Many health care systems around the world have adopted broad-based population health strategies to improve patients’ experience of care, advance health, and lower per capita costs―the "Triple Aim" espoused by the Institute for Healthcare Improvement. But according to former Commonwealth Fund Harkness Fellows Geraint Lewis and Rhema Vaithianathan and colleagues, achieving these goals requires more targeted approaches. These experts argue that predictive modeling techniques are needed to identify and prioritize care for patient populations at risk of experiencing a "triple fail" event—one that is harmful, costly, and likely to result in poor patient satisfaction.
What the Study Found
Health care organizations should begin the predictive modeling process by analyzing medical and pharmacy claims, electronic health record information, and other administrative data to estimate the risk of individuals experiencing an event such as an unplanned hospital readmission within 30 days, or lower back surgery for patients not provided decision support services.
Once opportunities for improvement are identified, health care organizations prioritize care for patients who are most likely to respond to interventions or for those receiving suboptimal care. Providers then track and analyze the impact of their interventions to identify even more specific characteristics of patients likely to respond to them.
A stratified approach to the achieving the Triple Aim has many advantages over traditional population health strategies, but it requires careful planning, monitoring, and adaption for success. Pilot projects and more immediate access to Medicare data may also improve the use of predictive modeling techniques.