Using data from HealthInfoNet, Maine's health information exchange (HIE), researchers have created an electronic medical record-based online risk model to predict the healthcare resources patients would need six months out.
Such predictions can help health organizations better plan care, and are vital in the shifting U.S. healthcare payment system, according to an article published in the Journal of Medical Internet Research.
The researchers are touting success with the model after testing it at both the individual patient level and population level. The analysis involved aggregated data from more than 1 million patients about their care from the preceding 12 months.
The researchers found that a small proportion of patients consumed a relatively large amount of healthcare resources, but for a variety of reasons, prompting the authors to recommend focusing resources on them, but treating them separately.
Predictive analytics were integrated into the Maine State HIE system with an online population risk surveillance dashboard to allow real-time surveillance by accountable care organization field staff and population health managers.
University of Chicago researchers have developed a computational model comparing resources allocated to study a specific disease with its relative burden on society.
Meanwhile, Humana is using predictive analytics to identify gaps in care, and Aetna, along with its plans to merge with Humana, is focused on replacing its fee-for-service reimbursement model with a value-based model.
To learn more:
- read the article