Recent Research Reveals the Impact of Drug Rebates on Formulary Design in Medicare Part D module

Drug prices for patients enrolled in Medicare Part D are determined by private negotiations and undisclosed rebates between insurance companies (or their intermediaries, Pharmacy Benefit Managers) and drug manufacturers. These rebates are essentially discounts off the list price of a drug. A key feature is that these discounts are "formulary-contingent," meaning manufacturers pay larger rebates if their drug is placed on a more favorable tier of the insurance plan's list of covered medications, its formulary.  Placing a drug on a “preferred" tier means lower out-of-pocket costs for patients and higher demand for the drug.

HSI faculty researcher, Prof. Mert Demirer collaborated with Prof. Alexander L. Olssen from Wharton to study these confidential rebates, uncovering their effects on health plan design, patient costs, and overall well-being. The researchers focused on prices of branded statins (cholesterol-lowering drugs, such as Crestor and Lipitor) in 2010. Statins were a significant part of Part D prescriptions, and the market included both branded and generic options, making it an ideal case study.

Instead of trying to model the complex negotiations, the researchers built a model of how patients choose plans and statins, and how insurers, wanting to maximize profits, choose their formularies. By observing which formulary choices insurers actually made, they could infer the range of rebates that would make those choices profitable for the insurers. This method, known as "moment inequalities," allowed them to estimate the rebates without knowing the exact negotiation process.

Key Findings on Rebates and Their Impact:
1. Rebate amounts are significant:
    ◦ The study estimates that the average per-unit rebate for placing branded statins on the preferred tier was around 37.3% for Crestor and 36.2% for Lipitor.
    ◦ These are considerably higher than the average branded drug rebates across all Part D drugs in 2010 (estimated at 13.8%).

2. And they have a significant impact on formularies:
    ◦ Rebates have a large effect on formulary design. For example, increasing Lipitor rebates from 35% to 50% would lead to an 18 percentage point increase in large plans placing Lipitor on the preferred tier.
    ◦ Sometimes, increasing a rebate for one drug can negatively affect another drug. If Crestor rebates increased, some plans would move Lipitor to a less preferred tier, potentially harming Lipitor users.

3. They also have a significant impact on Beneficiary Costs and Well-being:
    ◦ Reducing or Eliminating Rebates: If rebates were prohibited entirely, the share of plans covering both Crestor and Lipitor on the preferred tier would fall drastically by 66 percentage points. This would lead to a significant increase in out-of-pocket costs for branded statins (over 50% increase) and a 5.21% decrease in consumer surplus. The researchers used consumer surplus, the difference between what someone would be willing to pay for a good and what they actually pay, as a measure of consumer well-being. So, as the out-of-pocket costs to the consumers increase, consumer well-being decreases. People would also use branded statins less (Crestor down 25%, Lipitor down 35%).

    ◦ Increasing Rebates: If rebates were increased enough to match the lower prices paid in Canada (a 48.4% rebate for both), beneficiaries buying branded statins would see their cost-sharing fall by about 7-9%. This would lead to a 1.8% increase in consumer well-being.

4. As well as an impact on Premiums:
    ◦ Changes in branded statin rebates have a smaller effect on monthly insurance premiums compared to their effect on out-of-pocket costs. For instance, eliminating rebates might increase premiums by about 5.63%, which is much less than the increase in cost-sharing.

Does the Current System Work Well? 
The researchers conclude that, for statins, the current private negotiation system in Medicare Part D works well for consumers. The estimated rebates are consistent with the system achieving more than half of the potential consumer well-being in the statin market. This is partly because there are already sufficient choices for branded statins on preferred tiers. However, they caution that this finding may not apply to other drug markets with different competitive structures.

This research highlights that drug rebates are a powerful force in Medicare Part D, significantly influencing which drugs insurers favor, how much patients pay out-of-pocket, and ultimately, consumer well-being and insurer profits. Policy changes to these rebates would have substantial, and sometimes complex, effects on the drug insurance landscape.