Long-term trends of mortality and suicide among persons with schizophrenia

Among the general population in China, life expectancy was 69.7 years for the total population in 1994, with women having longer life expectancy (71.9 years, mortality rate 104.72 per 1000) than men (67.7 years, mortality rate 148.53 per 1000).^26,27 Between 1994 and 2015, 38.4% of our cohort of people with schizophrenia died. The life expectancy of our participants was 58.5 years for women and 50.6 years for men, which was 13.4 and 17.1 years shorter than that for women and men in the general population in China during the same time period. Although there is a lack of studies examining mortality and suicide in the Chinese population in the 1990s, comparing our results with the weighted average life expectancies in persons with schizophrenia (overall, 64.7 years; female, 67.6 years; male, 59.9 years) in a recent meta-analysis^{Reference Hjorthoj, Sturup, McGrath and Nordentoft2} shows that the life expectancy of persons with schizophrenia in our study was lower by around 9 years. The gender gap in life expectancy identified in our study, where males showed a shorter life expectancy than females, is consistent with prior research.^{Reference Tiihonen, Lonnqvist, Wahlbeck, Klaukka, Niskanen and Tanskanen1,Reference Hjorthoj, Sturup, McGrath and Nordentoft2,Reference Laursen5,Reference Crump, Winkleby, Sundquist and Sundquist28} We also noted that the age at diagnosis in our study is older relative to the literature for both males and females, which may be attributed to unique barriers associated with early detection and treatment in rural China, including poverty, homelessness, lack of health insurance, lack of informed choices of treatment and stigma associated with mental illness.^{Reference Tiihonen, Lonnqvist, Wahlbeck, Klaukka, Niskanen and Tanskanen1–Reference Chesney, Goodwin and Fazel3}

The suicide rate in China was estimated to be 23 per 100 000 people between 1995 and 1999, with the female suicide rate was 25% higher than that among males.^{Reference Phillips, Li and Zhang13} A recent study found that suicide rates in China significantly decreased between 2002 and 2015 (by 16.0% per year until 2006, then by 3.4% per year until 2015), although rates among males were substantially higher than those among females after 2006.^{Reference Jiang, Niu, Hahne, Hu, Fang and Shen29} In our study, 13.8% of the all-cause deaths among persons with schizophrenia were attributed to suicide, which is higher than the rate for the general population. The estimated lifetime suicide risk was 5.3% in persons with schizophrenia in this study, which is similar to previous estimates (4.9%).^{Reference Palmer, Pankratz and Bostwick9} A cause-deleted life table calculated the gain in life-years after a hypothetical prevention of a suicide, allowing us to measure the significance of the lifetime risk of suicide for mortality among persons with schizophrenia. The result showed that life expectancy of the sample would have significantly increased in the absence of a suicide burden. Our findings extend previous studies and further address the large gap in mortality and suicide rates between persons with schizophrenia and the general population.^{Reference Phillips, Li and Zhang13,Reference Phillips, Yang, Li and Li14,Reference Ran, Chan, Chen, Tang, Lin and Li24}

Predictors of mortality and suicide among persons with schizophrenia

Our results show significant effects of age and gender on mortality and suicide in persons with schizophrenia, and unique predictors for all-cause mortality and suicide. Poor family care attitudes towards participants was found to predict greater mortality and death due to other causes, but not death due to suicide. The effect of poor family attitudes on increasing rates of all-cause mortality and death due to other causes was unique for rural China, where family caregiving and support play a major role in patients’ treatment and care.^{Reference Ran, Wong, Yang, Ho, Mao and Li30,Reference Ran, Chui, Wong, Mao, Lin and Liu31} Evidence shows that having poorer familial attitudes towards persons with schizophrenia might be associated with living without family caregivers, poorer mental status, lower rates of remission, less access to antipsychotic medications and lower social functioning,^{Reference Ran, Chui, Wong, Mao, Lin and Liu31} which in turn increase the risks of all-cause mortality and death by accidents and other causes. Evidence shows that socioeconomic development has shaped the treatment status of persons with schizophrenia in rural China.^{Reference Ran, Weng, Liu, Zhang, Yu and Peng32} Compared with patients in 1994, persons with schizophrenia in 2015 were more likely to have fewer or no family members and family caregivers, which in turn was associated with never being treated.^{Reference Ran, Weng, Liu, Zhang, Yu and Peng32} Besides, poor family attitudes could increase the internalised stigma in persons with schizophrenia, which might worsen their long-term outcome in rural China.^{Reference Ran, Zhang, Wong, Yang, Liu and Liu16} This is particularly relevant to the Chinese context, given that the majority of persons with schizophrenia (over 90%) are cared for by family caregivers at home.^{Reference Ran, Weng, Chan, Chen, Tang and Lin18,Reference Ran, Chan, Ng, Guo and Xiang33}

