As Medicare searches far and wide for ways to reduce costs, there's the perennial question: Is the Medicare Prescription Drug Plans (Part D) program saving money or needlessly running up costs?

Based on some superficial evidence, Part D seems to be a cost saver. But if you poke beneath the surface, the savings are less clear-cut and deserve more examination. A recent study by Georgetown Professor Jack Hoadley for the Kaiser Family Foundation asks some questions that Medicare needs to explore further.

Part D was one of the first hybrid offerings introduced into Medicare as part of the Medicare Modernization Act of 2003. Along with Medicare Advantage Plans (Part C), the idea behind these subprograms was to introduce private competition within Medicare’s fee-for-service (FFS) structure and bring costs down. Part D, which was started six years ago, expanded Medicare benefits to include prescription medications. For-profit companies, which received subsidies for participating in the plan, offered drugs at discounts.   

Around the time of Part D’s inception, the Congressional Budget Office (CBO) made some estimates as to how much the program would cost. The CBO numbers would provide a future baseline to determine whether the drug program — through the private competitive model — would save Medicare money.

Recent evaluations of Part D's performance have been embraced by groups who favor keeping the current model of the plan.

“When was the last time anyone heard of a government program coming in under budget year after year and generating nearly a 90 percent satisfaction rate?" asked Rep. Cory Gardner (R-Colo.) in a recent statement. "That’s exactly what’s happening right now in Part D,” Gardner said. “It is a federal program coming in under projected costs." 

Gardner was part a group of more than 100 members of Congress who are urging Health and Human Services (HHS) Secretary Kathleen Sibelius to "protect and preserve the competition that has made Medicare Part D work."

The most recent tally of the drug plan's performance is that Part D's net operating costs are some 30 percent lower than what the CBO projected in 2003. While at first that seems like good news and gives the impression that competition within the plan is keeping costs low, Hoadley takes a more nuanced view.

It would be easy to jump to the conclusion that the head-to-head competition among companies was responsible for the lower-than-expected costs for Part D, but Hoadley isn't entirely convinced. Here are some of his insights:

• Enrollment was lower than expected. Part D is voluntary, so beneficiaries can choose not to sign up for it, and that may have impacted the program’s total costs. The CBO overestimated enrollment; it projected in 2003 that 87 percent of Medicare beneficiaries would sign up, but as of 2012, only 73 percent have done so. Lower enrollment means lower costs, so competition may not have been a factor.
• Lower overall drug spending had an impact. The CBO also guessed incorrectly on total drug spending. It projected 12 percent average per capita growth in prescription expenditures by 2006, then a 9 percent rate thereafter. The actual rates have been 10 percent and 4 percent, respectively.  
• Generics played a big role. While this factor may not have been fully explored by the CBO, patients increasingly took advantage of patent expirations on brand-name drugs that then became available as generics at a lower cost. As a result, Part D increased from a 67 percent usage rate of generics in 2007 to more than 75 percent in 2010 (the current rate, not cited, is likely to be higher).
• Fewer drugs have been approved for Part D. Over the years, fewer patented drugs have been in use due to the introduction of more and more generics. That lowers cost over time — even more so in recent years as patents have expired on brand-name medications used by Medicare patients.

"Based on the available evidence," Hoadley concluded, "the claim that Part D spending is lower than originally projected due to private plans seems overstated, given the multiple factors that influence drug spending trends."

What Hoadley didn't explore is the profile of beneficiaries currently using Part D. Do they tend to be younger and healthier than the overall Medicare population? What about the people who didn't sign up for Part D? Were they getting better deals on drugs elsewhere, perhaps through Medicare Supplement Insurance (Medigap) plans with prescription coverage? Wouldn't it be worth asking them why they didn't sign up for Part D?

What would be even more useful for the Centers of Medicare and Medicaid Services, as it evaluates the relative costs of running Part D, is to use single-payer drug plans as comparative models. How does Medicare's plan compare to Medicaid or the Veterans Administration? Do these programs offer more or less generics in their plans? What is the average discount or rebate offered in the non-Medicare programs?

What if Medicare required specific discounts in Part D or looked at how medication purchased at huge discounts through Medicare Medical Insurance (Part B) could be applied to Part D? Part B, for example, buys large quantities of chemotherapy and other drugs used in clinical or hospital settings. How can this discounting system be applied to Part D?

It would also be worthwhile to look at Canada's single-payer plan or other countries that buy drugs in bulk at huge discounts.

In that context, it's worth exploring another Kaiser report published earlier this month, entitled "Prescription Drug Procurement and the Federal Budget." The study notes that by applying Medicaid rebates to drugs bought by low-income beneficiaries, rather than those prices negotiated by Part D plans, the government could save over $100 billion over 10 years (see link below).

One of the greatest flaws in any large-scale evaluation of a program as large as Part D is that it should be compared to something akin to what researchers call a "control group" or comparative model. The 30 percent savings cited above were based on the CBO's projections (which are often off) and not compared to a working system like Medicaid, which employs large-scale purchasing. That's why even more questions should be posed as to the efficacy of the hybrid private–public model.