应用累积比数Logit模型对PANSS评分表衍生调查问卷的初步分析

Using the cumulative odds Logit model to explore the application of the questionnaire derived from PANSS

目前临床上应用最多的抗精神分裂症药物阿立哌唑、奥氮平以及利培酮尽管具有较好的疗效,但并不能完全治愈精神分裂症,患者仍需长期服药,因此需要家属对出院后患者的疾病状态更好地了解和掌握。临床上用于评价患者疾病状态的PANSS评分表专业性强,普适性较差,作者在之前的研究中,由PANSS总表开发出了可用于家属评分的调查问卷。本研究将问卷结果通过累积比数Logit概率模型与PANSS评分相对应的5种不同的疾病状态进行了相关性分析,最终模型结果显示问卷评分在目前临床应用最多的阿立哌唑、奥氮平以及利培酮3种药物中对这5种疾病状态均具有较好的预测性能,提示问卷具有较为广泛的临床应用范围。

Currently, aripiprazole, olanzapine and risperidone are three anti-psychiatry agents commonly used in the treatment of schizophrenia. Although the efficacy of these drugs is good, schizophrenia cannot be completely cured yet. Patients need long-term medication. The family members of patients may play a key role to understand the disease status of patients after patient discharge from hospital. PANSS is a commonly used scale in the clinic to evaluate the disease status and drug effects of anti-psychiatry agents. It was professionally written, and is not user friendly to amateurs. In the previous study, we developed a questionnaire for patient＇s family members to monitor the disease status. In this study, we explored the correlations between the results of questionnaire and 5 kinds of disease state corresponding to different PANSS score interval using the cumulative odds Logit model. The final results show that the model had relatively good prediction ability for aripiprazole, olanzapine and risperidone, suggesting that the questionnaire has an extensive prospect of clinical applications.