Managing the health risks of member populations proactively is and should be a top priority for all at-risk organizations. Their ability to do so effectively is largely determined by the analytics underpinning and the accuracy of their risk adjustment and stratification models. What innovative risk stratification teams and data scientists are also discovering is that the data sources feeding those analytics are almost as important, if not more so, than the analytics themselves.
Healthcare organizations may be unaware that standard risk adjustment and stratification models are not a one-size-fits-all models. Depending on their goals and objectives, they may find that most standard models provide limited insight. The addition of data from non-traditional sources, plus sophisticated analytics that extract predictive intelligence from the data, can substantially improve the accuracy, effectiveness and value of their risk adjustment and stratification initiatives, and in turn, their efforts to improve health outcomes and reduce costs.
Standard risk adjustment models, like the CMS-HCC Risk Adjustment Model, are primarily used for financial purposes, like health plan reimbursement. These models seek to calculate the expected future average cost of a patient population based primarily on historical claims data and limited demographics. Models based solely on chronic condition or high cost methodologies, for example, fail to account for a wide range of specific drivers of future risks, making them insufficient, on their own, to provide reliable and accurate intelligence for population management decisions.
Fortunately, the solution is contained in a single, simple tactical adjustment—implementing the use of traditional risk adjustment models in combination with advanced predictive analytics. The fusion and analysis of financial, socioeconomic and other non-medical data with clinical-based data enables transparency and insight into the health risks of individuals and populations, now and in the future. The true power of predictive analytics is derived from the synergistic effect created by the combination of multiple data types culled from various sources.
A key alternative source of information that deserves emphasis is socioeconomic data, such as those sourced from public records. When viewed from various angles, public records often contain hidden bits of information that can be pieced together to reveal strong indicators of potential health risks, which are undetectable through standard risk adjustment techniques. For example, credit score dips and bankruptcy filings may indicate a loss of employment or other socioeconomic struggles, which can lead to depression, substance abuse or other mental health issues. Divorces, income drops and certain types of address changes are all strong indicators of major life changes, which can also be predictors of future risk.
The point is that real events in peoples’ lives can have a significant impact on their mental and physical health. More information about individuals’ lives provides more clarity about their health and better opportunities to prevent or fill gaps in their health care. And the predictive power derived from this information source can significantly increase the accuracy with which populations are stratified and lined up for care and disease management programs.