) 6050 83.5 32,097 77.7 14.7 Bipolar affective disorder (F30–F31)

1196 16.5 9236 22.3 −14.7

Use of ECT during index admission

172 2.4 1030 2.5 −0.6

Length of hospital stay 1st (1–21 days) 1614 22.3 7993 19.3 7.4 2nd (22–40 days) 1300 17.9 8418 20.4 −6.4 3rd (41–64 days) 1227 16.9 8581 20.8 −10.0 4th (65–89 days) 1287 17.8 8002 19.4 −4.1 5th (90–180 days) 1818 25.1 8339 20.2 11.7

Abbreviations: ECT, electroconvulsive therapy; ICD, international classification of diseases.

Y. Okumura et al. Psychiatry Research 270 (2018) 490–495

491

history of intensive care unit [ICU] admission), and characteristics of index admissions (type of hospital, unit, and admission; principal di- agnosis, use of electroconvulsive therapy, and length of stay). These covariates were selected based on evidence derived from previous studies (Beadles et al., 2015; Kurdyak et al., 2018; Lin and Lee, 2008; Marcus et al., 2017).

The type of hospital was classified as either general or non-general; general hospitals have ≥ 100 beds with at least the following five specialties: internal medicine, surgery, obstetrics and gynecology, oto- laryngology, and ophthalmology. The type of unit was classified as ei- ther acute care or non-acute care; acute care units were defined as the eight types of psychiatric units listed in Table S1. The type of admission was classified as either voluntary or involuntary.

2.7. Statistical analyses

First, we assessed covariate balance using standardized differences, in which an absolute value greater than 10% indicates an important imbalance in the prevalence of a covariate between the groups (Austin, 2011). Second, we fitted a Poisson regression model and compared risks between the groups. All potential covariates were si- multaneously entered into the models. Risk ratios (RR) and their 95% confidence intervals (CI) were derived from the model. Third, we conducted a subgroup analysis to examine whether the association varied across the levels of all covariates. We assessed the statistical significance of interaction terms with a significance level of 0.05. Fourth, we conducted a sensitivity analysis using a propensity matching method. We estimated propensity scores by using a logistic regression

Fig. 1. Follow-up visits and psychiatric readmission during the 180-day period (between 31 and 210 days) after the index discharge. *p< .05; CI, confidence interval; ECT, electroconvulsive therapy; ICU, intensive care unit; RR, risk ratio; LOS, length of stay.

Y. Okumura et al. Psychiatry Research 270 (2018) 490–495

492

model with the following algorithm: nearest neighbor matching method, matching ratio of 1:1, caliper width of 0.2 of the standard deviation of the logit of the propensity score, and no replacement (Austin, 2011). Within the propensity score matched cohort, we used the method of Agresti and Min to compare readmission rates (Agresti and Min, 2004). Fifth, we conducted sensitivity analyses based on another time frame in which a timely follow-up visit was defined as an outpatient visit to a psychiatrist within 60 days after the index dis- charge, and outcomes were defined as psychiatric readmissions during the 180-day period (between 61 and 240 days) and the 90-day period (between 61 and 150 days) after the index discharge. Statistical ana- lyses were performed using R version 3.4.1 (R Foundation for Statistical Computing, Vienna, Austria) with the MatchIt package.

3. Results

3.1. Study population

The study cohort included 48,579 patients (Fig. S1). Among these, 7246 patients (14.9%) had no follow-up visit within 30 days after discharge. Of the patients receiving timely follow-up visits, only 325 (0.7%) received home-visit services by psychiatrists. A between-groups comparison of the sample characteristics showed major imbalances in the number of prior psychiatric visits within 180 days before admission (Table 1).

Fig. 2. Follow-up visits and psychiatric readmission during the 90-day period (between 31 and 120 days) after the index discharge *p< .05; CI, confidence interval; ECT, electroconvulsive therapy; ICU, intensive care unit; RR, risk ratio; LOS, length of stay.

