基于离散选择实验的老年人对人工智能辅助慢性病管理咨询服务的偏好研究

打开文本图片集

【Abstract】 ObjectiveA discrete choice experiment （DCE）was employed to quantitatively analyze the atribute preferences older adultsforAI-assstedchronic disease management consultation services.Methods Sawtooth Stware wasused to construct DCE scenarios.Experimental designs were generated using the D-optimal criterion and the % ChoicEff macro program，resulting in multiple-choice tasks.Each DCE task included two alternative options （Option Aand Option B）andan“opt-out/exit”option tosimulatereal-world decision-making among olderadults.To assess internal consistency，each questionnaire includedone repeated-choice task to evaluaterespondent stability，although thesewere excluded from formal data analysis.scenarios were divided into16 versions，and respondents were randomlyassigned to one these versions.Results“Service cost”was the most critical factor for older adults，with a relative importance 40.0% ，indicating high sensitivity to financial burdens.“Complication warning”ranked second （30.6% ），reflecting concerns about early detection and intervention for health risks.“Service provider” （10.7% ）and“response speed”（8.5 % ）were moderately important， indicating that secondaryconsideration was given to who delivers theserviceand its timeliness.“Medication reminders” （5.2 % ）and“interaction mode”（5.1 % ）were the least influential attributes.ConclusionService cost is the primarydecision-making factor.Ahuman-computercollaboration model ispreferred.Real-time warning functions arehighlyvalued.Moderateresponsespeed is moreacceptable.Significantheterogeneity exists in preferences. se insightsunderscore the need for cost-sensitive，risk-focused，and flexible AI-asisted chronic disease management solutions tailored to the needs older adults.

【Key words】Artificial inteligence; Chronic diseases; Disease management； Discrete choice experiment

随着人工智能（artificial intelligence，AI）技术在医疗领域的飞速发展，其在疾病诊断、监测和治疗等核心环节发挥着日益关键的作用，同时也在风险因素识别、健康资源分配和精准医疗干预等多方面发挥着越来越重要的作用[1，2]。（剩余15697字）