Application of Machine Learning Techniques to Examine Social Service Needs Among Hispanic Family Caregivers of Persons with Dementia.
Application of Machine Learning Techniques to Examine Social Service Needs Among Hispanic Family Caregivers of Persons with Dementia.
复制标题
DOI：
10.3233/shti220776
复制
发表时间：
2022-06-29
影响因子：
--
通讯作者：
Luchsinger, Jose A
中科院分区：
文献类型：
--
作者：
Yoon, Sunmoo;Mendes, Alexandra;Burgio, Louis;Mittelman, Mary;Dunner, Ilana;Levine, Jed A;Hoyos, Carolina;Tipiani, Dante;Ramirez, Mildred;Teresi, Jeanne A;Luchsinger, Jose A
我们应用机器学习算法来研究人口统计学与社会工作服务结果之间的关系，这些社会工作服务由纽约市招募的痴呆症患者的西班牙裔家庭护理人员使用。社会工作服务需求主要集中在获得医疗保健系统的工具性支持，而不是其他具体服务（例如，住房或食品方案）或解决收入相对较高的照顾者的心理需求。机器学习方法的一项发现是，在接受医疗相关社会工作服务的人中，拥有高家庭朋友支持（>4）的频繁用户（≥ 10次）比没有这种支持的频繁用户更有可能解决他们的问题（准确率：81.9%，AUC：0.82，F-测量：0.86 J 48）。尽管一半的参与者多次接受社会工作服务，但除非他们经常寻求社会工作服务（十次以上），否则照顾者的需求仍然得不到满足。
We applied machine learning algorithms to examine the relationship between demographics and outcomes of the social work services used by Hispanic family caregivers of persons with dementia recruited for a clinical trial in New York City. The social work service needs were largely concentrated on instrumental support to gain access to the healthcare system rather than other concrete services (e.g., housing or food programs) or to address psychological needs among the caregivers with relatively higher income. A finding from the machine learning approach was that among those who receive medical-related social work services, frequent users (≥ 10 times) with high family friend support(>4) were more likely than frequent users without such support to have their issues resolved (Accuracy: 81.9%, AUC: 0.82, F-measure: 0.86 by J48). Even though half of the participants received social work services multiple times, the needs of the caregivers remained unmet unless they sought social work services frequently (more than ten times).