Nirali Shah, Alexis Castellanos, Yen T Chen, John D Piette, Amy Bucher, Susan L Murphy
{"title":"A Scoping Review on AI-Supported Interventions for Non-Pharmacological Management of Chronic Rheumatic Diseases.","authors":"Nirali Shah, Alexis Castellanos, Yen T Chen, John D Piette, Amy Bucher, Susan L Murphy","doi":"10.1002/acr.25612","DOIUrl":null,"url":null,"abstract":"<p><p>This review summarizes AI-supported non-pharmacological interventions for adults with chronic rheumatic diseases, detailing their components, purpose, and current evidence base. We searched Embase, PubMed, Cochrane, and Scopus databases for studies describing AI-supported interventions for adults with chronic rheumatic diseases. Eligible interventions targeted clinical outcomes (pain, function, disability, fatigue), psychological measures (depression, anxiety), or behavioral outcomes (physical activity, nutrition). All publication types (journal articles, conference abstracts, protocols) published in English language until January 19, 2025, were considered, and interventions of any duration, frequency, country of origin, or setting (inpatient, outpatient, community, and home setting) were included. Two reviewers independently screened studies and one extracted data on study characteristics, intervention components, AI methodologies, and outcomes. Fifteen AI-supported interventions were identified, primarily targeting osteoarthritis (OA) (73%) and focusing on education and exercise advice (67%). The most common AI tool was rule-based expert systems (40%), followed by natural language processing systems (33%) and machine learning algorithms (27%). The interventions ranged from 3 weeks to 12 months, while sample sizes ranged from 7 to 427 participants reflecting huge variability across studies. Most interventions demonstrated high usability, engagement, and adherence. Improvements in exercise compliance, physical activity, and symptoms such as pain and physical function were reported, though effects varied across studies and were sometimes not sustained long-term. AI-supported interventions show promise in promoting education, exercise, and behavioral guidance for adults with chronic rheumatic diseases. There is evidence for high usability and engagement but the clinical impact on long-term symptom management is uncertain.</p>","PeriodicalId":8406,"journal":{"name":"Arthritis Care & Research","volume":" ","pages":""},"PeriodicalIF":3.3000,"publicationDate":"2025-07-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Arthritis Care & Research","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1002/acr.25612","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"RHEUMATOLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
This review summarizes AI-supported non-pharmacological interventions for adults with chronic rheumatic diseases, detailing their components, purpose, and current evidence base. We searched Embase, PubMed, Cochrane, and Scopus databases for studies describing AI-supported interventions for adults with chronic rheumatic diseases. Eligible interventions targeted clinical outcomes (pain, function, disability, fatigue), psychological measures (depression, anxiety), or behavioral outcomes (physical activity, nutrition). All publication types (journal articles, conference abstracts, protocols) published in English language until January 19, 2025, were considered, and interventions of any duration, frequency, country of origin, or setting (inpatient, outpatient, community, and home setting) were included. Two reviewers independently screened studies and one extracted data on study characteristics, intervention components, AI methodologies, and outcomes. Fifteen AI-supported interventions were identified, primarily targeting osteoarthritis (OA) (73%) and focusing on education and exercise advice (67%). The most common AI tool was rule-based expert systems (40%), followed by natural language processing systems (33%) and machine learning algorithms (27%). The interventions ranged from 3 weeks to 12 months, while sample sizes ranged from 7 to 427 participants reflecting huge variability across studies. Most interventions demonstrated high usability, engagement, and adherence. Improvements in exercise compliance, physical activity, and symptoms such as pain and physical function were reported, though effects varied across studies and were sometimes not sustained long-term. AI-supported interventions show promise in promoting education, exercise, and behavioral guidance for adults with chronic rheumatic diseases. There is evidence for high usability and engagement but the clinical impact on long-term symptom management is uncertain.
期刊介绍：
Arthritis Care & Research, an official journal of the American College of Rheumatology and the Association of Rheumatology Health Professionals (a division of the College), is a peer-reviewed publication that publishes original research, review articles, and editorials that promote excellence in the clinical practice of rheumatology. Relevant to the care of individuals with rheumatic diseases, major topics are evidence-based practice studies, clinical problems, practice guidelines, educational, social, and public health issues, health economics, health care policy, and future trends in rheumatology practice.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
