Troisi Alfonso, Kustermann Stefano, Di Genio Massimo, Siracusano Alberto
Department of Neurosciences, University of Rome Tor Vergata, Rome, Italy.
J Clin Psychiatry. 2003 Dec;64(12):1460-4. doi: 10.4088/jcp.v64n1210.
A critical step for improving the prediction of on-ward violence is the identification of variables that are not only consistently associated with an increased risk of aggression but also easily evaluated during the admission interview. The goal of this prospective study was to assess the predictive utility of hostility during admission interview.
The sample consisted of 80 newly admitted male patients with heterogeneous DSM-IV psychiatric diagnoses recruited from the psychiatric ward of an urban public hospital. Psychiatric symptoms at admission were assessed with the Brief Psychiatric Rating Scale (BPRS). Aggressive behavior during the first week of hospitalization was measured with the Modified Overt Aggression Scale. Data were collected between January and June 1998.
In a multiple regression model, BPRS items hostility and tension-excitement emerged as significant predictors of verbal aggression, whereas thinking disturbance (high) and suspiciousness-uncooperativeness (low) emerged as significant predictors of aggression against objects. In contrast, when aggression was treated as a binary dependent variable in a logistic model, hostility during the admission interview had no utility in predicting on-ward aggressive behavior.
This study confirms the importance of distinguishing between different types of aggression to improve the accuracy of predictions of violence. The findings suggest that the question whether hostility is a useful short-term predictor of aggression in psychiatric inpatients cannot be answered conclusively. The predictive utility of hostility was relatively high for predicting verbal aggression but was negligible for predicting other types of aggressive behavior.
改善对未来暴力行为预测的关键一步是识别那些不仅始终与攻击风险增加相关，而且在入院访谈期间易于评估的变量。这项前瞻性研究的目的是评估入院访谈期间敌意情绪的预测效用。
样本包括从一家城市公立医院精神科病房招募的80名新入院的男性患者，他们被诊断患有不同类型的DSM-IV精神疾病。入院时的精神症状用简明精神病评定量表（BPRS）进行评估。住院第一周的攻击行为用改良的明显攻击量表进行测量。数据收集于1998年1月至6月期间。
在多元回归模型中，BPRS项目中的敌意和紧张-兴奋是言语攻击的显著预测因素，而思维紊乱（高）和猜疑-不合作（低）是针对物体攻击的显著预测因素。相比之下，当在逻辑模型中将攻击行为作为二元因变量时，入院访谈期间的敌意情绪对预测未来攻击行为没有效用。
本研究证实了区分不同类型攻击行为对于提高暴力行为预测准确性的重要性。研究结果表明，关于敌意情绪是否是精神科住院患者攻击行为的有用短期预测因素这一问题，无法得出确凿答案。敌意情绪在预测言语攻击方面的预测效用相对较高，但在预测其他类型攻击行为方面可忽略不计。
