基于KAP实证分析的化疗患者用药依从性调查及列线图预测模型的建立

Investigation into Medication Compliance of Chemotherapy Patients and Establishment of Nomogram Prediction Model on Empirical Analysis of KAP

目的:了解肿瘤患者化疗的用药依从性现状,探讨影响患者用药依从性的危险因素并建立风险模型。方法:以2019年68月在中国科学院合肥肿瘤医院住院化疗的肿瘤患者为研究对象。采用自行设计的一般资料调查表、化疗知识-态度-行为(KAP)调查问卷及Morisky服药依从性量表等对患者的基本信息、患者对抗肿瘤药合理使用的KAP、用药依从性及其影响因素进行调查。采用Logistic回归分析影响患者服药依从性的危险因素,然后将筛选出的独立危险因素使用R软件建立列线图预测模型,并对模型的预测性及准确度进行验证。结果:Logistic回归分析结果显示,化疗次数>2[OR=2.540,95%CI(1.423,4.531)]、尚不能完全承受药费[OR=1.613,95%CI(1.125,2.313)]、用药种数>2[OR=1.836,95%CI(1.089,3.095)]、每日用药次数>3[OR=1.810,95%CI(1.169,2.804)]、KAP知识得分≤10[OR=2.090,95%CI(1.204,3.629)]、KAP态度得分≤16[OR=1.937,95%CI(1.151,3.261)]及KAP行为得分≤5[OR=1.785,95%CI(1.059,3.007)]为化疗患者用药依从性差的独立危险因素,具有统计学意义(P<0.05)。Bootstrap法验证化疗患者用药依从性差的列线图模型的预测效能,一致性指数(C-index)为0.760[95%CI(0.733,0.787)],且校正曲线趋近于理想曲线,说明本研究列线图模型具有良好的精准度和区分度。结论:对肿瘤患者及时考虑化疗次数、能否承受药费、用药种数、每日用药次数、KAP知识、KAP态度、KAP行为等因素综合评估化疗患者用药依从性差的发生率,能够提高对化疗患者用药依从性差的诊断效能,具有较高的临床应用价值,值得进一步推广使用。

Objective:To understand the current status of medication compliance of cancer inpatient,discuss the risk factors affecting patients’medication compliance and establish a risk model.Methods:The patients who were hospitalized in the oncology department of our hospital from June to August 2019 were taken as the research object.The self-designed general data questionnaire,questionnaire on chemotherapy knowledge-attitude-practice(KAP)and Morisky medication compliance scale were used to investigate the basic information of patients,patients’KAP about rational use anti-tumor drug,medication compliance and influencing factors.According to the classification into good compliance group and poor compliance group,Logistic regression analysis was used to analyze the risk factors of patients’medication compliance.Futhermore,the independent risk factors screened out were used to establish a nomogram prediction model using R software,and the predictability and accuracy of the model were verified.Results:Logistic regression analysis showed that times of chemotherapy>2(OR=2.540,95%CI 1.423 to 4.531),in ability to afford the drug expenses(OR=1.613,95%CI 1.125 to 2.313),kinds of drugs>2(OR=1.836,95%CI 1.089 to 3.095),times of daily drugs>3(OR=1.810,95%CI 1.169 to 2.804),KAP knowledge scores≤10(OR=2.090,95%CI 1.204 to 3.629),KAP attitude scores≤16(OR=1.937,95%CI 1.151 to 3.261)and KAP behavior scores≤5(OR=1.785,95%CI 1.059 to 3.007)were independent risk factors for noncompliance of drug use in tumor patients with statistically significant differences(P<0.05).Bootstrap internal verification method according to the collinear nomogram model for concurrent drug non-compliance in cancer patients was used,the C-index was 0.760(95%CI 0.733 to 0.787),and the calibration curve was close to the ideal curve,indicating that the nomogram model in this study had good accuracy and differentiation.Conclusion:Timely consideration of chemotherapy times,affording the expenses for medicine,drug kinds,medication daily times,and KAP knowledge,attitude and behavior in cancer patients can be used to comprehensively evaluate the incidence of poor compliance and improve the diagnosis of poor compliance in cancer patients,which shows high clinical application values,and is worthy of further promotion.