Prevalence of violence to others among individuals with schizophrenia in China: A systematic review and meta-analysis

BACKGROUND: Violence to others (hereinafter referred to as “violence-TO”) is common in individuals with schizophrenia. The reported prevalence of violence-TO among schizophrenics ranges widely in existing studies. Improved prevalence estimates and identification of moderators are needed to guide future management and research. METHODS: We searched EBSCO, EMBASE, Medline, PubMed, Science Direct, Web of Science, CNKI, VIP, WANFANG data, and CBM for relevant articles published before June 5, 2022. Meanwhile, violence-TO was summarized into four categories: (a) violence-TO on the reviews of official criminal or psychiatric records (type I); (b) less serious forms of violence-TO (type II); (c) physical acts causing demonstrable harm to victims (type III); (d) homicide (type IV). We did meta-analysis for the above types of violence-TO, respectively, and applied subgroup analyses and meta-regression analyses to investigate the source of heterogeneity. RESULTS: A total of 56 studies were eligible in this study and 34 of them were high-quality. The prevalence of type I to type IV in individuals with schizophrenia in China was 23.83% (95% CI: 18.38–29.75%), 23.16% (95% CI: 8.04–42.97%), 17.19% (95%CI: 8.52–28.04%), and 0.62% (95% CI: 0.08–1.54%) respectively. The results of the subgroup analysis showed that the prevalence of type I was higher among subjects in the inland than in the coastal non-economic zone, while the prevalence of type III was the highest in the coastal economic zone, followed by the inland region and the lowest in the coastal non-economic zone. The results of multivariate meta-regression analyses showed that: patient source in type I (β = 0.15, P < 0.01), patient source (β = 0.47, P < 0.01), and proportion of male (β = 0.19, P < 0.01) in type II, age (β = 0.25, P < 0.01), and GDP per capita (β = 0.05, P = 0.01) in type III were statistically significant. CONCLUSION: The prevalence of different types of violence-TO and their influencing factors varied. Therefore, the authorities should take different management measures. In addition to individual factors, regional factors may also affect violence-TO, which suggests the need for a multi-sectorial approach to prevention and treatment for subjects in different regions and adopting targeted control strategies. SYSTEMATIC REVIEW REGISTRATION: [www.ClinicalTrials.gov], identifier [CRD42021269767].