背景：目前我国老年人口规模庞大，人口老龄化形势日益严峻。相较于单纯的认知障碍或衰弱，认知衰弱会显著增加跌倒、住院、失能失智、死亡等不良健康结局发生风险。加强社区认知衰弱老年人的健康管理，尽早识别认知衰弱迹象并及时采取针对性的干预策略可有效防控失能失智的发生，为促进健康老龄化提供新思路。基于移动健康技术的健康干预和管理能够提高医疗服务的可及性，为老年人提供个性化、灵活的干预选择。鉴于老年人在应用移动健康方面的偏好具有异质性，影响其使用行为，因此有必要全面地了解认知衰弱老年人对于移动健康干预的偏好特征。

目的：（1）识别认知衰弱老年人移动健康干预的偏好及其异质性，为不同偏好类型的认知衰弱老年人移动健康干预策略的构建提供指导。（2）整合认知衰弱老年人干预的高质量证据，结合不同偏好特征，构建有针对性的认知衰弱老年人的移动健康干预策略。

方法：研究分为三部分。第一部分为横断面研究，基于文献回顾、半结构式访谈、专家咨询和重要性排序，设计并实施离散选择实验，便利选取北京市西城区德胜社区卫生服务中心、丰台区方庄社区卫生服务中心、海淀区某干休所社区服务站和朝阳区某干休所社区服务站服务范围内社区的424名认知衰弱老年人作为研究对象，测量其移动健康干预偏好，建立离散选择模型对总体偏好、边际支付意愿、模拟情景预测和偏好异质性进行分析。第二部分为定性用户画像研究，采用目的抽样法并按照最大差异化原则，选取参与第一部分偏好研究的三类不同偏好类型的社区认知衰弱老年人作为研究对象，通过半结构式访谈法提取分析认知衰弱老年人对移动健康技术使用经历和需求特征，构建经济敏感型、效果优先型和情感支持型三类不同移动健康干预偏好的定性用户画像。第三部分为干预策略构建研究，采用文献研究法，系统回顾能够改善认知衰弱老年人认知功能和衰弱状态的高质量证据，结合认知衰弱老年人移动健康干预偏好，对干预要素进行归纳、总结，形成分类别的干预策略。

结果：第一部分研究共确定核心目标、单次干预时长、每月花费、交互模式、远程评估方式和健康教育形式6个属性和对应的16个水平。共收集352份有效的离散选择实验问卷，纳入研究的6个属性的各水平均在统计学意义上有显著性影响（P＜0.05），单次干预时长和核心目标是认知衰弱老年人最为重视的非经济属性，尤其是“单次干预时长-20分钟”（β=1.231，P＜0.001）和“核心目标-缓解负性情绪”（β=1.083，P＜0.001）；而交互模式（β=0.218，P=0.027）相较于其他属性对认知衰弱老年人的影响程度最低。边际支付意愿分析结果显示，影响认知衰弱老年人移动健康干预偏好的最重要非经济因素是单次干预时长，如果将单次干预时长由40分钟精简至20分钟，他们愿意为此支付151.247元（95% CI：0.066，0.138）。情景预测分析结果显示相较于基线水平，如果月均花费分别增至50元/月和100元/月，预测认知衰弱老年人的选择概率将分别下降17.2%和29.8%；而如果将提升身体功能转变为改善认知功能、缓解负性情绪，选择概率将增加18.3%、22.2%。偏好异质性结果显示，基于潜类别Logit模型可将认知衰弱老年人的移动健康干预偏好分为经济敏感型、效果优先型和情感支持型三类。第二部分共访谈25例认知衰弱老年人，根据访谈内容分析并归纳出认知衰弱老年人对移动健康干预偏好的定性用户画像的四个标签维度，分别是整体特征、健康行为要素、技术需求和功能需求。通过以上标签维度对偏好研究结果呈现的三类移动健康干预偏好进行描述，并绘制定性画像剪影和词语图。第三部分将针对轻度认知障碍人群和衰弱人群的干预与管理的高质量证据作为干预策略的证据支撑，分析其在移动健康干预中可实施的转化，针对三类不同移动健康干预偏好特征，最终构建包括干预内容、健康教育和功能评估三个维度的干预策略。

结论：认知衰弱老年人对移动健康干预的偏好受多种因素影响，其中每月花费、核心目标、单次干预时长、交互模式、远程评估方式和健康教育形式均是影响社区认知衰弱老年人移动健康技术干预偏好的关键属性，但影响程度各不一致，按照偏好特征可分为经济敏感型、效果优先型和情感支持型。不同类型的移动健康干预偏好的用户画像特征各异，因此干预策略的设计应兼顾偏好的同质性和异质性，并结合现有高质量证据，构建既符合人群价值观念又遵循循证依据的干预策略，促进健康老龄化管理的推动。

Background: At present, the scale of the older adults population in China is huge, and the situation of population aging is becoming more and more serious. Compared with cognitive impairment or frailty alone, cognitive frailty increases the risk of adverse health outcomes such as falls, hospitalization, disability and dementia, and death. Strengthening the health management of the older adults with cognitive frailty in the community, identifying the signs of cognitive frailty as early as possible and taking targeted intervention strategies in time can provide new ideas for the promotion of healthy aging and the prevention and control of disability and dementia. Health intervention and management based on mobile health technology can improve the accessibility of medical services and provide personalized and flexible intervention options for the older adults. The older adults have heterogeneous preferences for mobile health technology application, which affects the use behavior. Therefore, it is necessary to comprehensively understand the preferences of the older adults with cognitive frailty for mobile health intervention.

