No abstract

No abstract

No abstract

No abstract

Decades of research have demonstrated that diabetes affects racial and ethnic minority and low-income adult populations in the U.S. disproportionately, with relatively intractable patterns seen in these populations' higher risk of diabetes and rates of diabetes complications and mortality (1). With a health care shift toward greater emphasis on population health outcomes and value-based care, social determinants of health (SDOH) have risen to the forefront as essential intervention targets to achieve health equity (2-4). Most recently, the COVID-19 pandemic has highlighted unequal vulnerabilities borne by racial and ethnic minority groups and by disadvantaged communities. In the wake of concurrent pandemic and racial injustice events in the U.S.,

Machine learning is used increasingly in clinical care to improve diagnosis, treatment selection, and health system efficiency. Because machine-learning models learn from historically collected data, populations that have experienced human and structural biases in the past—called protected groups—are vulnerable to harm by incorrect predictions or withholding of resources. This article describes how model design, biases in data, and the interactions of model predictions with clinicians and patients may exacerbate health care disparities. Rather than simply guarding against these harms passively, machine-learning systems should be used proactively to advance health equity. For that goal to be achieved, principles of distributive justice must be incorporated into model design, deployment, and evaluation. The article describes several technical implementations of distributive justice—specifically those that ensure equality in patient outcomes, performance, and resource allocation—and guides clinicians as to when they should prioritize each principle. Machine learning is providing increasingly sophisticated decision support and population-level monitoring, and it should encode principles of justice to ensure that models benefit all patients.

Several themes emerged in the statement, including a need for basic science, population-based, translational and health services studies to explore underlying mechanisms contributing to endocrine health disparities. Compared to non-Hispanic whites, non-Hispanic blacks have worse outcomes and higher mortality from certain disorders despite having a lower (e.g. macrovascular complications of diabetes mellitus and osteoporotic fractures) or similar (e.g. thyroid cancer) incidence of these disorders. Obesity is an important contributor to diabetes risk in minority populations and to sex disparities in thyroid cancer, suggesting that population interventions targeting weight loss may favorably impact a number of endocrine disorders. There are important implications regarding the definition of obesity in different race/ethnic groups, including potential underestimation of disease risk in Asian-Americans and overestimation in non-Hispanic black women. Ethnic-specific cut-points for central obesity should be determined so that clinicians can adequately assess metabolic risk. There is little evidence that genetic differences contribute significantly to race/ethnic disparities in the endocrine disorders examined. Multilevel interventions have reduced disparities in diabetes care, and these successes can be modeled to design similar interventions for other endocrine diseases.

No abstract