Abstract:
|
Research from the early 1980s produced sophisticated tools, such as the DxCG relative risk score (RRS) for predicting a year's health care costs (and other outcomes) for individuals and groups of people from each person's age, sex and medical problems (diagnoses) coded on electronic claims (or encounter) records during a year. For some years now, MassHealth (our State's Medicaid program) has used this RRS in calculating payments to accountable care and managed care organizations (MCOs) to ensure that a plan gets the right amount of money to care for the needs of each person who signs up - more for sicker people, less for healthy ones. My team worked with the State to add social determinants of health data (SDH) and to address other problems not previously accounted for in its current payment formulas. I will discuss several practical issues, such as how to adjust for factors that we cannot explicitly include as predictors, such as high levels of long-term supportive services (LTSS) use and how to identify "issues of concern" after fixing whatever problems we can through risk adjustment.
|