Health Affairs: Medicare subsidy payment needs refinement

Medicare payment plans do not sufficiently reimburse insurers for high medication use among its beneficiaries. This led to plans avoiding or dropping the Medicare population who receive low-income subsidies; however, fine-tuning the risk-adjustment approach and monitoring the incentives of Part D could help mitigate this problem, according to a study published online Oct. 28 in Health Affairs.

To better understand whether Medicare payments sufficiently compensate Medicare beneficiaries within the Part D low-income subsidy program, John Hsu, MD, MBA, director of the program for clinical economics and policy analysis at the Mongan Institute for Health Policy and Massachusetts General Hospital in Boston, and colleagues used data from Medicare prescription drug plans to determine payment adjustments.

“Since 2006, numerous insurers have stopped serving the low income segment of the Medicare Part D program, forcing millions of beneficiaries to change prescription drug plans,” the authors wrote.

Hsu and colleagues said that one-third of all Part D beneficiaries are enrolled in the low-income subsidy program, and that Medicare increases payments for beneficiaries who receive these subsidies by applying a multiplier that reflects greater medication use even after risk adjustment. Medicare currently pays 8 percent for beneficiaries who receive the full subsidy and 5 percent for those who receive a partial subsidy—9.6 million beneficiaries received a full or partial subsidy in 2009.

Hsu and colleagues examined individual drug spending of beneficiaries receiving the subsidy and compared them to those who didn't receive the subsidy. They also compared the actual and expected ratios of plan spending for subsidy to nonsubsidy beneficiaries under the current risk-adjustment approach.

The average risk score for nonsubsidy beneficiaries was 1.0 compared to the average risk score of 1.3 for low-income subsidy beneficiaries, “reflecting their greater disease burden,” the authors wrote. Average drug spending for full-subsidy beneficiaries compared with nonsubsidy beneficiaries was 1.21. The rate for partial-subsidy beneficiaries compared with nonsubsidy beneficiaries was 1.09.

“Plans spend 21 percent more on average for a full subsidy beneficiary than for a nonsubsidy beneficiary,” Hsu and colleagues wrote. This amount was much higher than the 8 percent in the current Centers for Medicare and Medicaid Services Part D payment approach.

The authors said that one way of reducing the incentive of plans to avoid beneficiaries under the subsidy would be to incorporate prior-year medication use into the risk adjustment model, which would help to improve payment accuracy.

“In short, payment accuracy, particularly in the low-income subsidy market, is important not only to reduce plans’ incentives to dump low-income patients and target less costly patients—a practice known as 'cream skimming'— but also to maintain plans’ participation in the subsidy market and minimize the growth of premiums for the entire Part D program,” the authors wrote.

Additionally, Hsu and colleagues said that Medicare could raise the existing low-income subsidy multiplier or increase total Part D dollars allocated to low-income beneficiaries. They said improving the accuracy of the risk-adjustment approach by including prior-year drug use would “help close the spending gap between subsidy and nonsubsidy beneficiaries.

“According to our findings, the current payment approach does not adequately account for the higher drug spending of the beneficiaries intended to be covered by the subsidy,” they  wrote. “This creates a perverse incentive for plans to avoid these beneficiaries, which the plans can do by increasing premiums above regional benchmarks.

“Combined with ongoing changes in the Part D low-income subsidy market, our findings highlight the critical need to monitor and refine the actual incentives of Part D, as well as to watch for unintended consequences. In effect, postmarket surveillance of Medicare policy is needed,” the authors concluded.