Y. Okumura et al. Psychiatry Research 270 (2018) 490–495

493

3.2. Main analyses

Patients who received timely follow-up visits had lower readmission rates during the 180-day (21.7% vs. 37.5%; adjusted RR, 0.54 [95% CI, 0.52–0.57]) and 90-day (12.3% vs. 29.0%; adjusted RR, 0.40 [95% CI, 0.38–0.42]) periods than patients who did not receive follow-up visits (Figs. 1 and 2). Subgroup analysis showed that the strength of the as- sociations was modified by several factors, although the direction of the associations did not vary by subgroup. For example, the strength of the association was weaker among patients with a history of psychiatric admissions (adjusted RR, 0.70) than among those without (adjusted RR, 0.50) (Fig. 1). The strength of the association was similar among pa- tients with bipolar disorder and those with schizophrenia (Fig. 1). Si- milar findings were observed in the secondary outcome results (Fig. 2).

3.3. Sensitivity analyses

We identified a propensity score-matched cohort of 14,454 patients from the entire cohort (Table 2; Table S3). The covariate balance was considerably improved (Table 2). Follow-up visits were associated with reduced readmission rates during the 180-day (21.2% vs. 37.4%; RR, 0.57 [95% CI, 0.54–0.60]) and 90-day (12.7% vs. 28.9%; RR, 0.44 [95% CI, 0.41–0.47]) periods (Table S4). The strength of the association between follow-up visits and readmission was weaker when follow-up visits were defined as ≤ 60 days rather than ≤ 30 days (Tables S4 and S5).

4. Discussion

Our study found a moderate association between timely follow-up visits after psychiatric hospitalization and risk of readmission in pa- tients with schizophrenia or bipolar disorder using the nationwide claims database. The strength of this association appears to be much higher in Japan than in the United States and Canada (Kurdyak et al., 2018; Marcus et al., 2017), although the risk of readmission in patients receiving follow-up visits was similar in Japan (180-day readmission rate: 22%; 90-day readmission rate: 12%) and in the United States (90- day readmission rate: 10%–13%) and Canada (180-day readmission rate: 22%). Follow-up rates after psychiatric hospitalization and base- line risk of readmission (i.e., risk of readmission in patients receiving no follow-up visits) might mutually influence the strength of the associa- tion between timely follow-up visits and readmission. Patients with no follow-up visits who are living in a country with high follow-up rates are likely to have a higher risk of readmission than those living in a country with low follow-up rates.

We found that the follow-up rate within 30 days after psychiatric hospitalization was 85%, which was much higher than that reported in previous studies. For example, Marcus et al. reported that the 30-day follow-up rate was 64%–65% in a schizophrenia cohort and 62%–73% in a bipolar cohort of commercially and Medicaid insured adults (Marcus et al., 2017). Fontanella et al. also found that the 30-day follow-up rate was 70% in a mood disorder cohort of Medicaid enrolled youth (Fontanella et al., 2016). Kurdyak et al. reported that the 30-day follow-up rate for psychiatrists or primary care physicians was 65% in a schizophrenia cohort of Canadian adults (Kurdyak et al., 2018).

The baseline risk of readmission within the subsequent 180 days was higher in our study (38%) than that reported by Kurdyak et al. (26%) (Kurdyak et al., 2018). Similarly, the baseline risk of readmission within the subsequent 90 days was also much higher in our study (29%) than that reported by Marcus et al. (13% in schizophrenia and 11% in bipolar cohorts) (Marcus et al., 2017).

Our study has several limitations. First, we could not measure im- portant potential covariates, such as history of suicide attempts, fi- nancial status, marital status, education level, patient satisfaction with treatment, insight into illness, and type of discharge (e.g., discharge on/ against medical advice) (Donisi et al., 2016). For example, patients who have better insight into their illness are more likely to receive follow-up visits and to have better adherence to medications. As a result, those patients might have lower risk of readmission. Future studies are needed to confirm the balance pertaining to these unmeasured covari- ates between the groups. Second, our data did not include claims solely covered by public funds, comprising approximately 19% of psychiatric discharges (Niimura et al., 2017). Third, our data could not exclude healthy-user bias because we focused only on patients who were dis- charged to the community within 180 days after admission and those who enrolled in the database at least 210 days after discharge. Fourth, we could not conduct subgroup analyses by diagnostic subtype (i.e., manic, depressed, or mixed episode) due to concerns about coding ac- curacy. Fifth, we focused on limited aspects of the exposures and out- comes. It would be valuable to further investigate the effectiveness of timely follow-up visits on the risk for readmission in the short term

Table 2 Sample characteristics of the propensity score-matched cohort.

