精神病前驱期问卷16项版本评估求助者精神病风险的效度和信度

Validity and reliability of the 16-Item Version of the Prodromal Questionnaire in screening psychosis risk of help-seekers

目的：测试精神病前驱期问卷16项版本（PQ-16）中文版识别求助者精神病风险时的效度和信度.方法：从精神科门诊选取101名求助者完成PQ-16调查,分别对条目分和痛苦分2种评价模式进行内容效度和内部一致性信度检验.选用精神病风险综合征结构式访谈（SIPS）作为效标,检验区分效度和效标效度.结果：SIPS结果阳性与阴性间的PQ-16条目分和痛苦分的差异均有统计学意义（均P＜0.001）.PQ-16各条目分与总分的Spearman相关系数在0.29 ～0.64之间（均P＜0.01）,与痛苦分的相关系数在0.27 ～0.68之间（均P＜0.01）.PQ-16条目分与痛苦分呈正相关（r=0.80,P＜0.001）,PQ-16条目分与痛苦分均与SIPS得分呈正相关（r =0.39 ～0.80,P＜0.001）.PQ-16条目分为7分时可以得到最大的敏感度（66％）和特异度（86％）之和,阳性预测值为54％;而痛苦分为8分时可以获得最佳的敏感度（78％）和特异度（91％）,阳性预测值为68％.PQ-16的条目分和痛苦分均与SIPS诊断相关（AUC=0.81、0.91,均P＜0.001）.PQ-16条目分的Cronbach α系数为0.75,折半信度为0.76;痛苦分的α系数为0.86,折半信度为0.87.结论：精神病前驱期问卷16项版本（PQ-16）中文版评估求助者精神病风险有良好的效度和信度,采用痛苦分的评价模式识别能力更强.

Objective：To examine the validity and reliability of the different evaluated models in the 16-Item Version of the Prodromal Questionnaire （PQ-16） for identifying psychosis risk in the help-seekers.Methods：The PQ-16 was administrated in 101 help-seekers from psychiatry outpatient to analyze the content validity and the internal consistency reliability.The Structured Interview for Prodromal Syndromes （SIPS） was used as criterion to test the discriminant validity and the criterion validity.Results：There were significant difference between the two groups identified as positive or negative diagnosis by the SIPS both in Item Score model （P ＜0.001） and Distress Score model （P ＜0.001）.There were significant correlations between item scores and total scores of PQ-16 both in Item Score model （r =0.29-0.64,Ps ＜0.01） and Distress Score model （r =0.27-0.68,Ps ＜0.01）.The Item Scores statistically positively correlated with the Distress Scores （r =0.80,P ＜ 0.001）,both of which positively correlated with SIPS and it＇ s subscales （r =0.39-0.80,Ps ＜ 0.001）.A cut-off Item Score of 7 and Distress Score of 8 were used to differentiate between those with a diagnosis on the SIPS versus those with no SIPS diagnoses.The cut-off value in Item Score yielded 66％ sensitivity,86％ specificity and a positive predictive value （PPV） of 54％,while 78％ sensitivity,91％ specificity and a PPV of 68％ in Distress Score method.The area under the ROC curve （AUC） both in Item Score （AUC =0.81,P ＜ 0.001） and Distress Score model （AUC =0.91,P ＜ 0.001） significantly predicted the SIPS diagnosis.The Cronbach α was 0.75 and the split-half coefficient was 0.76 by the Item Scores; and the Cronbach α was 0.86 and the split-half coefficient was 0.87 by the Distress Scores.Conclusion：The PQ-16 shows good reliability and validity,and the Distress Score evaluation model may be better than the Item Score in identifying the psychosis risk.