Objectives: (1) To identify the preferences and heterogeneity of mobile health intervention for cognitively frail older adults, and to provide guidance for the construction of mobile health intervention strategies for cognitively frail older adults with different preferences. (2) Construct mobile health intervention strategies for cognitively frail older adults by integrating high-quality evidence of intervention for cognitively frail older adults and combining preference characteristics.

Methods: The study was divided into three parts. The first part was a cross-sectional study, which implemented a discrete choice experiment based on literature review, semi-structured interviews, expert consultation and importance ranking. A total of 424 older adults people with cognitive frailty from Desheng Community Health Service Center in Xicheng District, Fangzhuang Community Health Service Center in Fengtai District, a cadre retreat community service station in Haidian District and a cadre retreat community service station in Chaoyang District of Beijing were selected as the research objects by convenience sampling, and their preferences for mobile health intervention were measured. A discrete choice model was established to analyze the aggregate preference, marginal willingness to pay, simulation scenario prediction and preference heterogeneity. The second part was qualitative user portrait study. Purposive sampling method was used and according to the principle of maximum differentiation, three types of community cognitively frail older adults people with different preference types who participated in the former part of the preference study were selected as interview objects. The characteristics of the experience and needs of the cognitively frail older adults people using mobile health technology were extracted and analyzed by semi-structured interview method. To construct qualitative user portraits of three types of mobile health intervention preferences, namely economically sensitive, effect preferred and emotional support. The third part was the study on the construction of intervention strategies. The literature review method was used to systematically review the high-quality evidence that could improve the cognitive function and frailty state of the older adults with cognitive frailty, and the intervention elements were summarized and summarized in combination with the intervention preferences of the older adults with cognitive frailty to form classified intervention strategies.

Results: In the first part of the discrete choice experiment, six attributes and their corresponding 16 levels were determined through the steps of literature review, in-depth interview, expert consultation and importance ranking. A total of 352 valid discrete choice experiment questionnaires were collected. All levels of the six attributes included in the study had statistically significant effects (P<0.05). Core goals and single intervention duration were the most important non-economic attributes of the older adults with cognitive frailty, especially "single intervention duration -20 minutes" (β=1.231,  P<0.001) and "core goal-relieving negative emotions" (β=1.083, P < 0.001); Interaction mode (β=0.218, P=0.027) had the lowest impact on cognitive frailty compared with other attributes. The results of marginal willing-to-pay analysis showed that the most important non-economic factor affecting the preference of mobile health intervention in the older adults with cognitive frailty was the duration of a single intervention. If the duration of a single intervention was reduced from 40 minutes to 20 minutes, they were willing to pay 151.247 yuan (95% CI: 0.066, 0.138). The results of scenario prediction analysis showed that, compared with the baseline level, if the average monthly cost increased to 50 yuan/month and 100 yuan/month, the choice probability of the older adults with cognitive frailty would decrease by 17.2% and 29.8% , respectively. If the improvement of physical function was transformed into the improvement of cognitive function and the relief of negative emotions, the choice probability would increase by 18.3% and 22.2% , respectively. The results of preference heterogeneity showed that based on the latent class Logit model, the mobile health intervention preference of the older adults with knowledge and frailty was divided into three categories: cost-sensitive type, effect preferred type and emotional support type. In the second part of qualitative user portrait study, semi-structured interviews were conducted with 25 cognitive frailty older adults people. According to the interview content, four label dimensions of qualitative user portrait of cognitively frail older adults people's preference for mobile health intervention were analyzed and summarized, which were overall characteristics, health behavior elements, technology needs and functional needs. The above label dimensions were used to describe the three types of mHealth intervention preferences presented by the preference research results, and the qualitative portrait silhouette and word map were drawn. In the third part of the intervention strategy construction study, the high-level evidence of intervention and management for people with mild cognitive impairment and frailty was used as evidence support for intervention strategies, and the transformation that could be implemented in mobile health intervention was analyzed. Combined with the preference characteristics, three types of intervention strategies with different mobile health intervention preferences were finally constructed from the three dimensions of intervention content, health education and functional assessment.

Conclusions: The preference of the older adults with cognitive frailty for mobile health intervention is affected by many factors, among which monthly cost, core goal, duration of single intervention, interaction mode, remote assessment method and health education form are the key attributes that affect the preference of mobile health technology intervention for the older adults with cognitive frailty in the community, but the degree of influence is not consistent. According to the preference characteristics, it can be divided into economic sensitive type, effect preferred type and emotional support type. Different types of mobile health interventions have different characteristics of user portraits. Therefore, the design of intervention strategies should take into account the homogeneity and heterogeneity of preferences, and combine with the existing high-quality evidence to construct intervention strategies that are in line with population values and follow evidence-based evidence.