Successful application of big data analytics to personalized medicine often involves probabilistic risk assessment models to apply individual patient characteristics to treatment decisions. However, these models have been criticized for insufficient “contextual specificity” and for their vulnerability to incorrect interpretations of the predictive model results. Contextual specificity reflects the rich diversity of individuals who interact with the healthcare system as characterized by demographics, cultures, lifestyles, preferences, socio-economic status, and genetic endowments. Moreover, the reality is that many datasets available to healthcare systems, researchers, payers, and regulatory decision makers do not fully reflect all groups or experiences within healthcare systems. In addition, studies using healthcare data do not always interpret race and ethnicity within the context of the often-racialized environment in which healthcare utilization occurs. What are the tradeoffs associated with including race as a covariate in prediction models? Considering the measures of race available in healthcare system data, are these data ‘fit for purpose’ when the goal is to guide healthcare delivery in racially, ethnically, and culturally diverse patient populations? How can HEOR inform and guide the use of healthcare system data while advancing health equity goals? This session will discuss how we can appropriately utilize healthcare data for probabilistic risk assessment as well as illustrate HEOR’s evolving and increasing role in addressing these issues.

*Speakers to be added as confirmed!