Model Predicts Healthcare Utilization With EMR Information

Researchers in China and the United States have developed a predictive model capable of forecasting which patients will account for either small or large proportions of healthcare resource utilization in the next 6 months.

Researchers in China and the United States have developed a predictive model capable of forecasting which patients will account for either small or large proportions of healthcare resource utilization in the next 6 months.

In a study published in the Journal of Medical Internet Research, the researchers used data from Maine’s health information exchange and constructed a retrospective cohort of more than 1.2 million patients using the preceding 12-month electronic medical record data.

They then scored patient healthcare resource utilization for the next 6-month period and developed a predictive model that was later integrated into the Maine health information exchange to predict the next 6-month risk of resource utilization for more than 1.3 million patients.

“The trend of increasing spending on health care demands focused attention, which should include analyzing health care resource utilization drivers and predicting future care resource utilization,” the authors wrote. “An effective prediction of future resource utilization can help improve care resource allocation and care delivery, supporting the transition from a volume-based incentive system to a value-based system.”

The researchers used mean outpatient, emergency department, and inpatient days to reflect the trend of healthcare resource utilization rather than cost.

Patients with high risks of resource utilization were mostly elderly, had inpatient or emergency department visits, or had chronic diseases. In particular, high-risk patients were likely to have hypertension, diabetes, heart disease, or asthma or chronic obstructive pulmonary disease. Patients without any chronic diseases has the lowest risk of utilization and the lowest next 6-month mean resource utilization.

The purpose of the predictive model is to assist providers so they can apply appropriate care management and optimize resource allocation for the care of high-risk patients.

“Knowing in advance the projected care service usage associated with chronic medical conditions can help providers make more informed decisions on the allocation of care management services with the objective of decreasing unnecessary utilization associated with treating chronically ill patients,” the authors concluded.