Artificial intelligence for patient scheduling in the real-world health care setting: A metanarrative review

Objectives

The application of artificial intelligence (AI) and machine learning (ML) to scheduling in medical practices has considerable implications for most specialties. However, the landscape of AI and ML use in scheduling optimization is unclear. We aimed to systematically summarize up-to-date evidence about application of AI and ML models for scheduling optimization in clinical settings.

Methods

We systematically searched multiple databases from inception through August 2020 to identify studies that described real-world application of AI and ML in health care scheduling and reported outcomes. Eligible studies included those conducted in any health care setting using ML or predictive modeling through AI to optimize patient scheduling processes in real-time, real-world settings. Outcomes of interest included assessing impact on stakeholders (i.e., providers, patients, health systems), including impact on workload, burden, burnout, cost, utilization, patient and provider satisfaction, waste reduction, and quality. Data were extracted and reviewed in duplicates, independently and blindly by two reviewers. The results were synthesized and summarized using a metanarrative approach.

Results

The initial search strategy yielded 3,415 citations, of which 11 eligible studies were included. Outcome measures for studies on missed appointments covered patient double-booking volume, missed appointments, service use, and missed appointment risk. Resource allocation outcomes assessed wait time, disease-type matching performance, schedule efficiency revenue, and new patient volume wait time. Other outcomes included visit requests, examination length prediction, and surgical case time.

Conclusions

Available evidence shows heterogeneity in the stages of AI and ML development as they apply to patient scheduling. AI and ML applications can be used to decrease the burden on provider time, increase patient satisfaction, and ultimately provide more patient-directed health care and efficiency for medical practices. These findings help identify additional opportunities in which AI platforms can be developed to optimize patient scheduling.

Public Interest Summary

Artificial Intelligence (AI) and machine learning (ML) can help many aspects of health care. Patient scheduling has significant implications for the cost benefits of improved technology. The longstanding use of technology in medicine serves as a strong foundation for future AI applications. Here, we present an up-to-date review of the current use of AI and ML for schedule optimization in the health care clinic setting. Current evidence shows a wide variety of stages in the development, function, and application of AI and ML in patient scheduling. Given the current gaps of knowledge, future studies should address feasibility, effectiveness, generalizability, and risk of AI bias in patient scheduling.