Research uncovers how provider offices discriminate against patients seeking primary care

New research published in Economics Letters by the Tulane School of Public Health and Tropical Medicine examines how differences in questions asked and information provided by a physician’s office contributes to inequalities in whether or not an individual is accepted as a new patient.

The research was conducted by Dr. Janna Wisniewski and Dr. Brigham Walker, both research assistant professors in the Department of Health Policy and Management, as well as by researchers at Portland State University and Oakland University.

While discrimination based on sex, race, and ethnicity is outlawed, the American Medical Association’s Code of Ethics states that “physicians are not ethically required to accept all prospective patients.” Physicians have the discretion to reject patients based on financial factors or concerns that the patient will be difficult to treat. Previous research has demonstrated that physicians display the same level of implicit racial bias as the general population, indicating that the tastes and preferences of physicians may lead to discrimination.

According to Wisniewski and Walker, the mechanisms that lead to appointment refusals, other than financial factors, have not been well documented. Their team conducted an audit of primary care offices and found that Black and Hispanic patients were more likely to be asked about their insurance.

Data was collected between 2013–2016 in 50 states and Washington, DC, yielding 11,030 completed calls to practicing primary care physicians. They found that the rates and stated reasons for denial of appointment offers differ substantially across patient groups.

Overall, Medicaid patients were offered 27.6% fewer appointments than privately insured patients. Hispanic (−6.4 percentage points, comparable to −9.8%) and Black women (−3.9 percentage points, comparable to −6.0%) were at the most significant disadvantage compared to White women. Male groups received fewer appointments (between 3.0 percentage points / 4.6% and 3.6 percentage points / 5.5% less) than White women. No differences by sex or race were found among patients with Medicare and Medicaid, but among self-pay patients, Hispanic women and men were offered fewer appointments than White women. All groups – Black men and women, Hispanic men and women, and White men – were less likely to have their insurance accepted compared with White women, and all groups but Black men were more likely to be told that the practice was not accepting new patients compared with White women.

“What is striking about these results is that provider offices deny some patients on the stated basis of their insurance, despite all patients having the same insurance on average,” said Walker.

Medicaid status alone was sufficient for many physicians to reject patients outright. However, their research suggests that physicians may be screening non-Medicaid patients using observable characteristics, using the excuse that the practice is “not taking new patients” to justify some prospective patients’ refusal.

The paper helps to disentangle an elusive distinction between taste-based discrimination and statistical discrimination. Both are problematic, but the discriminations are motivated differently. Practices exhibiting taste-based discrimination make choices based on animus to their economic detriment. Practices exhibiting statistical-based discrimination make assumptions to drive toward their economic goals.

Both may be at play in this study. However, given that significant disparities occurred among self-pay patients whose ability to pay is most uncertain compared to those covered by insurance suggests a significant statistical discrimination role. This distinction’s relevance is that these tendencies may be best disrupted in different ways, some of which the researchers plan to explore in future, related research.