基于BP神经网络模型的河南省严重精神障碍患者服药依从性影响因素分析

A study on influencing factors of drug compliance in Henan serious mental disorder patients based on BPNN

目的:探讨BP神经网络模型在河南省严重精神障碍患者服药依从性影响因素分析中的应用。方法:采用多阶段分层随机抽样的方法抽取患者,利用自编调查问卷收集患者和监护人基线资料及患者服药依从性的情况,采用BP神经网络模型及logistic回归模型分析服药依从性的影响因素。结果:共收集1872份有效问卷,服药依从患者961例。BP神经网络和logistic回归模型的正确预测率分别为72.1%和71.4%,ROC曲线下面积(95%CI)分别为0.772(0.752~0.791)和0.769(0.750~0.788),二者相比,差异无统计学意义(P=0.687)。根据BP神经网络模型中自变量的重要性标准化后的结果,影响服药依从性前3位的因素是家庭收入水平、所患精神障碍类型和发病后监护人采取措施。结论:河南省严重精神障碍患者服药依从性较差。运用BP神经网络构建河南省严重精神障碍患者服药依从性模型可行,模型拟合性能良好。

Aim:To understand the present situation of drug compliance in Henan serious mental disorder patients,and to study its influencing factors based on BP neural network(BPNN).Methods:Using stratified random sampling method,the baseline data and the patients′drug compliance were collected by self-designed questionnaires.BPNN model and logistic regression model were applied to analyze the influencing factors of drug compliance.Results:A total of 1872 valid questionnaires were collected,and there were 961 patients with drug compliance.The correct forecast rate of trained BPNN model and logistic regression model was 72.1%and 71.4%,and AUC(95%CI)were 0.772(0.752-0.791)and 0.769(0.750-0.788)(P=0.687).After standardization,the first three influencing factors of drug compliance according to BPNN model were the family income level,disease category and the guardian to take measures after onset.Conclusions:The present situation of drug compliance in Henan serious mental disorder patients was relatively poor.It is feasible to construct a prediction model of drug compliance in Henan serious mental disorder patients based on BPNN,and the model fits in good condition.