Previous suicide attempts were identified as a predictor of 21-year suicide in this study, which is consistent with previous studies.^{Reference Ran, Zhang, Wong, Yang, Liu and Liu16,Reference Ran, Weng, Liu, Zhang, Yu and Peng32} Not being admitted to hospital was also found to be a predictor of 21-year suicide. Studies in Western countries have suggested that death by suicide among people with severe mental illness peaks during the first year after hospital admission, partly due to the failure of the health system to identify and treat physical diseases during admission for mental disorder.^{Reference Ösby, Correia, Brandt, Ekbom and Sparén7,Reference De Hert, Cohen, Bobes, Cetkovich-Bakmas, Leucht and Ndetei34} However, many persons with schizophrenia are not admitted in rural China (over 60%), owing to family poverty and poor mental health services, and only some with more severe symptoms are likely to be admitted to hospital.^{Reference Ran, Chen, Conwell, Chan, Yip and Xiang12,Reference Ran, Yu, Peng, Liu, Zhang and Hu17,Reference Ran, Weng, Chan, Chen, Tang and Lin18} Many persons with schizophrenia with aggressive and suicidal behaviour are not treated appropriately (e.g. by admission to hospital, regular medication).^{Reference Ran, Zhang, Wong, Yang, Liu and Liu16} The results of this study show the importance of hospital admission and antipsychotic medication for prevention of suicide in persons with schizophrenia in rural China.^{Reference Ran, Chen, Conwell, Chan, Yip and Xiang12,Reference Ran, Weng, Chan, Chen, Tang and Lin18}

Positive family history of psychosis, however, was found to be related to lower suicide rates, which was not observed in a 14-year follow-up study in rural China.^{Reference Ran, Xiao, Zhao, Zhang, Yu and Mao11} This suggests that family history of psychosis may have significant effect on reducing suicide risk in the longer term. Family history of psychosis was found to be associated with younger age at onset of psychosis, although such difference did not hold for long-term psychiatric symptoms.^{Reference Ran, Xiao, Zhao, Zhang, Yu and Mao11} It is possible that with longer duration of psychosis, patients and their families gained greater awareness of illness and risks of suicide, which might act as a protective factor against their likelihood of death by suicide.

The results of this study showed that males consistently showed higher rates of mortality and suicide than females, and the gap had been widening over the 21 years, a finding consistent with previous literature in Western countries.^{Reference Hjorthoj, Sturup, McGrath and Nordentoft2,Reference Lee, Liu, Tu, Palmer, Eyler and Jeste6,Reference Ösby, Correia, Brandt, Ekbom and Sparén7,Reference Brown, Kim, Mitchell and Inskip35} The poor long-term prognosis for men in this study might be explained by the higher rates of being single or divorced, poor family economic status, earlier age at onset of schizophrenia, lack of family caregivers, poor treatment status and higher criminal behaviours.^{Reference Ran, Mao, Chan, Chen and Conwell19,Reference Ran, Chui, Wong, Mao, Lin and Liu31} Higher rates of suicide among males than females support the alarming sign of suicide among men in previous literature, including our 14-year follow-up studies.^{Reference Phillips, Yang, Li and Li14,Reference Ran, Mao, Chan, Chen and Conwell19,Reference Lester36} This 21-year follow-up study provided further evidence that the large gap in suicides by gender was widening in the long term among rural residents in China.

The results of this study indicated that older persons with schizophrenia had significantly higher all-cause mortality rates than their younger counterparts, although the risk of suicide decreased with age. This study supports the conclusion that age predicts mortality and suicide differently, which is consistent with previous findings in Western countries.^{Reference Olfson, Gerhard, Huang, Crystal and Stroup4,Reference Ösby, Correia, Brandt, Ekbom and Sparén7,Reference Palmer and Pankratz V37} Differences in mortality and suicide risk estimates are likely related to variations in disease distributions by age: older adults with schizophrenia may have greater exposure to physical diseases, such as cardiovascular disease, diabetes mellitus and respiratory disease.^{Reference Olfson, Gerhard, Huang, Crystal and Stroup4,Reference De Hert, Cohen, Bobes, Cetkovich-Bakmas, Leucht and Ndetei34,Reference De Hert, Correll, Bobes, Cetkovich-Bakmas, Cohen and Asai38,Reference Taipale, Tanskanen, Mehtälä, Vattulainen, Correll and Tiihonen39}

