March 19, 2018 – A new PhRMA-funded study released by Milliman suggests that 340B hospitals have higher drug spending for commercially insured patients than non-340B hospitals and claims that it is the result of inappropriate drug use. However, methodological flaws in the study raise serious questions about the findings. In fact, Milliman’s report includes data that indicate those differences are due to the fact that 340B hospitals treat a sicker and more expensive patient cohort.
The study, which was released two days before a scheduled Senate committee hearing on the 340B program, reviewed drug spending over a one-year period in the commercial payer setting. The authors said they replicated a 2015 Government Accountability Office (GAO) analysis that examined Medicare Part B drug spending per patient and found such spending was “substantially higher” in 340B hospitals than in non-340B hospitals. The GAO suggested that the observed spending difference was the result of 340B hospitals using more drugs or more expensive drugs than necessary.
For this PhRMA study, Milliman applied a similar methodology to data from private insurers and reported similar results. They reported average outpatient drug spending per patient in 340B DSH hospitals was roughly three times the average spending in non-340B hospitals. The Milliman researchers said they controlled for differences in health status (sickness) between 340B and non-340B patients and found that differences in health status were minimal and did not explain the spending differences.
However, the Milliman study looked at data from all patients treated in hospitals, including patients who had not been given an outpatient drug. This caused their spending estimates to be inflated. If Milliman had, instead, looked only at patients who had received a drug, the results would have been quite different. In fact, the report acknowledges that when they compared drug spending using only patients who received a drug, they found patients in 340B hospitals had health status risk scores 28 percent higher than patients treated in non-340B hospitals. This is consistent with the findings of a 2017 analysis that shows Medicare drug spending per patient is higher in 340B DSH hospitals because they treat patients who are sicker, and therefore more expensive to treat.
Similarly, the Milliman study finds that the difference in oncology-specific drug spending between 340B hospitals and non-340B hospitals is “less pronounced.” The study found that the average spending on oncology drugs in 340B DSH hospitals per oncology patient is much more similar to spending levels in non-340B hospitals. Yet the study authors suggest that 340B hospitals are maximizing a financial incentive to use more drugs or more expensive drugs than necessary. If this were the case, one would expect to see the largest spending differences to be in the oncology space, given that cancer drugs are some of the most expensive drugs on the market and would, presumably, present the greatest financial incentive for overuse by 340B hospitals to generate greater financial margins. However, the study finds the exact opposite to be true – the spending difference is significantly smaller in the oncology space. This finding further supports the notion that spending differences are not the result of inappropriate drug use in 340B hospitals.
Rather than support the conclusion that 340B hospitals are using drugs inappropriately, this study appears to strengthen a growing body of evidence that 340B hospitals are unique safety-net providers — hospitals that treat significantly more low-income and rural patients who tend to be sicker and more expensive to treat.