For years, the Medicare prescription drug benefit Part D has been credited with positively impacting national trends in health outcomes and medical services. But a recent study led by Becky Briesacher, associate professor at Northeastern University, and Stephen Soumerai, Harvard Medical School professor of population medicine, challenges that assumption and suggests that the U.S. Congressional Budget Office’s new costing method, based on assumed cost savings, may be “premature.”
Since its implementation in 2006, Part D has substantially increased access to prescription drugs for the nearly 50 million Medicare subscribers. That increased access, though, has not led to a clear decrease in emergency room visits, hospital stays, inpatient costs or mortality, according to the research by Briesacher and Soumerai, which included colleagues from HMS and Harvard Pilgrim Health Care Institute, where Soumerai is based.
“We are concluding that Medicare Part D did not save the [Medicare] program any money overall,” said Briesacher, a health services researcher in the Northeastern School of Pharmacy. “You have to be realistic about the fact that giving people access to medication is important, but it’s not always going to substantially save money in other parts of the health care system or keep a significant number of people out of the hospital.”
The team presented its results in a paper published in Annals of Internal Medicine.
About one year after Medicare Part D was launched, early studies were conducted among Medicare beneficiaries who either had no prescription drug coverage or poor coverage prior to Part D. Those early studies found these specific subgroups saw statistically significant decreases in non-drug related medical spending and hospitalizations.
But, the researchers say, these selected subgroups do not represent the experiences of Medicare subscribers at large, many of whom already had some type of drug coverage prior to Part D.
“All too often early studies of new treatments and policy interventions tend to show dramatic positive health effects that diminish or disappear as more rigorous studies are conducted,” Soumerai said. “In general, basing treatment and policy decisions on studies of small or potentially biased populations can lead to mistaken conclusions about causes and effects.”
To see if that was the case with Medicare Part D, the team widened the scope of the earlier analysis and looked at 11 years’ worth of data from the Medicare Current Beneficiary Survey, an annual face-to-face panel survey of about 12,000 Medicare subscribers.
The team found no significant change in subscribers reporting they were in poor to fair health five years after Part D was implemented. In 2006 that figure was 26.6 percent, while in 2010 it was 24.6 percent, a statistically insignificant finding that Briesacher attributes to pre-existing historical trends.
Also, overall emergency department visits and inpatient services stayed at about 13 percent for the entire study period.
The previously accepted early studies of Part D led the Congressional Budget Office, tasked with determining the cost of legislation, to adopt an algorithm based on the belief that increases in prescriptions filled across the Medicare population results in overall cost offsets.
According to the paper, the CBO methodology estimates Medicare spending on medical services is now routinely reduced by 0.2 percent for each 1 percent increase in drug prescriptions filled.
The budget impact of the flawed methodology is not trivial, according to the study. The researchers point to provisions in the Affordable Care Act to decrease Part D cost sharing that are based on the CBO’s prediction that Medicare’s non-drug spending will be reduced by $35 billion through cost savings in medical services, primarily decreases in hospitalization.
“We’d like the Congressional Budget Office to re-examine the policy,” Briesacher said. “It’s about properly scoring the legislation so it doesn’t assume these cost offsets that we can’t find.”