Inform disease prevalence, healthcare utilization, and treatment patterns
EHRs under-represent people who cannot readily access or afford healthcare
Track outcomes for many diseases
Algorithms and machine learning technology contain in-built bias of the system and the inherent biases of developers
Influence public health policy, disease and hospitalization prediction models
EHRs are less often used in rural & impoverished areas where healthcare systems are less centralized and less funded
Guide payers’ recommendations on optimizing disease prevention and treatment plans
Skewed output may affect payers’ decision-making about the medical services and drugs that are necessary to the Black community
Identify potential patients for clinical trials or advanced therapies
Algorithms used by trial investigators to find candidates can exclude Black patients due to negative descriptors used in EHRs