Drug Costs Tied to Part D Spending Differences

Regional variations in Medicare Part D spending can be mainly attributed to a difference in the cost of drugs, rather than the volume of drugs used, according to the results of a new study. The study’s findings were recently reported in the New England Journal of Medicine [2012;366(6):530-538]. Although regional variation in Medicare Part D spending has been noted in previous research, there has been little investigation into why this variation occurs. In this study, researchers used 2008 Medicare data to determine whether this variation was due to differences in volume or medication choice by measuring regional variation in overall pharmaceutical expenditures as well as the variation in 3 drug categories. Researchers assessed variation in pharmaceutical spending for angiotensin-converting enzyme (ACE) inhibitors and angiotensin-receptor blockers (ARBs), 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors (statins), and 2 forms of antidepressants (selective serotonin reuptake inhibitors [SSRIs] and serotonin-norepinephrine reuptake inhibitors [SNRIs]). As part of the study, researchers decomposed differences in regional expenditures across 306 hospital-referral regions (HRRs) by calculating annual prescription volumes and cost per prescription. They calculated the figures by using data gathered from a 40% random sample of the 2008 Medicare Denominator file. Researchers identified approximately 4.7 million beneficiaries who were continuously enrolled in fee-for-service Medicare and stand-alone Part D prescription drug plans (PDPs). Those enrolled in Medicare Advantage PDPs were excluded. Once beneficiaries had been placed in HRRs based on their zip codes, researchers assessed per capita annual prescription-drug expenditure, per capita annual number of prescriptions filled, cost per prescription filled, and ratio of branded-drug prescriptions to total prescriptions filled for the HRRs. Primary outcomes of the study included annual prescription volume and cost per prescription. Researchers found that after adjusting for demographic, socioeconomic, and health-status differences among beneficiaries, the difference in per capita prescription-drug spending ranged from $2413 in the lowest quintiles of HRRs to $3008 in the highest HRR quintiles, a variation of 24.7%. They also found that 75.9% of this regional variation for all prescription-drug spending could be attributed to the cost per prescription filled, while 24.1% was related to volume differences. This trend was also seen for ACE inhibitors and ARBs, where 87.5% of regional variation could be explained by cost per prescription, and again for statins, where 56.3% of regional variation could be explained by cost per prescription. But, this was not the case for SSRIs and SNRIs. Researchers found that just 36.1% of regional spending variation in this drug category could be explained by cost per prescription, while 63.9% was explained by differences in prescription volume. The study also found that there was a high correlation between cost per prescription at the regional level and the ratio of branded-drug prescriptions to total prescriptions. When assessing the ratio of branded drugs to total prescriptions across HRRs, the study’s authors found a range for overall drug use of 0.24 to 0.45. They also noted individual ratios for ACE inhibitors and ARBs (ranging from 0.24 to 0.55), statins (ranging from 0.29 to 0.60), and SSRIs and SNRIs (ranging from 0.15 to 0.51). Researchers concluded by noting that there could be potential for cost savings if some quintiles use fewer brand name drugs. They estimated that the Medicare program and beneficiaries could have saved $4.5 billion if branded-drug use in all HRRs was similar to the use in the lowest quintile. However, they acknowledged several limitations to their study. For instance, researchers assessed variation in 3 drug categories regardless of the clinical indication for use. In addition, figures were not adjusted for Part D plan characteristics, and the risk-adjustment measure used in the study may not have adequately adjusted for health status. Other limitations were a study design that only included people with 12 months of continuous enrollment in PDPs and also did not factor in any information about possible rebates with pharmaceutical manufacturers.