Prescription drug benefit doesn't save money for Medicare

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 her team, which included col­leagues from Har­vard Med­ical School.

"We are con­cluding that Medicare Part D did not save the (Medicare) pro­gram any money overall," said Briesacher, a health ser­vices researcher in the School of Phar­macy with nationally-recognized exper­tise in drug policy and med­ica­tion use in older adults. "You have to be real­istic about the fact that giving people access to med­ica­tion is impor­tant, but it's not 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 non­drug med­ical spending and hospitalizations.

But, as Briesacher explains, 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.

Briesacher and her team widened the scope of the analysis and looked at 11 years worth of data from the Medicare Cur­rent Ben­e­fi­ciary survey, which is an annual face-to-face panel survey of about 12,000 Medicare subscribers.

They found no sig­nif­i­cant change to 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, which is sta­tis­ti­cally insignif­i­cant and which Briesacher con­tributes to pre-existing his­tor­ical trends.

Also, the emer­gency depart­ment trips and inpa­tient ser­vices stayed the same 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 that works off the belief that increases in pre­scrip­tion fills 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 Con­gres­sional Budget Office's pre­dic­tion that Medicare's non­drug 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."