Racial bias detected in cancer treatment approvals

Prior authorization; the process by which a health insurance company denies or approves coverage for a health care service before the service is performed; became standard practice beginning with Medicare and Medicaid legislation in the 1960s.

Although research has found disparities in prior authorization for cancer patients based on race, little is known to date about the role of prior authorization in increasing or decreasing these disparities.

To learn more, Benjamin Ukert, PhD, assistant professor of health policy and management in the Texas A&M University College of Public Health, and a Penn State colleague conducted a retrospective study of data provided by a major national commercial insurance provider on 18,041 patients diagnosed with cancer between January 1, 2017, and April 1, 2020.

Data on prior authorization from providers and insurers are difficult to access and analyze, but this research could provide valuable insights into equity in the specialty care prior authorization process for patients, health care providers and plan administrators, policymakers, and employers.

Benjamin Ukert, PhD, Associate Professor, Health Policy and Management, School of Public Health, Texas A&M University

For the study, Ukert described the racial and ethnic makeup of the data used in terms of prior authorization process outcomes for self-insured and fully insured adults diagnosed with the 13 most common cancers other than basal cell carcinomas, which typically do not require prior authorization. Subjects had at least two claims for Evaluation and Management office visits with a cancer diagnosis or a cancer diagnosis during an emergency department or hospital stay during the study period.

For prior authorization data, Ukert looked at the length of days from cancer diagnosis to prior authorization, the decision to deny or approve the service, and whether the denial resulted from medical necessity.

Independent variables were self-reported race/ethnicity, provided by employers and electronic medical records, and extracted from sociodemographic data on covered individuals that were available from the insurer. Racial categories were non-Hispanic white, non-Hispanic Asian, non-Hispanic black, and Hispanic (Hispanic-white or Hispanic-black).

For covariates, Ukert used a broad set of sociodemographic control variables identified from medical claims and the American Community Survey. Others included sociodemographic information, including information on health insurance coverage and duration of health plan enrollment before cancer diagnosis. After measuring the degree of any comorbidities during the six months before cancer diagnosis, Ukert merged block group characteristics on household income and education level from the 2017 five-year American Community Survey. She then used linear regression models to assess whether disparities by race or ethnicity emerged in prior authorization process outcomes.

The sample was 85 percent white, 3 percent Asian, 10 percent black, and 1 percent Hispanic; 64 percent were women; and the average age was 53. The average prior authorization denial rate was 10 percent, and the denial rate specifically for medical necessity was 5 percent. Those who identified as Hispanic had the highest prior authorization denial rate, 12 percent, while those who identified as black had the lowest prior authorization denial rate, 8 percent.

“In summary, we found no racial or ethnic disparities in prior authorization outcomes for individuals who identify as Black and Hispanic, compared to whites,” Ukert said. “Additionally, Asian patients had higher rates of prior authorization approvals compared to white patients.”