首页|人工智能电话随访在社区高血压患者健康管理中的应用现况及影响因素分析

人工智能电话随访在社区高血压患者健康管理中的应用现况及影响因素分析

扫码查看
目的:对人工智能(artificial intelligence,AI)电话随访的接听、信息采集、一致性情况等进行分析,以评价AI技术在社区慢性病随访管理中的应用价值.方法:于 2020 年 9 月 1 日—9 月 30 日采用多阶段整群随机抽样方法选取上海市静安区彭浦新村街道社区卫生服务中心慢性病管理对象中需随访的高血压患者作为调查对象,采用AI随访系统对高血压患者进行电话随访,对采集到的内容进行有效信息提取,自动生成随访结果.结果:本次研究所覆盖的高血压患者为 4 070 例,其中 3 887 例进行了应答,电话接通率为95.50%,有效应答人数为 3 529 例.二分类及多分类条目的信息采集率及一致率与询问次序呈低度负相关,定量类条目的信息采集率及一致率与询问次序呈低度正相关.AI电话随访系统共采集了52 852 条信息,语音信息采集完整率为71.31%,其中 49 677 条信息与人工复核结果一致,语音识别信息一致率为93.99%.单因素分析结果显示,数据类型、年龄及AI电话随访拨打时间是对语音识别信息一致率差异具有统计学意义的影响因素(P<0.05).结论:建议通过加载方言语音包、设置更为智能化的交互问询、更加标准化的答案设置及个性化的电话随访拨打时间等提升AI电话随访的信息采集率及一致率.
Analysis of the application status and influencing factors of AI telephone follow-up in the health management of hypertensive patients in the community
Objective:To analyze the artificial intelligence(AI)telephone follow-up answering,information collection,and consistency to evaluate the application value of AI technology in community chronic disease follow-up management.Methods:From September 1 to 30,2020,the multi-stage cluster random sampling method was used to select hypertensive patients who needed follow-up among the chronic disease management subjects in Pengpuxincun Community Health Service Center as the survey objects,and the AI follow-up system was used to conduct telephone follow-up of hypertensive patients,and the contents collected were effectively extracted,and automatically generated follow-up results.Results:A total of 4 070 hypertensive patients were covered in this study,of which 3 887(95.50%)responded and 3 529 responded effectively.The information collection rate and agreement rate of dichotomous data and multi-categorical data were negatively correlated with the order of inquiry,while the quantitative data were positively correlated.A total of 52 852 pieces(71.31%)of the information were collected by the AI telephone follow-up system,of which 49 677 pieces(93.99%)were consistent with the results of manual review.The results of univariate analysis showed that data type,age and dialing time of AI telephone follow-up were statistically significant influencing factors for the difference in the consistency rate of speech recognition information(P<0.05).Conclusion:It is suggested to improve the information collection rate and consistency rate of AI telephone follow-up by loading dialect voice packs,setting up more intelligent interactive inquiry,more standardized answer settings,and personalized telephone follow-up dialing time.

hypertensive patientAI telephone follow-upanalysis of influencing factor

张琳、彭德荣、罗元欣、金长琴、程旻娜、王思源、蒋云

展开 >

上海市静安区彭浦新村街道社区卫生服务中心健康管理部,上海 200435

上海市静安区彭浦新村街道社区卫生服务中心,上海 200435

上海市疾病预防控制中心慢性病与伤害防治所 健康管理科,上海 200336

高血压患者 人工智能电话随访 影响因素分析

上海市卫生健康委员会科研项目

20214Y0488

2024

上海医药
上海医药行业协会

上海医药

影响因子:0.781
ISSN:1006-1533
年,卷(期):2024.45(16)
  • 14