Racial-ethnic, insurance-mediated bias common when recommending diabetes devices


Ebekozien reports being a member of the Medtronic Diabetes Health Equity Advisory Board and receiving compensation through T1D Exchange. Odugbesan reports no relevant financial disclosures.

We were unable to process your request. Please try again later. If you continue to have this issue please contact [email protected]

More than 60% of diabetes providers display implicit insurance-mediated bias, and more than one-third have racial-ethnic bias when recommending diabetes technology, according to study data published in Diabetes Technology & Therapeutics.

Ori Odugbesan

“Provider bias exists as it relates to insurance and race/ethnicity,” Ori Odugbesan, MD, MPHassociate director of health equity and quality improvement at T1D Exchange in Boston, told Healio. “We’d like every patient to receive standard care, which includes offering appropriate technology to all patients regardless of their race, regardless of their insurance status, family structure or educational level.”

Continuous Glucose Monitor_381066634
Source: Adobe Stock

Odugbesan and colleagues conducted a multicenter study of 109 providers from seven diabetes clinics in the T1D Exchange Quality Improvement Network as part of an ongoing health equity project (82% women; mean age, 40.7 years). The Diabetes Provider Implicit Bias Tool was used to evaluate bias, with modifications made for adult providers and to analyze implicit racial-ethnic bias. The tool, developed by co-author Ananta Addala, DO, MPHincluded a hypothetical clinical vignette in which providers were randomly assigned questions varying by insurance status and patient names, which were used as a proxy to racial identity. Participants ranked race/ethnicity, family income, age, HbA1c, type of insurance coverage, self -monitoring blood glucose and patient or family preference in order of consideration for recommending diabetes technology. Providers were also asked whether they were aware of their own explicit bias.

Implicit racial-ethnic-mediated bias was identified if a provider offered more technology to a hypothetical patient with a non-Hispanic white name than to patients with common Hispanic or common Black names, or if they ranked race and ethnicity as one of the top two factors for recommending technology. Insurance-mediated implicit bias was identified if patients with private insurance were offered more technology than those with public insurance in the vignettes or if insurance was ranked as one of the top two factors for recommending technology.

One-third of providers display racial-ethnic bias

Of the cohort, 56% were physicians, 69% were pediatric diabetes providers and 82% were women. Racial-ethnic mediated bias was observed in 34% of the cohort. Providers who said they were able to recognize their own bias were more likely to have more racial-ethnic bias than those who were not aware of their own bias (adjusted OR = 4.66; 95% CI, 1.6-17.09; P = .009).

Osagie Ebekozien

“There’s a lot more awareness about some of these racial-ethnic biases, and that can explain why a lot of people are recognizing they have their own biases,” Osagie Ebekozien, MD, MPH, CPHQ, population health researcher and executive vice president and chief medical officer at T1D Exchange, told Healio. “A lot of providers also have to go through training on bias as part of their diversity, equity and inclusion series. There’s value in that, but one of the things we’re pointing out in the study is if you have only interventions that tackle awareness building, it doesn’t address some of the more implicit or unconscious bias that have been baked in for much longer.”

Insurance-mediated bias common

Bias against public insurance was observed in 61% of the cohort. Providers who did not have insurance-mediated bias were younger (mean age, 38.3 years vs. 42.3 years) and had fewer years in practice (5.3 years vs. 9.3 years; P = .006) than those who had bias.

Pediatric providers ranked technology preference as the most important factor for recommending continuous glucose monitoring and insulin pump therapy, whereas race and ethnicity was ranked as least important. Insurance coverage was the third-most important factor for recommending CGM and fourth-most important for recommending insulin pump therapy. Adult providers had similar results, except insurance coverage was ranked as third-most important for recommending both CGM and insulin pump therapy.

Odugbesan and Ebekozien said more studies are needed to evaluate the effectiveness of interventions that target implicit bias as well as examining the root of bias.

“When did this bias start? Is it at medical school or is it while they are practicing?” Odugbesan said. “This would really help to think about how we can address biases. Would it be helpful to bake some of this training into the medical curriculum? Would it be helpful to bake it into continuing medical education? We also need to measure and try to quantify contributors to inequities, because we know this is just one of the contributors.”

“There’s also a big role in how we empower patients to be advocates for themselves,” Ebekozien added. “Patient-centered care and patients are really being the drivers for improved outcomes for using technology. We’re hoping there will be more research on those aspects.”

For more information:

Osagie Ebekozien, MD, MPH, CPHQ, can be reached at [email protected]

Ori Odugbesan, MD, MPHcan be reached at [email protected]

Leave a Comment