Among people with various mental illnesses, those with schizophrenia have the highest incidence of aggressive behavior, which not only disturbs the medical environment, but also poses a great threat to family stability and social harmony. Previous regression analysis studies have shown that childhood trauma, self-esteem and resilience are influencing factors of aggressive behavior in schizophrenia patients, but few studies have explored the internal relationship between these four variables.

To explore the relationship between childhood trauma, self-esteem, resilience and aggressive behavior in schizophrenic patients using a self-constructed chain mediation model, and to clarify its internal mechanism.

A convenience sample of stable schizophrenia inpatients were selected from Daqing Third Hospital and Chifeng Anding Hospital from March 2021 to July 2022. A survey was conducted with them using a questionnaire consisting of general information questionnaire, Positive and Negative Symptom Scale (PANSS) , Childhood Trauma Questionnaire-Short Form (CTQ-SF) , Connor-Davidson Resilience Scale (CD-RISC) , Self-esteem Scale (SES) , and Aggressive Questionnaire (AQ) . Pearson correlation analysis was used to explore the correlation of AQ score with CTQ-SF, CD-RISC and SES scores. Multiple linear regression analysis was used to explore the influencing factors of aggressive behavior. AMOS 24.0 was used to construct the chain mediation model. The mediation effect of the chain mediation model was tested using Bootstrap method.

Three hundred and ten cases attended the survey, and 300 of them (96.8%) who responded effectively were finally included for analysis. The average scores of PANSS, CTQ-SF, SES, CD-RISC and AQ were (18.4±3.7) , (49.7±13.5) , (26.9±6.9) , (55.2±15.5) and (84.3±19.8) , respectively. Multiple linear regression analysis showed that childhood trauma〔B=0.551, 95%CI (0.421, 0.681) , P<0.001〕, self-esteem〔B=-0.661, 95%CI (-0.908, -0.413) , P<0.001〕and resilience〔B=-0.448, 95%CI (-0.561, -0.335) , P<0.001〕 were influencing factors of aggressive behavior. Pearson correlation analysis revealed that CTQ-SF scores were positively correlated with AQ scores (r=0.695, P<0.01) and negatively correlated with SES scores and CTQ-SF scores (r=-0.524, P<0.01; r=-0.565, P<0.01) ; SES scores were negatively correlated with AQ scores (r=-0.607, P<0.01) and positively correlated with CD-RISC scores (r= 0.522, P<0.01) ; the CD-RISC score was negatively correlated with the AQ scores (r=-0.681, P<0.01) . The chain mediation model test results showed that the goodness-of-fit of the model was good. Bootstrap test showed that the 95%CI of each mediation path did not include 0, and the mediation effect was significant. The overall effect of childhood trauma on aggressive behavior was significant〔β=0.810, 95%CI (0.743, 0.871) , P<0.001〕. The indirect effect of childhood trauma on aggressive behavior, that is, acting through the mediators of self-esteem and resilience, was significant〔β=0.473, 95%CI (0.362, 0.598) , P<0.001〕. The direct effect of childhood trauma on aggressive behavior was significant〔β=0.377, 95%CI (0.179, 0.485) , P<0.001〕.

Self-esteem and resilience play a chain mediating role between childhood trauma and aggressive behavior in patients with schizophrenia. In the future, a risk prediction model of aggressive behavior in schizophrenia patients can be constructed to assess the risk of aggressive behavior, which is contributive to early identification and management of those at risk of aggression, thereby reducing or preventing the occurrence of aggressive behavior.

在各类精神疾病中，精神分裂症患者攻击行为发生率最高，攻击行为不仅扰乱医疗环境，也对家庭稳定和社会和谐造成巨大威胁。既往的回归分析研究表明童年创伤、自尊、心理弹性是精神分裂症患者攻击行为的影响因素，但当前较少有研究探讨上述4个变量之间的内在联系。

构建链式中介模型探讨精神分裂症患者童年创伤、自尊、心理弹性及攻击行为的关系，阐明其内在作用机制。

依据便利抽样原则，选取2021年3月—2022年7月在大庆市第三医院和赤峰市安定医院住院的稳定期精神分裂症患者为研究对象。对研究对象进行问卷调查，问卷内容包括一般情况调查表、阳性与阴性症状量表（PANSS）、童年创伤量表（CTQ-SF）、心理弹性量表（CD-RISC）、自尊量表（SES）、攻击行为问卷（AQ），采用Pearson相关分析探讨CTQ-SF、CD-RISC、SES和AQ得分的相关性，并采用多元线性回归分析探究精神分裂症患者攻击行为的影响因素。采用AMOS 24.0进行链式中介模型的构建，并采用Bootstrap方法对所构建的链式中介模型进行中介效应检验。

本研究共发放问卷310份，回收有效问卷300份，有效回收率为96.8%。患者PANSS、CTQ-SF、SES、CD-RISC、AQ平均得分为（18.4±3.7）、（49.7±13.5）、（26.9±6.9）、（55.2±15.5）、（84.3±19.8）分。多元线性回归分析结果显示，精神分裂症患者童年创伤〔B=0.551，95%CI（0.421，0.681），P<0.001〕、自尊〔B=-0.661，95%CI（-0.908，-0.413），P<0.001〕、心理弹性〔B=-0.448，95%CI（-0.561，-0.335），P<0.001〕是攻击行为的影响因素。Pearson相关分析结果显示，CTQ-SF得分与AQ得分呈正相关（r=0.695，P<0.01），与SES得分、CD-RISC得分呈负相关（r=-0.524，P<0.01；r=-0.565，P<0.01）；SES得分与AQ得分呈负相关（r=-0.607，P<0.01），与CD-RISC得分呈正相关（r=0.522，P<0.01）；CD-RISC得分与AQ得分呈负相关（r=-0.681，P<0.01）。链式中介模型检验结果显示模型的拟合度和适配度良好。Bootstrap分析结果显示，各中介路径95%CI均不包括0，中介效应显著。童年创伤对攻击行为的总效应显著〔β=0.810，95%CI（0.743，0.871），P<0.001〕；童年创伤对攻击行为的间接效应，即自尊、心理弹性的中介效应显著〔β=0.473，95%CI（0.362，0.598），P<0.001〕；童年创伤对攻击行为的直接效应显著〔β=0.377，95%CI（0.179，0.485），P<0.001〕。

自尊与心理弹性在精神分裂症患者童年创伤与攻击行为之间起链式中介作用。未来可构建精神分裂症患者攻击行为风险预测模型，对患者进行攻击行为风险评估，早期识别有攻击风险的患者并对风险人群进行管理，以减少和预防攻击行为的发生。