High Cost Sharing and Specialty Drug Initiation under Medicare Part D: A Case Study in Newly Diagnosed Chronic Myeloid Leukemia Patients

Monday, June 13, 2016: 10:15 AM
G60 (Huntsman Hall)

Author(s): Jalpa Doshi; Pengxiang Li; Hairong Huo; Amy R. Pettit; Rishab Kumar; Brendan Weiss; Scott Huntington

Discussant: Amy J. Davidoff

Specialty drugs often offer significant medical advances for patients with chronic and/or life threatening diseases but frequently carry high out-of-pockets costs that may impede access. This is particularly true in the case of Medicare Part D, as CMS allows plans to place drugs exceeding a cost threshold ($600/month from 2010-2015) on “specialty tiers.” Today, virtually all Part D plans using tiered benefit structures have a specialty tier. After meeting a deductible (where required), patients typically face specialty tiers with 25%-33% coinsurance during an initial coverage phase, followed by ~50% coinsurance during the coverage gap phase, before entering the catastrophic phase with 5% coinsurance. Yet no studies have examined the impact of such high cost sharing on specialty drug initiation under Part D.

Tyrosine kinase inhibitors (TKIs) have revolutionized the treatment of chronic myeloid leukemia (CML), making it an apt case study. Utilizing 2011-2013 100% Medicare claims, we examined rates of TKI initiation and time to initiation among fee-for-service Medicare Part D patients newly diagnosed with CML (captured via diagnostic and lab test codes and absence of previous TKI claims), as compared to their counterparts who were receiving low-income subsidies (LIS) and facing nominal cost sharing (≤$5). All TKIs were subject to 25%-33% coinsurance across all plans in our sample. Further, given median total costs of ~$6500 per 30-day TKI prescription, the first drug fill “straddled” Part D benefit phases, resulting in an out-of-pocket cost >$2500 for non-LIS patients. We found that non-LIS patients were less likely than LIS patients to have a TKI claim within 6 months of diagnosis (44.9% vs. 67.8%, p<0.05) and on average, took twice as long to fill one when they did (mean 50.1 vs. 23.9 days, p<0.05). Cox regressions controlling for sociodemographic, clinical (e.g., CML severity), and plan characteristics (e.g., utilization management tools) confirmed descriptive findings (HR=0.61, p=0.001). Sensitivity analyses using 6 alternate algorithms to identify new CML patients, and using different model specifications (i.e., plan ID fixed effects to ensure LIS and non-LIS patients were facing the same plan-level restrictions aside from cost sharing), and using a different definition of index date (i.e., using the second claim as the index date to count time to TKI initiation) showed consistent findings.

In summary, high cost sharing among non-LIS patients was associated with reduced and/or delayed access to TKIs under Part D. Although the Part D coverage gap will be phased out by 2020, 25%-33% specialty tier cost sharing will remain and may create financial barriers that inadvertently discourage use of cost-effective treatments.