Factors associated with death, hospitalization, resignation, and sick leave from work among patients with schizophrenia in Japan: a nested case–control study using a large claims database

The lifetime prevalence of schizophrenia is estimated to be 0.48% worldwide [1] and 0.59% in Japan [2]. Patients with schizophrenia have diverse symptoms including hallucination and delusion (positive symptoms), inactivity and social withdrawal (negative symptoms), cognitive impairment, and depressive symptoms. These symptoms have a negative impact on patients’ lifestyle, increase their physical disease risk, limit their social functioning, and result in loss of work productivity.

Schizophrenia is associated with approximately 14.5 years of potential life loss compared to the general population [3]. Relapses in schizophrenia occur frequently, and about one-fifth of patients with schizophrenia are hospitalized in Japan [4]. Thus, the employment rate of patients with schizophrenia is very low compared to the general population [5]. A recent Japanese study reported that 48.9% of patients with schizophrenia do not have a job [2]. Furthermore, patients with schizophrenia experience difficulties sustaining a job, and 51% resign within one year [6, 7]. Thus, treatments of schizophrenia to reduce premature mortality, prevent relapse, and maintain social functioning and an independent social life are required [8]. Death, hospitalization, and continuous employment are important clinical outcomes for evaluating the treatment of schizophrenia, and clarifying the risk factors for these outcomes is key to improving the treatment approach.

Treatments of schizophrenia are mainly based on pharmacotherapy with antipsychotics to improve psychotic symptoms. Continuous antipsychotic treatment is known to be an important factor for lowering risk of death and hospitalization for patients with schizophrenia [8, 9]. On the other hand, other symptoms and impairments, such as cognitive impairment, depressive symptoms, and various physical symptoms, including drug-induced ones, may increase mortality risk and cause declines in social functioning [10‐13].

Depression in schizophrenia is a critical symptom, and can occur at any phase throughout the course of schizophrenia illness – at the prodromal, first-episode, acute, and chronic stable phases [14]; its prevalence among patients with schizophrenia was estimated as 32.6% by a meta-analysis [15]. Depression in schizophrenia is reported to be a risk factor for suicide [11, 16], which is associated with more life-years loss than other causes of death in patients with schizophrenia [10]. It has been reported that depressed patients with schizophrenia are more likely to use relapse-related mental health services compared to non-depressed patients with schizophrenia, and improvement of depressive symptoms also tended to improve these problems [17]. Recently, a cross-sectional study in Japan reported that patients with schizophrenia with moderate or severe depressive symptoms had higher levels of work productivity loss than those with mild or no depressive symptoms [2]. Therefore, depression in schizophrenia is presumed to be a risk factor for death, hospitalization, and lack of continuous employment.

The risk factors for these important clinical outcomes remain unclear, and these outcomes depend on the medical and social environments in individual countries. Therefore, it is important to accumulate clinical evidence by utilizing real-world data in the targeted country. To our knowledge, there is no large-scale longitudinal study in Japan that evaluated potential risk factors, including depression, for death, hospitalization, and lack of continuous employment in patients with schizophrenia.

In this study, we evaluated risk factors for the important clinical outcomes of death, hospitalization, and lack of continuous employment in patients with schizophrenia using a large claims database in Japan.

In this study, risk factors for the important clinical outcomes of patients with schizophrenia were evaluated by nested case–control design using a large-scale, health insurance-based database in Japan. Depression was identified as a risk factor for death, hospitalization, and sick leave. CCI score and prescription for laxatives were identified as risk factors for death and sick leave, indicating that physical comorbidities and constipation are risk factors for these important clinical outcomes. Constipation was also found as a risk factor for hospitalization, and extrapyramidal symptoms (EPS), indicated by anticholinergics prescription, was determined as a risk factor for hospitalization and resignation.

In this study, the proportions of patients with depression in the death, hospitalization, resignation, and sick leave cohorts were 24.5%, 23.2%, 29.9%, and 30.1%, respectively. The proportions were similar to a previous report that evaluated the prevalence of depression by meta-analysis [15]. As the study did not include Japanese data, this suggests that depression among patients with schizophrenia is a common symptom worldwide. A previous study reported that depression is a critical symptom of schizophrenia and a risk factor for suicide [11]. Suicide is associated with many life-years loss, but it is not the major cause of death in patients with schizophrenia [10, 13, 21]; therefore, it is notable that depression for patients with schizophrenia was identified to be a risk factor for all-cause death, including suicide, in this study. This finding suggests that depression affects deaths caused by accidents and physical disease, as depression lowers the self-management abilities of patients with schizophrenia and leads to an unhealthy lifestyle. In our study, hypertension and diabetes, which are well-known to lead to cardiovascular events, were not detected as risk factors for death, which may be because the eligible patients in our study were under 60 years old. On the other hand, our results indicated that dyslipidemia was associated with a reduced risk of death. This finding was unexpected, as dyslipidemia is also known to be a risk factor for cardiovascular events. Further studies are needed to confirm this result.

