利用优劣尺度法确定陈述性偏好研究的属性:以抗抑郁药选择偏好为例  

Using best-worst scaling for attribute selection in stated preference research:a case study on antidepressant preferences

作　　者：陶颖 刘世蒙[1,2] 任延峰 谢飞 陈英耀[1,2] TAO Ying;LIU Shimeng;REN Yanfeng;XIE Fei;CHEN Yingyao(School of Public Health,Fudan University,Shanghai 200032,China;NHC Key Laboratory of Health Technology Assessment(Fudan University),Shanghai 200032,China;Shanghai Mental Health Center,Shanghai Jiao Tong University School of Medicine,Shanghai 200030,China)

机构地区：[1]复旦大学公共卫生学院,上海200032 [2]国家卫生健康委员会卫生技术评估重点实验室(复旦大学),上海200032 [3]上海交通大学医学院附属精神卫生中心,上海200030

出　　处：《健康发展与政策研究》2024年第6期482-491,共10页Health Development and Policy Research

基　　金：国家自然科学基金面上项目(72074047);国家自然科学基金青年项目(72404059)。

摘　　要：目的探索将优劣尺度法应用于陈述性偏好研究的属性开发过程,为属性选择提供方法学参考。方法采用平衡不完全区组设计生成优劣尺度法(best-worst scaling object case,BWS-1)问卷,以有抑郁倾向的社会公众和确诊抑郁症患者为研究对象,通过线上调查获取研究参与者对抗抑郁药物的选择偏好,采用计数法和条件logit模型进行数据分析,并分析偏好异质性。结果计数法和模型法的BWS-1结果具有较高的一致性。肝脏或肾脏损伤(重要性排序均为第1位)、头痛或头晕(重要性排序分别为第2位和第3位)、疾病复发率(重要性排序分别为第4位和第2位)在患者组和抑郁倾向组的重要性排名均靠前,月自付费用在2组中均存在较大偏好异质性。基于定性研究和BWS-1定量结果,最终纳入后续陈述性偏好研究中的属性包括出现肝脏或肾脏损伤的风险、出现头痛或头晕的风险、出现胃肠道不良反应的风险、睡眠障碍、缓解率和月自付费用,其中多数是在BWS-1结果中重要性排名靠前的属性。结论BWS-1可为陈述性偏好研究中属性确定提供有效的定量证据,但仍需结合研究问题、决策环境、利益相关方意见等,通过定性研究补充并论证BWS-1未考虑的属性,以提高研究的严谨性。Objective To explore the use of Best-Worst Scaling(BWS-1,object case)in the attribute selection process of stated preference research,providing methodological guidance for researchers.Methods A BWS-1 questionnaire was created using a balanced incomplete block design.The study surveyed depressed and depression-prone individuals online to assess their preferences for antidepressant selection.Data were analyzed using counting methods and a conditional logit model.The preference heterogeneity was also examined.Results The BWS-1 results obtained from the counting method and modelling approach showed high consistency.Liver or kidney injury(both ranked first in importance),headache or dizziness(ranked second and third,respectively)and relapse rates(ranked fourth and second,respectively)were highly prioritized by both depressed patients and depression-prone respondents.Monthly out-of-pocket costs exhibited substantial preference heterogeneity in both groups.Based on qualitative research and BWS-1 quantitative findings,six attributes were ultimately included in the subsequent stated preference study:the risk of liver or kidney injury,the risk of headache or dizziness,the risk of gastrointestinal adverse effects,sleep disturbances,remission rate,and monthly out-of-pocket costs.Most of these attributes were highly ranked in the BWS-1 results.Conclusions BWS-1 can provide valid and reliable evidence for attribute selection in stated preference research.However,it should be complemented by qualitative research that considers research questions,decision-making contexts,and stakeholder opinions to address attributes not covered by BWS-1,thereby enhancing the study’s rigor.

关 键 词：抑郁症患者 陈述性偏好 优劣尺度法 属性选择 抗抑郁药物 

分 类 号：R195[医药卫生—卫生统计学]