Ward Catherine L, Flisher Alan J, Zissis Chrisostomos, Muller Martie, Lombard Carl J
J Child Adolesc Ment Health. 2004 Jul 1;16(1):31-7. doi: 10.2989/17280580409486561.
Objective - This study aimed to assess the reliability of an adaptation of the Harvard Trauma Questionnaire for use in epidemiological investigations of adolescents' exposure to violence and related post-traumatic stress disorder (PTSD) symptoms. Method - The exposure items of the Harvard Trauma Questionnaire were adapted for the South African context. The symptoms scale was left intact. Self-report questionnaires were administered on two occasions to 104 students in four high schools in Cape Town, South Africa. Test-retest reliability was assessed using Cohen's kappa (k) and observed agreement, as well as a concordance correlation coefficient for the symptoms scale. Cronbach's alpha was used to assess internal consistency of the symptoms scale. Results - All but one item showed at least fair test-retest reliability. Cronbach's alpha for the symptoms scale was 0.92, and the concordance correlation coefficient between the total symptoms score at time 1 and time 2 was 0.64 (95% CI = 0.51-0.74). Conclusions - Results suggest that the instrument may be sufficiently reliable for use in epidemiological studies of adolescent populations, although this sample is limited in several key respects. Further studies should investigate reliability in broader samples, as well as examining validity.
目的——本研究旨在评估改编后的哈佛创伤问卷在青少年暴力暴露及相关创伤后应激障碍（PTSD）症状流行病学调查中的可靠性。方法——哈佛创伤问卷的暴露项目针对南非背景进行了改编。症状量表保持不变。对南非开普敦四所高中的104名学生进行了两次自填问卷调查。使用科恩kappa系数（κ）和观察一致性评估重测信度，同时使用一致性相关系数评估症状量表的信度。使用克朗巴哈α系数评估症状量表的内部一致性。结果——除一项外，所有项目均显示出至少尚可的重测信度。症状量表的克朗巴哈α系数为0.92，第一次和第二次总症状评分之间的一致性相关系数为0.64（95%CI = 0.51 - 0.74）。结论——结果表明，尽管该样本在几个关键方面存在局限性，但该工具在青少年人群的流行病学研究中可能具有足够的可靠性。进一步的研究应在更广泛的样本中调查信度，并检验效度。