Identified risk factors for hospitalization to a psychiatric ward were not only psychiatric conditions, such as depression and sleep disturbance, but also physical conditions such as constipation and EPS. Our findings support a previous study that depression increases the risk of hospitalization and rehospitalization in patients with schizophrenia [17] and a previous systematic review that reported that sleep disturbance was associated with positive symptoms and poor clinical outcome [22]. EPS and constipation are common side effects of antipsychotic medication. In clinical practice, patients who experience side effects are required to reduce dosage or switch to other antipsychotics [23]. Therefore, patients with EPS or constipation may have difficulty managing psychotic symptoms. EPS induces and overlaps with depressive symptoms [24], and constipation is common in depressed individuals [25]. Moreover, sleep disturbance is strongly associated with depression [26]. Thus, depression may be a key risk factor for hospitalization, although these are independent risk factors. Lifestyle diseases, such as hypertension and diabetes, were not identified as risk factors for hospitalization. The results are clinically understandable because patients with schizophrenia are unlikely to be hospitalized to a psychiatric ward due to lifestyle diseases in Japan.

Cognitive function impairment, negative symptoms, and education were reported to be predictors of employment status in patients with schizophrenia [27]. However, few studies have examined risk factors in patients with schizophrenia for continuous employment-related outcomes such as sick leave and resignation. A case–control study in Germany reported that schizophrenia severity, assistance or care in everyday life, age, and typical antipsychotic treatment were associated with early retirement in patients with schizophrenia; however, depression was not included as a candidate risk factor in the analysis [28]. Our study found that depression was not associated with resignation, but was a significant risk factor for sick leave from work. To our knowledge, this is the first study to identify depression as a risk factor for sick leave in patients with schizophrenia. However, the results were unexpected because we hypothesized that depression would be both a risk factor for resignation and sick leave. There may be differences of the characteristics between patients who took sick leave and those who resigned. Although we did not evaluate duplicate cases in sick leave and resignation, the ratio of depression was higher in cases of sick leave than resignation (37.8% vs. 29.3%), and the number of the resignation cases (n = 811) was only about one-fifth of the sick leave cases (n = 4,590). Similar to our results, a cohort study composed of employees in 11 Japanese large private companies reported that among the employees with schizophrenia who had sick leave, the resignation rate was only 28.6% within 3.5 years and 66.7% returned to work [29]. In the Netherlands, which has a different medical and social environment from Japan, 76% of employees taking sick leave due to schizophrenia returned to work within 2 years [30]. Thus, most employees who take sick leave due to schizophrenia are able to return to work. These findings may explain why depression was a risk factor for sick leave but not resignation in the present study. We also identified EPS as a risk factor for resignation. The result is similar to the previous German case–control study that reported that the proportion of patients with EPS was about twice that among patients who took early retirement compared to those that did not [28].

Medical and social environments are markedly differ among countries, and are thought to affect our focusing outcomes, especially employment-related outcomes. Thus, in this study, we investigated risk factors in patients with schizophrenia for resignation and sick leave from work using a Japanese database, and found that depression was a risk factor for sick leave. Furthermore, sleep disturbance, constipation, and physical comorbidities (CCI score), which may be related to depression, were indicated to be risk factors for sick leave.

Since this study is based on medical claims data, there are several limitations due to data availability and characteristics. The first limitation is the definition of resignation. We defined resignation based on the reason for withdrawal from the private company’s health insurance. Thus, patients who quit the company and work in another company may have been misclassified, which may have led to underestimating the association of potential risk factors. This may explain why depression and constipation were not associated with resignation despite being strong risk factors for the other three outcomes. Second, we could not evaluate the following factors, which are thought to affect the outcomes that we investigated: social environmental factors, such as continuous employment support service, employment systems, and living alone, and lifestyle factors, such as physical activity and eating habits; and psychotherapy or counseling intervention. Third, since adherence to antipsychotics was suggested to be a strong risk factor for death, hospitalization, and early retirement [8, 9, 28], we included only people with a diagnosis of schizophrenia and prescription data for antipsychotics as case or control, but we could not evaluate their medication adherence status due to data characteristics, and we did not focus on the details of antipsychotics, such as type, dosage, and the number of antipsychotics prescribed to the patients. Fourth, we could not distinguish detailed depressive features, such as post-psychotic depression and depressed symptoms due to negative symptoms in patients with schizophrenia.

This study suggests that depression in patients with schizophrenia is a risk factor for important clinical outcomes such as death, hospitalization, and sick leave from work. Also, physical conditions, such as constipation and EPS, are risk factors for the important clinical outcomes that we investigated. These findings suggest that healthcare professionals should carefully pay attention to both psychiatric symptoms and physical symptoms for the treatment of patients with schizophrenia.