Machine learning in general practice: scoping review of administrative task support and automation

Area of Science:

• Healthcare Informatics
• Machine Learning Applications
• General Practice Management

Background:

• Artificial intelligence (AI) is increasingly used in general practice for diagnostics and treatment recommendations.
• AI applications for administrative task automation in general practice are currently limited.
• This review summarizes research on machine learning (ML) for administrative tasks in general practice.

Purpose of the Study:

• To conduct a scoping review of machine learning methods applied to general practice administrative tasks.
• To identify research gaps and future directions in this field.

Main Methods:

• Searched healthcare and engineering databases (PubMed, Embase, CINAHL, Cochrane, Scopus, IEEE Xplore) from April to June 2022.
• Screened 1158 records based on eligibility criteria, extracting data on nine attributes.
• Included 12 studies in the final analysis.

Main Results:

• Most studies focused on scheduling tasks using supervised ML with minimal general practitioner (GP) involvement.
• Four studies utilized advanced ML methods, but data varied significantly in setting, type, and availability.
• The research landscape for ML in general practice administration is nascent.

Conclusions:

• There is a significant need and high potential for ML in automating general practice administrative tasks.
• Limited research is attributed to a lack of open-source data and a focus on diagnostic AI.
• Future research should prioritize open-source data, advanced ML techniques, and explicit GP involvement for replicability.