BackHome
Refresh: OffView Personal Schedule
Individual Submission Summary
Share...

Direct link:

Leveraging Diagnosis and Biometric Data to Uncover Disparities in Obesity Diagnosis

Sat, August 9, 2:00 to 3:30pm, East Tower, Hyatt Regency Chicago, Floor: Concourse Level/Bronze, Randolph 1A

Abstract

Despite extensive efforts to standardize definitions of obesity, clinical practices of diagnosing obesity vary widely. This study examined (1) discrepancies between biometric body mass index (BMI) measures of obesity and documented diagnoses of obesity in patient electronic health records (EHRs) and (2) how these discrepancies vary by patient gender and race and ethnicity from an intersectional lens. Methods: Observational study of 383,380 participants in the National Institutes of Health All of Us Research Program dataset. Results: Over half (60%) of participants with a BMI indicating obesity had no clinical diagnosis of obesity in their EHRs. Adjusting for BMI, comorbidities, and other covariates, women’s adjusted odds of diagnosis were far higher than men’s (95% confidence interval 1.66-1.75). However, the gender gap between women’s and men’s likelihood of diagnosis varied widely across racial groups. Overall, Non-Hispanic (NH) Black women and Hispanic women were the most likely to be diagnosed and NH-Asian men were the least likely to be diagnosed. Conclusion: Men, and particularly NH-Asian men, may be at heightened risk of underdiagnosis of obesity. Women, and especially Hispanic and NH-Black women, may be at heightened risk of unanticipated harms of obesity diagnosis, including stigma and competing demand other health concerns. Leveraging diagnosis and biometric data from this unique public domain dataset from the All of Us project, this study revealed pervasive disparities in diagnostic attribution by gender, race, and ethnicity.

Authors