Dive Brief:
- A study published in the Journal of Medical Internet Research highlights promising results from an online risk model that utilized EHRs in order to predict patient needs.
- The model, which was formed using information from almost 1.3 million patients through Maine's health information exchange, successfully predicted patients' healthcare needs a full six months out.
- According to iHealthBeat, the model also allowed accountable care organizations (ACOs) and population health managers to conduct real-time surveillance on patients' health services use.
Dive Insight:
The model worked by examining patients' healthcare utilization over a 12-month period before forming a retrospective resource use scale and a predictive decision tree. This was then used to predict healthcare use for 1.4 million patients over six months, between July 1, 2013 and December 31, 2013.
"Results demonstrated the strong correlation between its care resource utilization and our risk scores, supporting the effectiveness of our model," wrote the study authors, who predicted the model "will enable more effective care management strategies driving improved patient outcomes."
The business benefits of constructing a model that can accurately predict patients' prospective healthcare needs based on past EHR data are legion. If such a predictive model can be emulated in other states and localities, hospitals in the region can more precisely budget for their services and diminish the risk of being caught off-guard with an unexpectedly high tab for rendered services.
It might be more difficult to expand such a predictive model to a federal scale, however, considering the widespread geographic variations in healthcare needs and service costs.