Suicide, however, was a higher risk among younger adults, a result that is consistent with some prior studies conducted in the Western context.^{Reference Palmer, Pankratz and Bostwick9,Reference Pompili, Amador, Girardi, Harkavy-Friedman, Harrow and Kaplan40} Evidence has shown that persons with schizophrenia often do not receive regular antipsychotic treatment (e.g. clozapine) in rural China.^{Reference Ran, Zhang, Wong, Yang, Liu and Liu16} Severe self-stigma and lack of appropriate treatment during the first onset of illness, especially in rural China where the availability and accessibility of mental health services are limited,^{Reference Chan, Chan, Pang, Yan, Hui and Chang10,Reference Phillips, Li and Zhang13,Reference McGrath, Saha, Chant and Welham15} may be significant factors in suicide among the younger participants.^{Reference Ran, Chen, Conwell, Chan, Yip and Xiang12,Reference Ran, Chan, Chen, Mao, Hu and Tang20} The results of this study further address the importance of developing suicide prevention strategies through community-based mental healthcare to provide earlier diagnosis, antipsychotic treatment, anti-stigma intervention, family support and social functioning rehabilitation.^{Reference Ran, Chen, Conwell, Chan, Yip and Xiang12,Reference Ran, Chan, Chen, Tang, Lin and Li24}

Implications for policy and services

Our results have implications for healthcare policy and services.^{Reference Fleischhacker, Arango, Arteel, Barnes, Carpenter and Duckworth41} First, the findings can contribute to the understanding of long-term mortality and suicide of persons with schizophrenia, which is important for mental health professionals and policymakers to prevent and reduce mortality and suicide, and promote mental recovery of persons with schizophrenia. Specifically, we identified that men had higher rates of mortality than women, older adults had higher mortality rates and younger adults had higher suicide rates than their counterparts. These findings support efforts to take the demographic differences into account when developing interventions. Males with schizophrenia in rural China comprise a highly vulnerable group who, in addition to psychiatric treatment (e.g. hospital admission, regular antipsychotic treatment), need more support from family, community and society. It is thus important to integrate early diagnosis, treatment, community-based care and family support to improve mental recovery in the rural context. Healthcare practitioners, especially primary healthcare professionals, should be encouraged to strengthen gender-specific mental health intervention by providing targeted resources and treatment when designing intervention programmes in rural China.^{Reference Ran, Mao, Chan, Chen and Conwell19} Given the serious physical diseases associated with schizophrenia, mental health professionals are encouraged to incorporate physical care in settings treating persons with schizophrenia. Preventions and interventions that aim at improving monitoring of blood pressure and serum glucose level, and managing obesity, hypertension and other cardiovascular risk factors should be further supported to reduce the longevity gap between people with schizophrenia and the general population.^{Reference Olfson, Gerhard, Huang, Crystal and Stroup4,Reference Lee, Liu, Tu, Palmer, Eyler and Jeste6}

Second, it is critical to design culturally and contextually tailored community-based mental health interventions for persons with schizophrenia to reduce the mortality and suicide in rural China. Since poor family attitudes toward persons with schizophrenia were identified as important predictors for all-cause mortality and deaths due to other causes, culturally tailored interventions that target improving family support, reducing stigma of mental illness in patients and their family caregivers, enhancing positive contact and facilitating mental health promotion could be promising steps to increase the effectiveness of interventions.^{Reference Ran, Chui, Wong, Mao, Lin and Liu31,Reference Ran, Chan, Ng, Guo and Xiang33}

Third, our findings highlight the importance of early interventions for persons with schizophrenia in the efforts to reduce their risks of premature death and death by suicide. Given the limited health resources in rural China, primary healthcare professionals should be trained further to provide community-based mental health services in rural areas. Although duration of illness and never receiving treatment were not risk factors for long-term premature death and suicide in this study, it is crucial to improve early diagnosis, treatment, family-based intervention, community-based mental health services and recovery programmes for persons with schizophrenia in rural China to address the disparity in mortality and suicide.^{Reference Bellack42} Given that many persons with schizophrenia do not receive regular antipsychotic treatment in rural China, systematic and regular antipsychotic treatment (e.g. clozapine) should be a crucial aim for reducing suicide among these individuals.^{Reference Tiihonen, Lonnqvist, Wahlbeck, Klaukka, Niskanen and Tanskanen1,Reference Phillips, Yang, Li and Li14}

Strengths and limitations

The strengths of this study include the use of three waves of follow-up after a baseline investigation of community samples with schizophrenia, consistent diagnostic criteria over time and high follow-up rates during the 21-year duration.

The limitations include possible recall bias and loss of cases during the long-term follow-up with participants and informants, although such bias may have been reduced by the use of multiple follow-up data sources (i.e. trained psychiatrist interviewers, medical records, death certification). Second, informants might not have reported death by suicide owing to the stigma attached to suicide in the Chinese context. Cross-validation of data collected, however, might reduce such response bias. Third, the number of those within the study who died by suicide was low (n = 27). Thus, the results should be interpreted with caution because of the small numbers; furthermore, causality cannot be established owing to lack of a control group. Last, socioeconomic development in China over the 21 years has greatly changed the country's mental health services, including access to treatment and the nature of the treatments received, access to other services and quality of life more generally. Findings of this study may not apply to settings where such changes have not been observed.