The effect of Medicare Part D coverage on drug use and cost sharing among seniors without prior drug benefits

Abstract

This study evaluates the effect of Medicare Part D among seniors who previously lacked drug coverage, using time-trend analyses of patient-level dispensing data from three pharmacy chains. Of 114,766 seniors without drug benefits, 55 percent initiated drug insurance under Part D. After the penalty-free Part D enrollment period, use of statins, clopidogrel, and proton pump inhibitors stabilized at levels ranging from 11 percent to 37 percent above the trend that would have been expected if Part D had not been implemented. Patients reaching the Part D coverage gap (12 percent) experienced a decrease in essential medication use ranging from 5.7 percentage points per month for warfarin to 6.3 percentage points for statins.

EXHIBIT 1. Prescription Drug Coverage Among Elderly Medicare Beneficiaries Who Did Not Have Drug Coverage Before The Implementation Of Medicare Part D, By Month, January 2005–December 2006

SOURCE: Authors’ analysis of computerized pharmacy dispensing information from three pharmacy chains. NOTE: Subjects without drug coverage were identified as those who paid 60 percent or more of a drug’s price for 80 percent or more of drugs dispensed in 2005.

EXHIBIT 2. Characteristics Of Elderly Medicare Beneficiaries Who Continuously Filled Prescriptions In One Of Three Pharmacy Chains Over The Study Period, January 2005–December 2006

SOURCE: Authors’ analysis of computerized pharmacy dispensing information from three pharmacy chains. ^a According to patient ZIP code and national survey data.

EXHIBIT 3. Use Of Selected Essential Drugs Among Elderly Medicare Beneficiaries Who Did Not Have Drug Coverage Before The Implementation Of Medicare Part D, January 2005–December 2006

SOURCE: Authors’ analysis of computerized pharmacy dispensing information from three pharmacy chains. NOTES: Generally we were interested in studying the effects of Part D on the entire drug classes selected for this study. For statins as well as clopidogrel, generic formulations became available during the study period, which enabled us to assess the amount of generic substitution under Part D for those drug classes. Esomeprazole was broken out because it is therapeutically equivalent to omeprazole but priced higher; this allowed us to study potentially wasteful prescribing under Part D. DDDs are defined daily doses. PPIs are proton pump inhibitors.

EXHIBIT 4. Changes In Utilization And Copayment Trends Under Medicare Part D For Selected Medication Groups Among Elderly Medicare Beneficiaries With No Prior Drug Coverage, Relative To 2005 Baseline Trends

SOURCE: Authors’ analysis of computerized pharmacy dispensing information from three pharmacy chains. NOTES: Transition period was January–April 2006. Stable Part D period was May–December 2006. Baseline period was January–December 2005. The R² value was 0.89 or greater for all models. DDD is defined daily doses.

EXHIBIT 5. Out-Of-Pocket Spending For Selected Essential Drugs Among Elderly Medicare Beneficiaries Who Did Not Have Drug Coverage Before The Implementation Of Medicare Part D, January 2005–December 2006

SOURCE: Authors’ analysis of computerized pharmacy dispensing information from three pharmacy chains. NOTES: See Exhibit 3 Notes. DDDs are defined daily doses. PPIs are proton pump inhibitors.

EXHIBIT 6. Use Of Four Selected Essential Drugs Among Elderly Medicare Beneficiaries Who Did Not Have Drug Coverage Before The Implementation Of Medicare Part D, Who Reached The Part D Coverage Gap Before 1 September 2006

SOURCE: Authors’ analysis of computerized pharmacy dispensing information from three pharmacy chains. NOTES: Two data points were omitted from the analysis. We excluded the thirty-day interval in which the coverage gap was reached (interval 0), because in that interval, by definition every patient had one or more prescription fills to push the cumulative spending above the Medicare Part D threshold. Because, by design, we will find particularly high usage in interval 0, we would expect to find unusually low use in the interval immediately preceding (interval −1), because the average time between refills is about thirty days. PPIs are proton pump inhibitors.