Examination of private insurance claims data offers new insights

Until recently, almost all research on healthcare spending focused on publicly available Medicare data. Physicians and hospitals never revealed their transaction prices, and insurance companies did not share their claims data. The lack of transparency made it difficult to determine why healthcare costs have increased so much and why the U.S. spends nearly $3 trillion on healthcare without seeing better outcomes than other countries.

A recent report examining employer-sponsored private insurance claims data found that basing healthcare decisions on Medicare data might not be the best practice. The researchers found that the correlation between total spending per Medicare beneficiary and total spending per privately insured beneficiary was 0.14 in 2011, while the correlation for inpatient spending was 0.267.

“What that suggests is policy for Medicare doesn’t necessarily make better policy for the privately insured,” researcher Zack Cooper, PhD, assistant professor of publich health at Yale University, told Health Exec in a phone interview. “That was hugely surprising for us and a hugely important finding. Many of the places that policymakers have looked at as models for the nation [such as] Grand Junction, Colorado, and Lacrosse, Wisconsin, and Rochester, Minnesota, these are areas that are very cheap for Medicare, but they end up being some of the most expensive places in the nation to get private health insurance. I think that’s really, really important as we think about what policy to introduce going forward.”

The researchers were the first team to analyze data from the Healthcare Cost Institute (HCCI), a nonprofit organization established in 2011 to examine the causes of increased health spending in the U.S. Aetna, Humana, Kaiser Permanente and UnitedHealthcare provides HCCI with de-identified claims data that includes a unique provider identifier, a unique patient identifier, the date services were provided, the amount providers charged, providers’ negotiated transaction prices and payments that patients made to providers. As of October 2015, HCCI had commercial claims from more than 50 million people per year from 2007 through 2014.

In a 2014 HCCI report, which found healthcare spending per privately insured person increased 3.4 percent from the previous year, the organization had claims from more than 25 percent of people younger than 65 who had employer-sponsored insurance.

Cooper and fellow researchers Stuart Craig of the University of Pennsylvania, Martin Gaynor of Carnegie Mellon University and John Van Reenen of the London School of Economics worked on this project for four years. They had claims information on 27.6 percent of adults from 18 to 64 years old who had private employer-sponsored insurance in the U.S. between 2007 and 2011. The data included 92 billion health insurance claims from 88 million people covered by Aetna, Humana and UnitedHealthcare and accounted for approximately 5 percent of healthcare spending in the U.S. The Commonwealth Fund, National Institute for Health Care Management Foundation and Economic and Social Research Council funded the study.

The researchers evaluated prices in 306 hospital referral regions, which the Dartmouth Atlas of Health Care first identified for Medicare beneficiaries and has become the most commonly accepted geographic boundary in healthcare, according to Cooper.

They found that healthcare spending for the privately insured varied widely from a low of $1,707.39 per beneficiary in Honolulu, Hawaii, to $5,515.90 per beneficiary in Napa, Calif. And in many cases, there was little correlation in spending in regions when comparing Medicare and private insurance costs. For instance, they noted that Grand Junction, Colo., had the third lowest spending per Medicare beneficiary among the 306 regions in 2011 but also had that ninth highest average inpatient prices in the nation and the 43rd highest spending per person with private insurance.

Further, they found that transaction prices had a large role in the variation of inpatient spending for the privately insured, whereas prices had almost no role in spending variation among Medicare beneficiaries.

“Spending can be a function of price or quantity – how much care is provided in the area or what the price is in that area,” Cooper said. “For Medicare, prices don’t vary because Medicare sets the price for all healthcare services. The only thing that makes one region more expensive than another is it provides more stuff. Our goal was to see whether that applied to the private side. It turns out that private prices really dictate in large part whether an area is expensive, high spending or low spending.”

In fact, the researchers noted that hospitals’ and providers’ transaction prices varied across geographic areas. Hospitals that are for-profit, have more medical technologies and have a low proportion of Medicare patients are more likely to have higher prices. After adjusting for numerous factors, prices for hospitals that had monopolies in their regions were 15.3 percent higher than prices for areas with four or more hospitals.

The researchers plan on publishing more studies, including on healthcare spending growth over time, healthcare providers' prices growth over time, the causal relationship between market structure and hospital prices and the relationship between changes in Medicare reimbursements and hospitals' negotiated transaction prices.

Cooper noted that Medicare covers only 16 percent of people in the U.S. and accounts for 20 percent of healthcare spending. Thus, having access to private insurance claims data allows researchers to have a better understanding of healthcare spending and price variation.

“A lot of people in policy have suspected that prices play a role, but we hadn’t really had the evidence,” Cooper said. “It’s firmed up the importance of thinking about things like antitrust enforcement. That is, stopping hospitals from merging and how important that is to reigning in the price of healthcare. I think hospitals on the other hand have basically said historically that the reason that prices are so high is because Medicare underpays them or there are differences in quality that account for why one might be so expensive. I think our research highlights that that’s not the case. It really makes this industry much more transparent.”