In Journal of clinical epidemiology ; h5-index 60.0
OBJECTIVE : To describe the degree of automation in just-in-time, adaptive interventions (JITAIs) assessed in randomized controlled trials (RCTs) in any medical speciality, and to assess the completeness of intervention reporting.
STUDY DESIGN AND SETTING : Systematic scoping review. We searched PubMed, PsycINFO, and Web of Science, from 1 January 2019 to 2 March 2021, for reports of RCTs assessing JITAIs. We assessed whether study reports included the minimum information required to replicate the intervention based on JITAI frameworks. We described JITAIs according to their automation level using an established framework (partially, highly, or fully automated), and care workload distribution (requiring work from patients, healthcare professionals [HCPs], both, or neither).
RESULTS : We included 88 JITAIs (63%, n=56 supported chronic illness management, 12%, n=11 supported health behavior change). Overall, 77% (n=68) of JITAIs were missing some information required to replicate the intervention (e.g., n=38, 43% inadequately reported the algorithm used to select intervention components). Only fifteen (17%) JITAIs were fully automated and did not require additional work from HCPs nor patients. Of the remaining JITAIs, 36% required work from both patients and HCPs, and 46% required work from either patients or HCPs.
CONCLUSIONS : Most JITAIs are not fully automated and require work from HCPs and patients.
Oikonomidi Theodora, Ravaud Philippe, LeBeau Jonathan, Tran Viet-Thi
2022-Dec-12
artificial intelligence, clinical decision support systems, closed-loop, digital, ecological momentary interventions, just-in-time adaptive interventions