No follow-up visit (N=7227)

Follow-up visit (N=7227)

Characteristics n % n % Standardized difference, %

Age, years 0–19 202 2.8 237 3.3 −2.9 20–34 1609 22.3 1661 23.0 −1.7 35–49 2727 37.7 2662 36.8 1.9 50–64 2689 37.2 2667 36.9 0.6

Sex, female 3625 50.2 3659 50.6 −0.8 Number of psychiatric visits within 180 days before index admission

0 2358 32.6 2404 33.3 −1.5 1–3 1693 23.4 1593 22.0 3.3 4–6 1311 18.1 1209 16.7 3.7 7–12 1377 19.1 1454 20.1 −2.5 13 or greater 488 6.8 567 7.8 −3.8

History of psychiatric admission within 180 days before index admission

1141 15.8 1121 15.5 0.8

History of intensive care unit admission within 180 days before index admission

177 2.4 190 2.6 −1.3

Charlson index within 180 days before admission 0 4932 68.2 4931 68.2 0.0 1 1493 20.7 1492 20.6 0.2 2 461 6.4 464 6.4 0.0 3 or greater 341 4.7 340 4.7 0.0

Diagnosis of substance use disorders within 180 days before admission

335 4.6 347 4.8 −0.9

Type of hospital at admission General hospital 822 11.4 813 11.2 0.6 Non-general hospital 6405 88.6 6414 88.8 −0.6

Type of unit at admission Acute care unit 3941 54.5 3918 54.2 0.6

Non-acute care unit 3286 45.5 3309 45.8 −0.6 Type of admission

Involuntary 3136 43.4 3187 44.1 −1.4 Voluntary 4091 56.6 4040 55.9 1.4

Principal diagnosis (ICD-10 codes) Schizophrenia (F2) 6031 83.5 5994 82.9 1.6 Bipolar affective disorder (F30–F31)

1196 16.5 1233 17.1 −1.6

Use of ECT during index admission

169 2.3 188 2.6 −1.9

Length of hospital stay 1st (1–21 days) 1613 22.3 1611 22.3 0.0 2nd (22–40 days) 1300 18.0 1292 17.9 0.3 3rd (41–64 days) 1227 17.0 1206 16.7 0.8 4th (65–89 days) 1287 17.8 1271 17.6 0.5 5th (90–180 days) 1800 24.9 1847 25.6 −1.6

Abbreviations: ECT, electroconvulsive therapy; ICD, international classification of diseases.

Y. Okumura et al. Psychiatry Research 270 (2018) 490–495

494

(e.g., within 30 days after discharge) as well as in the long term (e.g., 360 days after discharge). In addition, studies are needed to determine the comparative effectiveness of type of follow-up visits with con- sideration of the costs and benefits. Sixth, the generalizability of our findings is uncertain because healthcare system structure differs among countries and it may influence the follow-up and readmission rates.

Nevertheless, our findings suggest that 15% of discharged patients with schizophrenia or bipolar disorder do not receive timely follow-up visits to a psychiatrist. These patients are at higher risk of psychiatric readmission. Therefore, timely follow-up visits after discharge could be helpful for reducing the readmission risk in patients.

Conflict of interest

For the past 3 years, YO has received personal fees from Merck & Co., Inc.; Janssen Pharmaceuticals, Inc.; the Medical Technology Association; Cando Inc.; and the Japan Medical Data Center. He has also received research grants from the Japan Agency for Medical Research and Development; the Ministry of Health, Labour and Welfare; the Japan Society for the Promotion of Science; the Institute for Health Economics and Policy; and Mental Health and Morita Therapy. Outside the submitted work, NS has received grants from the Ministry of Health, Labour and Welfare, along with personal fees and non-financial support from Otsuka Pharmaceutical, Janssen Pharmaceutical K.K., Eli Lilly Japan K.K., Pfizer Inc., Meiji Seika Pharma Co., Ltd., MSD K.K, and Daiichi-Sankyo Company, Ltd. TN has no conflicts of interest.

Acknowledgments

We would like to thank Editage (www.editage.jp) for English lan- guage editing. This work was funded by a grant from the Japan Society for the Promotion of Science (number: 18K09991).

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.psychres.2018.10.020.

