Abstract: Objective · To develop Acquired Capability for Suicide Scale and examine its reliability and validity in medical undergraduates. Methods · Convenient sampling was used. One thousand one hundred and eight medical students were selected and assessed with Acquired Capability for Suicide Scale, forty-two of which were reassessed after 2 months. SPSS 20.0 and AMOS 22.0 were used for analysis. Pearson correlation coefficient and content validity index (CVI) were used to examine the content validity; factor analysis and independent sample t-test were used to examine construct validity and differentiated validity. Reliability was assessed by internal consisitent reliability (Cronbach's α), split-half reliability (Spearman-Brown coefficient) and retest reliability (intraclass correlation coefficient, ICC). Results · The average score of Acquired Capability for Suicide Scale of the total 1 094 participants was (21.02±10.86). Three hundren and eight participants (28.15%) had suicide ideation, 68 had suicide attempt, and 38 used to have suicide behavior. Item-level CVIs (I-CVIs) were all higher than 0.80, and scale-level CVI (S-CVI) was 0.97. The correlation coefficients (r) of each subscale and its items ranged from 0.842 to 0.916. Kaiser-Meyer-Olkin (KMO) statistic was 0.691 and P value of Bartlett Test of Sphericity was less than 0.001, so the result showed the data was suitable for factor analysis. Principal component analysis and the maximum variance orthogonal rotation were used to extract 3 common factors with eigenvalue ≥ 1. The cumulative variance contribution rate was 76.01%. Each fitting index of the modified confirmatory factor analysis showed that the model fitted well [χ2/df=3.652, root mean square error of approximation (RMSEA)=0.070, comparative fit index (CFI)=0.977, goodness of fit index (GFI)=0.979, normed fit index (NIF)=0.969, incremental fit index (IFI)=0.977], and all the factors' loadings were over 0.6. The total score of suicide attempt group (33.85±11.42) was higher than that of non-suicide attempt group (27.60±9.35) (t=4.621, P=0.000), and the total score of suicide behavior group (36.18±12.60) was higher than that of non-suicide behavior group (27.97±9.37) (t=3.870, P=0.000). The Cronbach’s α coefficient of Acquired Capability for Suicide Scale was 0.794, the Spearman-Brown coefficient was 0.729, and ICC was 0.846. Conclusion · Acquired Capability for Suicide Scale includes 3 dimensions: self-perceived suicidal ability, impulsivity and tolerance of pain. The validity and reliability of Acquired Capability for Suicide Scale was good among Chinese medical undergraduates. The scale can be used to identify whether an individual, especially the people with suicide ideation, has the ability to commit suicide.

Key words: suicide, capability, reliability, validity, medical student