The break in FRAX: Equity concerns in estimating fracture risk in racial and ethnic minorities

Abstract

Although equitable care for racial and ethnic minorities has been a longstanding concern in medicine, Vyas et al. s recent article in New England Journal of Medicine highlighted a subset of clinical algorithms whose use of race may perpetuate inequities in clinical care. This includes the U.S. Fracture Risk Assessment Tool (FRAX) which, uniquely among Western countries, differentiates among four racial/ethnic groups: Caucasian, black, Hispanic, and Asian. Although FRAX has become useful for clinical decision-making, we see serious concerns in continuing to use FRAX to estimate fracture risk and determine osteoporosis treatment when it systematically underestimates the risk in non-white populations. FRAX was developed by the World Health Organization Collaborating Centre for Metabolic Bone Diseases at the University of Sheffield after identifying significant risk factors for fracture through meta-analyses of large observational cohorts. The group identified differences in fracture risk across countries and, within the United States, across racial/ethnic groups that led to the development of separate calculators. In the United States, the black, Hispanic, and Asian calculators include the same variables as the Caucasian calculator but also incorporate a correction factor which results in lower calculated fracture risk among these groups (Figure 1). These correction factors were derived from cohort studies conducted in the 1980s and 1990s, which showed a differential fracture risk by race and ethnicity, but have not been updated since. Previous studies have posited a relationship between race/ethnicity and bone mineral density (BMD), namely that BMD is greater in non-Hispanic black (NHB) men and women and lower in Asian populations compared with non-Hispanic whites (NHW). Such racial and ethnic differences in BMD have implied a biological or genetic relationship that can be proxied by race and ethnicity. However, areal BMD, as measured from twodimensional images using dual X-ray absorptiometry (DXA), does not account for three-dimensional bone size, shape, and body weight. Differences in areal BMD between racial and ethnic groups largely disappear when incorporating these factors. There is also an increasing consensus in medicine and sociology that race/ethnicity is a social construct. As such, race/ethnicity does not represent biological difference, but a complex relationship comprised of socioeconomic, political, geographic, and environmental factors. Many socioeconomic and environmental factors influence BMD and fracture risk. Lower income and educational status have been correlated with decreased BMD, thought to be operationalized through nutritional factors such as calcium and vitamin D status, physical activity, and access to medical care. Finally, recent relatively fast epidemiological shifts in fracture incidence over the last decade, such as decreased fracture risk in the NHW population and increased risk in the Hispanic population, also implicate environmental factors rather than genetic or evolutionary factors.10–12 It is also of concern that FRAX does not account for disparities in medical comorbidities that contribute to DOI: 10.1111/jgs.17316 Journal of the American Geriatrics Society