References

Agresti, A., Min, Y., 2004. Effects and non-effects of paired identical observations in

comparing proportions with binary matched-pairs data. Stat. Med. 23 (1), 65–75. Austin, P.C., 2011. An introduction to propensity score methods for reducing the effects of

confounding in observational studies. Multivar. Behav. Res. 46 (3), 399–424. Beadles, C.A., Ellis, A.R., Lichstein, J.C., Farley, J.F., Jackson, C.T., Morrissey, J.P.,

Domino, M.E., 2015. First outpatient follow-up after psychiatric hospitalization: does one size fit all? Psychiatr. Serv. 66 (4), 364–372.

Donisi, V., Tedeschi, F., Wahlbeck, K., Haaramo, P., Amaddeo, F., 2016. Pre-discharge factors predicting readmissions of psychiatric patients: a systematic review of the literature. BMC Psychiatry 16 (1), 449.

Fontanella, C.A., Hiance-Steelesmith, D.L., Bridge, J.A., Lester, N., Sweeney, H.A., Hurst, M., Campo, J.V., 2016. Factors associated with timely follow-up care after psychiatric hospitalization for youths with mood disorders. Psychiatr. Serv. 67 (3), 324–331.

Hermann, R., Mattke, S., Members of the OECD Mental Health Care Panel, 2004. Selecting indicators for the quality of mental health care at the health systems level in OECD countries. https://www.oecd.org/els/health-systems/33865630.pdf accessed Febrary 5 2018.

Kubo, S., Noda, T., Myojin, T., Nishioka, Y., Higashino, T., Matsui, H., Kato, G., Imamura, T., 2018. National Database of Health Insurance Claims and Specific Health Checkups of Japan (NDB): Outline and patient-matching technique. https://www.biorxiv.org/ content/biorxiv/early/2018/03/29/280008.full.pdf accessed 10 April 2018.

Kurdyak, P., Vigod, S.N., Newman, A., Giannakeas, V., Mulsant, B.H., Stukel, T., 2018. Impact of physician follow-up care on psychiatric readmission rates in a population- based sample of patients with schizophrenia. Psychiatr. Serv. 69 (1), 61–68.

Lin, H.C., Lee, H.C., 2008. The association between timely outpatient visits and the likelihood of rehospitalization for schizophrenia patients. Am. J. Orthopsychiatry 78 (4), 494–497.

Marcus, S.C., Chuang, C.C., Ng-Mak, D.S., Olfson, M., 2017. Outpatient follow-up care and risk of hospital readmission in schizophrenia and bipolar disorder. Psychiatr. Serv. 68 (12), 1239–1246.

Ministry of Health, Labour and Welfare, 2015. Methods for identifying one principle diagnose. http://www.mhlw.go.jp/file/06-Seisakujouhou-12400000-Hokenkyoku/ 0000090763.xls accessed 4 December 2017.

Ministry of Health, Labour and Welfare, 2016. Population Census in 2015. http://www.e- stat.go.jp/SG1/estat/GL08020103.do?_csvDownload_&fileId=000007809775& releaseCount=3 accessed 4 December 2017.

Ministry of Health, Labour and Welfare, National Center of Neurology and Psychiatry, 2016. 630 survey in 2014. https://www.ncnp.go.jp/nimh/seisaku/data/630/assets/ excel/h26_630.zip accessed 4 July 2018.

Niimura, J., Nakanishi, M., Yamasaki, S., Nishida, A., 2017. Regional supply of outreach service and length of stay in psychiatric hospital among patients with schizophrenia: national case mix data analysis in Japan. Psychiatry Res. 258, 295–298.

Okumura, Y., Sakata, N., Takahashi, K., Nishi, D., Tachimori, H., 2017. Epidemiology of overdose episodes from the period prior to hospitalization for drug poisoning until discharge in Japan: an exploratory descriptive study using a nationwide claims da- tabase. J. Epidemiol. 27 (8), 373–380.

Sundararajan, V., Quan, H., Halfon, P., Fushimi, K., Luthi, J.C., Burnand, B., Ghali, W.A., International Methodology Consortium for Coded Health, I., 2007. Cross-national comparative performance of three versions of the ICD-10 Charlson index. Med. Care 45 (12), 1210–1215.

Y. Okumura et al. Psychiatry Research 270 (2018) 490–495

495