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

Introduction: Ambulatory appointments are typically scheduled in fixed increments, resulting in suboptimal time utilization. Advanced analytics are rarely applied to address operational challenges in health care. We sought to develop a machine learning model that predicts the time pediatric urologists require to create a more efficient clinic schedule.
Methods: We prospectively collected data from January to April 2018. Variables included demographics and visit level covariates. The primary outcome was defined as in-room doctor time spent. Univariate analysis was performed. Data were split into train/test in a 4:1 ratio. Separate models using random forest were created for new and return visits. Two out-of-sample clinic days were used to compare the patient wait time between fixed-time visits and machine learning model. Patient punctuality simulation was performed 1,000 times for each day.
Results: A total of 256 visits (113 new/143 return) were included. Mean age at visit was 6.47 years. In univariate analysis, longer visits were significantly associated with new patients (p <0.01), testing (p <0.01), older patients and diagnoses like voiding dysfunction and neurogenic bladder. Conversely, morning clinic, previous urological surgery (p <0.01), recent postoperation (p <0.01) and diagnoses like penile complaints and hydrocele were associated with shorter visits. On average, our machine learning model predicted doctor time accurately to 3.6 (new patients) and 5.0 minutes (returning patients). In 1,000 simulated days with random patient punctuality machine learning reduced the wait time by 24% to 54%.
Conclusions: Pediatric urologists' clinic time can be accurately predicted with machine learning models. This insight can be incorporated into a robust scheduling model to minimize patient wait time, increase clinical efficiency and likely improve family satisfaction.