Overbooked and Overlooked: Machine Learning and Racial Bias in Medical Appointment Scheduling
41 Pages Posted: 23 Oct 2019 Last revised: 20 Jun 2020
Date Written: March 24, 2020
Problem definition: Machine learning is often employed in appointment scheduling to identify the patients with the greatest no-show risk, so as to schedule them into or right after overbooked slots. That scheduling decision maximizes the clinic performance, as measured by a weighted sum of all patients’ waiting time and the provider’s overtime and idle time. However, if the patients with the greatest no-show risk belong to the same demographic group, then that demographic group will be scheduled into or right after overbooked slots disproportionately to the general population.
Academic/Practical Relevance: That is problematic because patients scheduled in those slots tend to have a worse service experience than the other patients, as measured by the time they spend in the waiting room. Waiting time is undesirable because it increases patients’ frustration and dissatisfaction.
Methodology: Motivated by the real-world case of a large specialty clinic whose black patients have a higher no-show probability than non-black patients, we demonstrate that the state-of-the-art scheduling method that combines machine learning with scheduling optimization causes the black patients in our dataset to wait about 30% longer than non-black patients. To eliminate that disparity, we develop both “race-aware” and “race-unaware” solution methods: the former consider race explicitly when scheduling patients, whereas the latter do not.
Results: Our results suggest that the race-aware methodology is the only one capable of achieving both goals of eliminating racial disparity and obtaining the same schedule quality as that obtained by the state-of-the-art scheduling method. In contrast, the race-unaware methodologies fail to obtain both efficiency and fairness. We validate our findings both on simulated data and real-world data.
Managerial Implications: Our work allows healthcare schedulers to reap the benefit of machine learning without generating undesirable disparities.
Keywords: Appointment Scheduling, Machine Learning, Algorithmic Bias, Socio-economic Bias, Racial Bias
Suggested Citation: Suggested Citation