A Machine Learning Model to Maximize Efficiency and Face Time in Ambulatory Clinics

Area of Science:

• Health Informatics
• Machine Learning in Healthcare
• Pediatric Urology Operations

Background:

• Ambulatory clinic appointments are often scheduled in fixed time slots, leading to inefficient use of physician time.
• Advanced analytical methods are underutilized in addressing healthcare operational challenges.
• Developing predictive models for clinic visit duration is crucial for optimizing healthcare delivery.

Purpose of the Study:

• To develop and validate a machine learning model for predicting the time pediatric urologists need for patient visits.
• To enhance the efficiency of pediatric urology clinic scheduling.
• To reduce patient wait times and improve overall clinic flow.

Main Methods:

• Prospective data collection from January to April 2018, including patient demographics and visit-specific covariates.
• Development of separate random forest models for new and return visits using a 4:1 train/test split.
• Simulation of 1,000 clinic days to compare wait times between fixed-increment scheduling and the machine learning model.

Main Results:

• The machine learning model achieved accurate predictions of doctor time: 3.6 minutes for new patients and 5.0 minutes for returning patients.
• Factors associated with longer visits included new patient status, testing, older age, and diagnoses like voiding dysfunction.
• Simulations showed machine learning reduced patient wait times by 24% to 54% compared to fixed-time scheduling.

Conclusions:

• Machine learning models can accurately predict pediatric urologist appointment durations.
• Implementing these predictive models can lead to more efficient clinic scheduling.
• Optimized scheduling has the potential to significantly minimize patient wait times and enhance family satisfaction.