Role of Sacubitril/Valsartan in Improving Provider Performance in Managing Heart Failure Under Medicare Alternative Payment Models

This drug utilization study employed a retrospective cohort design using the 100% files of 2015-2018 Medicare enrollment and Part A, B & D claims data in conjunction with publicly available information on providers participating in Bundled Payments for Care Improvement (BPCI) and/or Medicare Shared Savings Program (MSSP).

BPCI Study Design The unit of analysis for addressing study objectives was a unique Congestive Heart Failure (CHF) episode of care, which encompassed an initial acute inpatient stay plus all Part A & B covered medical services during the 90 days post hospital discharge. Episode index date was defined as the date of a beneficiary's initial acute hospital stay with a qualifying Medicare Severity-Diagnosis Related Group (MS-DRG) for CHF, triggering a CHF episode of care. Beneficiaries were followed longitudinally from episode index date through episode end date to measure their utilization of sacubitril/valsartan and Part A & B expenditures during the CHF episode of care. A 12-month lookback period, based on when the episode was initiated, was utilized to capture beneficiaries' baseline characteristics. For example, episodes initiated on 2/1/2018 will have a 2/1/2017 to 1/31/2018 baseline lookback period. Balancing the need for data recency, all qualifying CHF episodes from 2016-2018 were included in the analysis. Given this is an episode-level analysis, a beneficiary may have contributed multiple CHF episodes to the data analysis. To account for potential within-subject clustering, a generalized linear model with a gamma distribution was applied for estimating the association between sacubitril/valsartan use and CHF episode costs.

MSSP Study Design The unit of analysis for addressing study objectives was a HF patient's Part A & B expenditures within the MSSP. Given the most recent year of data available at the time the study was conducted was 2018, the study population included HF patients for whom the MSSP participants were accountable in 2018 per Centers for Medicare & Medicaid Services (CMS) assignment. These beneficiaries' utilization of sacubitril/valsartan and annual Part A & B expenditures was measured from January 1, 2018 - December 31, 2018. Data from CY 2017 was leveraged for capturing these patients' baseline characteristics.

Inclusion criteria

• ≥18 years old during year of index date with valid gender and geographic information
• Continuous enrollment (CE) in the health plan with medical and pharmacy benefits for ≥6 months (180ii days) before the index date (pre-index period)
• CE with medical and pharmacy benefits for ≥18 months (540ii days) beginning on the index date (post-index period)
• All patients had uniform 6-month pre-index and 18-month post-index periods
• ≥1 medical claim with an MS diagnosis codeiii in any position during the pre-or post-index periods
• ≥1 claim for fingolimod after the index date (i.e., from index date +1 to 539 days post-index)
• By requiring ≥1 medical claim with an MS diagnosis code and ≥2 claims with an NDC for fingolimod, the inclusion criteria essentially incorporated the preferred/overall best performing MS case-finding definition validated by Culpepper et al. The criteria for that definition required at least 3 separate encounters from any of the following: MS-related ambulatory visits, MS-related inpatient stays, and DMT claims during a 12-month period. Slightly higher accuracy was found when a 24-month period was used.

Exclusion criteria

≥1 pharmacy or medical claim for any MS DMT during the pre-index period