Prevalence of Violence to Others Among Patients With Schizophrenia in China: A Systematic Review and Meta-Analysis

Background: Violence to others (hereinafter referred to as "violence") are common in people with schizophrenia. However, the reported prevalence of violence among patients with schizophrenia ranges widely in existing original studies. Improved prevalence estimates and identification of moderators are needed to enhance recognition and care, and to guide future management and research.Methods: In this systematic review and meta-analysis, with language limited to English or Chinese, we searched EBSCO, EMBASE, Medline, PubMed, Science Direct, Web of Science, CNKI, VIP, WANFANG data and CBM for relevant articles published before June 30, 2021. Meanwhile, violence behaviors were summarized into four categories: first, violence on the basis of reviews of official criminal or psychiatric records (hereinafter referred to as "type Ⅰ"); second, less serious forms of violence, such as minor physical acts (i.e., pushing) or verbal behavior (i.e., threats to harm someone) (hereinafter referred to as "type Ⅱ"); third, physical acts that caused demonstrable harm to victims (hereinafter referred to as "type Ⅲ"); fourth, homicide (hereinafter referred to as "type Ⅳ"). We did meta-analysis for the above four types of violence respectively using the high-quality studies and applied subgroup analyses and meta-regression analyses to investigate the source of heterogeneity. Findings: 51 studies were eligible for inclusion in this study and 31 of which were considered to be high-quality. The prevalence of type Ⅰ to type Ⅳ in patients with schizophrenia in China was 22·65% (95%CI: 17·61%-28·12%), 28·32% (95%CI: 11·02%-49·70%), 15·25% (95%CI: 6·83%-26·16%), 0·62% (95%CI: 0·08%-1·54%) respectively. The results of the subgroup analysis showed that the prevalence of each type of violence was higher among inpatients than among non-inpatients by source; the prevalence of type II and type III violence was higher among patients aged under 40; and the prevalence of type I violence was higher among inpatients in the inland than in the coastal non-economic zone, while the prevalence of type III violence was the highest in the coastal economic zone, followed by the inland region and the lowest in the coastal non-economic zone, both among inpatients and overall without differentiating the source of the patients. The results of multivariate meta-regression analyses showed that: patient source in type Ⅰ (β=0·1483, p=0·0087), age in type Ⅱ (β=0·4950, p<0·0001), age (β=0·4950, p<0·0001) and GDP per capital (β=0·0480, p=0·0140) in type Ⅲ were statistically significant.Interpretation: The prevalence of different types of violence and their influencing factors varied largely. Therefore, for different types of violence, the relevant government departments should take different management measures. In addition to the individual factors of patients, regional factors may also affect violence, which suggests the need for a multi-sectoral approach to prevention and treatment for patients in different regions and adopting targeted control strategies. At the same time, future epidemiological researches should further explore the correlation between regional factors and violence in people with schizophrenia.Registration Details: This study is registered in PROSPERO with number CRD42021269767.Funding Information: National Natural Science Foundation of China [grant number 81903414]; Science and Technology Program of Sichuan Province [grant number 2020YFS0216 and 2019YJ0076].Declaration of Interests: All other authors declare no competing interests.