Skip to main content

Department ofManagement & Entrepreneurship

News

Michael Santoro

Michael Santoro

Michael Santoro: AACSB 2021 Innovations That Inspire

Congratulations to Michael Santoro and colleagues for their work "Reducing Racial Disparities in Medical Appointment Scheduling" named by AACSB as one of their 2021 Innovations That Inspire.

Congratulations to Michael Santoro and colleagues for their work "Reducing Racial Disparities in Medical Appointment Scheduling" named by AACSB as one of their 2021 Innovations That Inspire.

Reducing Racial Disparities in Medical Appointment Scheduling

By employing artificial intelligence techniques in medical outpatient scheduling, researchers help maximize clinic efficiency while minimizing racial disparities in patient wait times.

Call to Action

An estimated 23 percent of appointments at outpatient medical clinics result in no-shows. Patient no-shows negatively affect the efficiency of outpatient clinics by introducing disruptions and reducing the provider’s productivity. A common way to counteract no-shows is to overbook appointments—to schedule more than one patient into the same appointment slot. It is well known that, in order to maximize clinic efficiency, clinics regularly overbook patients with the highest risk of no-show. However, those patients tend to disproportionately belong to racial groups associated with lower socioeconomic status.

Some reasons behind the correlation between race and no-show are inferior access to transportation, inability to afford child care, and difficulty taking time off of work to go to the appointment. Thus, maximizing clinic efficiency results in disproportionately scheduling patients belonging to the most vulnerable populations into the least desirable appointment slots. Consequently, those patients tend to experience longer wait times at the clinic when compared to other patients.

The goal of our project was to develop appointment-scheduling methodologies and software tools that maximize clinic efficiency while minimizing racial disparity. The project was done in collaboration with the Black Women’s Health Imperative and was carried out by a diverse and interdisciplinary research group, which includes researchers from Santa Clara University and Virginia Commonwealth University, with collective expertise in operations management, information systems, and business ethics.

Description

The project has four main innovative aspects. First, we developed an analytical model to prove that racial disparity arises when patients are scheduled using state-of-the-art appointment scheduling algorithms. The goal of this model was to mathematically prove that patients belonging to racial groups associated with a higher no-show risk tend to experience longer wait times at the clinic.

Second, we validated the analytical findings using data from a large specialty clinic whose Black patients have a higher risk of no-show than non-Black patients. In that dataset, state-of-the-art appointment scheduling algorithms result in a racial disparity of 33 percent; that is, Black patients wait 33 percent longer for their appointment than non-Black patients.

Third, we developed and implemented two types of methodologies to schedule patients efficiently while minimizing racial disparity: a race-aware approach and a race-unaware approach. The race-aware approach consists of using patients’ racial information to optimally schedule appointments in such a way that racial disparity is avoided. While the explicit use of race to schedule appointments is legal under current regulations, it may not be suitable for all clinics. Thus, a race-unaware approach was also developed: this methodology attempts to minimize racial disparity without explicitly using the patient’s race; instead, it uses proxies such as the patient’s no-show risk.

Fourth, we developed an open-source software package and posted it on GitHub. This tool allows clinics to choose a strategy for scheduling a set of patients given their individual no-show risk.

Impact

Compared to the state-of-the-art algorithm, the methodologies we developed in this project resulted in both high clinic efficiency and a small disparity between the Black and non-Black patients’ waiting times. In particular, the race-aware model results in no significant racial disparity and in an efficiency that is only 1.83 percent lower than that of the state-of-the-art algorithm. Compared to the state-of-the-art algorithm, the race-unaware model reduces the racial disparity from 33.12 percent to 7.70 percent, while reducing the clinic efficiency by only 3.56 percent.

Our work is the first project that studies and addresses racial disparity in medical appointment scheduling. The methodologies we developed as part of this study simultaneously achieve clinic efficiency and racial fairness. Our software tool can be readily used by practitioners to schedule appointments at their outpatient clinics. In July 2020, E. Tendayi Achiume, the United Nations special rapporteur on racism and xenophobia, cited our work in a report presented to the U.N. Human Rights Council. Our project was one of only six other studies on racial disparities in healthcare settings.

Reference Links:

MGMT, LSB, Home,