Venue: The Fuqua School of Business, Duke University, 1 Towerview Drive, Durham, NC 27708-0120
Presentation
Medicare Part D: Early Evidence on Prescription Drug Consumption, Hospitalization Offsets and Medicare Spending
Background: The Medicare drug benefit (Part D) implemented in January 2006 represents the most significant expansion of benefits since the program's inception more than 40 years ago. A major impetus for Medicare Part D was a concern about cost-related underuse of medications among the elderly. Presumably, the implementation of Part D will increase medication use through expanding drug coverage. This increased medication use may lead to cost offsets in Medicare Parts A and B if medication can better control chronic illness and lower rates of hospitalizations.
Objective: This study provides early evidence on the direct impact of Part D on medication use patterns, hospitalization offsets, and cost-savings for Medicare Part A and Part B.
Method: We collected all medical and pharmacy claims of 20,645 members from a large Medicare managed care plan between January 1, 2003 and December 31, 2006. We identified members who had the same medical plans but different pharmacy benefits (quarterly caps ranging from $150 to $1250 as well as a group with no quarterly cap). Their choices of generosity of drug plans before Part D solely depended on the availability of drug plans in their county of residence or through retiree health benefits but not on their choices based on the expected utilization. This valuable feature of our datasets limits the selection biases. We used a time series and comparison group design to measure the changes in medication use patterns and non-drug Medicare spending before and after the implementation of Medicare Part D. We used time series models to estimate the impact of Part D on changes in out-of-pocket drug expenditures, total drug expenditure, and total number of monthly prescription drug scripts, stratified by level of pharmacy benefits. We then identified a cohort whose demand for medication increased after the implementation of Part D, to examine the policy impact on hospitalization offset and other non-drug Medicare spending. To investigate the policy impact on long-term medication adherence and treatment discontinuation, we selected five therapeutic drug classes: antihypertensive, lipid-lowering, antidiabetic, antidepressant, and antipsychotic agents. We used generalized-estimating-equation (GEE) model to estimate the impact of Part D on proportion of days covered in the year for selected therapeutic drug classes, controlling for all individual characteristics.
Preliminary results: We found that Part D reduced out-of-pocket expenditures among the elderly, except for members who had no cap limits on pharmacy benefits before Part D. Out-of-pocket drug spending declined most among members who had less generous pharmacy benefits before Part D ($150 Cap, $250 Cap, and $350 Cap). Part D increased the demand of prescription drugs for members who had less generous drug plans with quarterly $150, $250, or $350 cap, but not for members without caps before Part D. We also found that the seasonal patterns associated with quarterly caps on medication before Part D were smoothed out. However we did not find that Part D offset non-drug medical spending. Subsequent analyses will examine potential medical cost offsets due to drug spending by chronic condition and by prescription drug limits. Conclusions: We found Part D decreased out-of-pocket pharmacy expenditures and increased demand for drugs for members who had less generous pharmacy benefits. In our analysis pooling several conditions together, we did not find significant savings from reduced non-drug medical services.