Navigate the healthcare industry
Healthcare Brew covers pharmaceutical developments, health startups, the latest tech, and how it impacts hospitals and providers to keep administrators and providers informed.
Doctors might see upward of 20 patients on an average day, which means scheduling needs to take top priority.
To minimize the time patients spend in the waiting room while not overbooking doctors, some clinics use software with machine learning algorithms. An ideal system should maximize scheduling efficiency while saving the clinic money on staffing. However, clinics might not take into account whether or not the tech is biased.
A scheduling algorithm used by a large specialty clinic forced Black patients to wait up to 30% longer to see their doctors, according to a 2021 study published in the academic journal Manufacturing and Service Operations Management.
The study, conducted between 2019 and 2021, looked at a clinic with a patient population of about 60% white patients and 30% Black patients, with the remaining 10% of patients falling into other ethnic groups.
The scheduling algorithm identified which patients had the greatest risk of not showing up to appointments, and those patients were placed in overbooked slots. The clinic’s Black patients were given higher no-show probability than non-Black patients.
“This no-show probability is a little bit like a credit score,” Michele Samorani, an associate professor of information systems and analytics at Santa Clara University and one of the study’s authors, told Healthcare Brew. “If you want to maximize efficiency, then you’re going to overbook the people with the highest no-show risk.”
Factors like a patient’s ability to take time off work and afford transportation or childcare go into deciding their no-show risk level. The algorithm determined that Black patients were more likely to have these risk factors than non-Black patients. “From a legal standpoint, you can use any characteristic to predict no-shows, even what we would think of as protected characteristics,” Samorani said. “While the use of protected characteristics is regulated in other fields, like loans or hiring, [...] in healthcare it’s quite unregulated.”
Patients scheduled in or right after overbooked slots tend to have a worse care experience because they have to wait longer, the study found.
“The disparity in waiting time between racial groups is especially unjust because of the evidence that people of color generally have inferior access to healthcare, receive poorer quality care, and experience worse healthcare outcomes,” the study’s authors wrote.
Hope is not lost, though. There are ways to fix the disparity.
“By far, the most effective approach is actually what we call the ‘race-aware’ [approach], which takes race into account when scheduling the appointment. Because by trying to equalize the waiting time of everybody, we find that that achieves [an] optimal level of fairness, and surprisingly, with no decrease in efficiency,” Samorani said. “There are simple ways to solve it.”
Has your clinic struggled with racial bias in its algorithms (scheduling or otherwise)? Email Maia Anderson at [email protected].