Beneficiaries who automatically qualify or successfully apply for the low-income subsidy and do not select a Part D plan are randomly assigned to a plan with a below average to average premium for the region. Researchers from the University of Pittsburgh compared an “intelligent reassignment” algorithm against Part D’s random placement and determined that intelligently assigning low-income beneficiaries using past drug claims would have saved Part D over $5 billion in 2009. Given that overall Part D spending totaled $52.5 billion in 2009, these findings present the potential for significant cost reduction. Furthermore, matching beneficiaries to the plan that best covered their drug needs reduced the number of prescriptions subject to utilization review and quantity limits. By law, however, Part D randomly assigns low-income beneficiaries who do not select a plan to prevent a single plan from receiving a disproportionate number of very sick beneficiaries. Thus, any efforts to intelligently assign beneficiaries in the future would require legal changes.
Medicare Part D provides a subsidy to beneficiaries with incomes below 150 percent of the federal poverty level. Enrollees with the low-income subsidy accounted for 75 percent of the $60 billion in total federal Part D spending in 2013. The government randomly assigns any new beneficiary who automatically qualifies for the subsidy, or who successfully applies for it without indicating a preferred plan, to a stand-alone Part D plan whose premium is equal to or below the average premium for the basic Part D benefit in the region. We used an intelligent reassignment algorithm and 2008–09 Part D drug use and spending data to match enrollees to available plans according to their medication needs. We found that such a reassignment approach could have saved the federal government over $5 billion in 2009, for mean government savings of $710 (median: $368) per enrollee with a low-income subsidy. Implementing that simple change to reassign beneficiaries would have also lowered the proportion of prescriptions that required utilization review from 29 percent to 20 percent, and the proportion of prescriptions with quantity limits from 27 percent to 19 percent.