自杀意图客观强度量表的效度和信度

Validity and reliability of Objective Scale for Suicidal Intent

目的：评估改编的自杀意图客观强度量表（OSSI）应用于自杀死亡和未遂研究的效度和信度。方法：纳入151例自杀死亡者（自杀组）和120例自杀未遂者（未遂组）。两组每个案例的同吃住家属及另一位亲友分别接受独立访谈,自杀未遂者另接受独立访谈,通过访谈搜集一般资料并由调查员填写OSSI。分析量表的结构效度、内部一致性、不同信息来源[家属、亲友、本人（未遂组）]之间的一致性,并比较不同性别、不同自杀结局和不同自杀方式之间的OSSI评分。结果：无论何种信息来源,量表均是二因子结构（自杀环境因子和自杀准备因子）。不同信息来源的量表总分及2个因子的内部一致性α系数分别为0.60~0.70,0.70~0.77,0.17~0.62。自杀组2个信息来源一致性组内相关系数（ICC）总分为0.36,2个因子为0.36、0.28;未遂组3个信息来源的ICC总分0.36,2个因子0.36、0.21。自杀组的量表总分及2个因子分均高于未遂组[（15.0±2.5）vs.（11.7±2.3）,（9.4±1.8）vs.（7.2±1.9）,（5.7±1.4）vs.（4.5±0.8）,均P〈0.05]。未遂组服用农药者的量表总分和自杀环境因子分均低于服用治疗药物者;自杀组服用农药者的量表总分和自杀环境因子分均低于使用刀枪绳者（均P〈0.05）。结论：自杀意图客观强度量表的信效度可以接受,可以用于自杀相关研究。

Objective： To evaluate the validities and reliabilities of Objective Scale for Suicidal Intent （OSSI） that had been adapted for suicide and attempted suicide prevention and research. Methods： One hundred and fiftyone suicides and 120 suicide attempters were recruited. For each identified suicides or suicide attempters, one family member proxy respondent and one associate proxy respondent （friend, neighbor） were interviewed separately by trained interviewers. For suicide attempters, subjects themselves were also interviewed by a third trained interviewer. The OSSI was rated by interviewers based on the detailed interview. Construct validity, internal consistency, inter-respondent （family members, associates, and attempters themselves）, consistency of the scale were rated. Objective suicidal intents were compared between males and females, between suicides and attempters, and a- mong suicides/attempters who committed suicidal behavior with different methods. Results： The 2-factor structure of the OSSI-suicidal environment and suicide preparing, was identified for whoever respondent was interviewed. The Cronbach acoefficients of internal consistency of the total scores of the scale, suicide environment subscale, suicide preparing subscale for different respondents interview were ranged 0. 60 -0. 70, 0. 70 -0. 77, and 0. 17 -0. 62, re- spectively. The intra-class correlation coefficients （ICCs） of the inter-respondent consistency for two respondents of suicides were 0. 36 for total scale and suicide environment subscale, 0. 28 for suicide preparing subscale. The ICCs for 3 respondents of attempted suicides were 0. 36 for total scale and suicide environment subscale, 0. 21 for suicide preparing subscale. In suicides, the scores of total scale, suicide environment subscale, and suicide preparing subscale were （15.0 ±2. 5）, （9.4± 1.8）, and （5.7 ± 1.4）, respectively. All of these scores in suicides were higher than those in suicide attempters, （11.7 ±2. 3）, （7.2 ± 1.9）, and （4. 5 ±0. 8）, respectively （Ps 〈0.05）. Among suicide attempters, those ingesting pesticide had lower total scores of the scale and scores of suicide environment subscale than those taking medications. And those suicide by pesticide had lower total scores of the scale and scores of suicide environment subscale than those suicide by knives, ropes and other violent methods. Conclusions： The validities and reliabilities of the Objective Scale for Suicidal Intent are acceptable. The scale could be used in suicide preven- tion and research in the future.