Second, population predictive models exist, are accurate and can predict ER visits and inpatient days for the next 12 months, along with costs and gaps in care. They use medical and prescription claims data. Since most people under age 65 are covered by commercial plans, the Health Insurance Portability and Accountability Act “healthcare operations” exception precludes the need for patient consent. For those 65 and older, valid predictive models also exist; it is simply a matter of getting the federal government to do something.
Third, real-time predictions are nonsense. Seventy to 80% of health-plan costs are driven by the 25% of people who have one or more chronic diseases and either receive half the care they need, or much more than they need (Medicare in particular). It took those people a long time to get as sick as they are; it will take a long time to help them get healthy. The issues are identification, stratification and engagement. Making the information available at the point of care is essential to avoid fragmentation of care.
The authors of the article need to get their heads out of the clouds and get into clinical practice.