Abstract:
To develop an instrument for assessing violence risk among Thai persons with schizophrenia in the community, the Thai violence risk scale (TVRS). The characteristics and circumstances associating with violence among persons with schizophrenia in the community were selected based on literature review and the Psychology of Criminal Conduct theory. Ten steps of scale development procedures by Crocker and Algina (1986) were used. The TVRS, a 17-item, was an alternate choice, yes-no question for the face-to-face interview instrument. Its construct validity was examined by both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). Findings from the EFA (varimax rotation, n = 300) showed that the TVRS was composed of 2 factors. Factor I was the characteristics (15 items) and factor II was the circumstances (2 items). The second order CFA (n = 604) provided the goodness of fit indices illustrating that the violence risk model fitted with the empirical data. These findings supported the construct validity of the TVRS. Predictive validity of the TVRS was studied in 128 persons with schizophrenia in the community. Its AUC was .88 (p<.001). With cut-off score of 23, the sensitivity was .80; the specificity was .79; and the positive predictive value was .64. The reliability, internal consistency, of the TVRS was also acceptable. Its Cronbach’s alpha coefficient was .89. With enough evidence to support its psychometric properties and it takes only 5 minutes to complete the TVRS, this instrument is suitable for assessing violence risk in persons with schizophrenia in the community. However, the TVRS is a newly development instrument, it requires further study to make it a standard instrument.
