Machine Learning and Medication Adherence: Scoping Review (Preprint)

Abstract

BACKGROUND This is the first scoping review broadly focused on machine learning and medication adherence. OBJECTIVE To categorize and summarize literature focused on using machine learning for medication compliance activities. METHODS PubMed, Scopus, ACM Digital library, IEEE, and Web of Science were searched to find works that meet the inclusion criteria. Study information of interest was systematically charted before inclusion in the final draft. Studies were placed into natural categories for additional analysis dependent upon the combination of medication adherence activities carried out. The protocol for this scoping review was created using the PRISMA-ScR guidelines. RESULTS Publications focused on predicting medication adherence have uncovered strong predictors that were significant across multiple studies. Studies that used machine learning to monitor medication compliance are generally still in early developmental stages and used a variety of sensor data to detect medication administration. Systems that combined medication monitoring with intervention were mostly concerned with detecting medication administration and only a few compared their system against more traditional approaches. CONCLUSIONS In general, this topic currently has relatively few publications but has been generating more interest over the last few years. Although important features for predicting adherence have been identified more work needs to be done to understand the complex interplay between these features. Systems used to monitor medication compliance also require further testing in more realistic environments and user acceptability evaluations. When interventions are attempted the effectiveness of the system should be evaluated against current systems used to encourage medication compliance. CLINICALTRIAL NONE