County-level variation in imaging utilization impacts practicing radiologists

Imaging utilization in the U.S. varies more at the county level than at the state level, according to a study published in the Journal of the American College of Radiology. In addition, researchers found poor ratings in economic measures, such as income and insurance coverage, independently correlated with higher imaging utilization—the opposite of what they expected.

According to co-author Andrew Rosenkrantz, MD, affiliate research fellow at the Neiman Health Policy Institute, these factors could affect radiologists, especially practicing physicians.  

“Radiologists need to keep these findings in mind if they’re taking on risk,” he said. “If they’re going to be accountable for a given population and controlling spending, they need to realize there’s enormous variation across regions driven by basic socioeconomic factors in the covered population,” he said.

Rosenkrantz and his colleagues used national databases to measure county-level Medicare imaging utilization, economic status and preformed multivariate regressions to tease out connections. They had originally anticipated greater imaging use in higher-income areas, postulating more financial resources would translate to greater access to health providers.

“Not only was this not the case, but an opposite effect was in fact observed on multivariate analysis: Higher uninsured and unemployed rates at the county level were both independently associated with higher Medicare imaging utilization and costs,” the authors wrote.

While this particular study didn’t delve into the reasons why, Rosenkrantz theorized it might have something to do with poor coordination of care.

“In certain counties where the population has low levels of insurance and care is fragmented, perhaps there’s more unnecessary imaging or repeat imaging, versus areas with more robust coverage,” he said.

These disparities have real world implications for radiologists and imaging practices. When care providers enter risk-bearing contracts and take responsibility for a given population, these county-by-county variations can have an outsize effect on predicted spending levels and resource utilization. This pattern may intensify as the net is tightened—even down to mere zip codes.

“The smaller the region, the more variation there could be,” said Rosenkrantz. “Other studies have shown wide variation at levels smaller than counties. On the other hand, some of this may get evened out at the statewide level.”

These findings should serve as a call to action for both researchers and practicing radiologists, said Rosenkrantz, but more transparency in data sharing is needed from both payors and providers.

“We linked data across databases, but we don’t have the fine details of the exact imaging exams that we’re done, we don’t know if they were done in hospitals, physician’s offices, or outpatient facilities,” he said. “Insurers should make coverage information available and all data should be accessible, including feedback reports to practices.”

Given these findings, further identification of potential drivers of imaging utilization and costs in the Medicare population are warranted, and practicing radiologists should have that data available to aid decision making, said Rosenkrantz.