Differences in Patients with Progressive Versus Non-Progressive AML, CLL, or NHL: Implications for Proposed Bundled Payment Strategies

Background: Increasingly, payers are redefining how payments for cancer care are structured. Demonstration programs and policy proposals by insurers and oncology organizations have bundled payments for various services, including chemotherapy. The impact of such payment policies may vary for patients (pts) with hematologic malignancies given their widely varying disease severity. Progression of disease (PD), and costs associated with it, may stress bundled payment programs, particularly if such programs do not accurately assess risk of progression across the insured cohort. To inform this ongoing development of new payment strategies for pts with hematologic malignancy, this study compared patterns of care in pts with AML, CLL or other forms of NHL who do, or do not, experience PD.

Methods: This retrospective studyused medical and pharmacy claims from a large national US health plan to identify commercially insured and Medicare Advantage (MA) pts age ≥18 years from 1/2007 - 8/2014 with ≥2 medical claims for AML (ICD-9-CM code 205.0x), CLL ( ICD-9-CM code 204.1x), or other NHL (ICD-9-CM codes 200.xx, 202.0x-202.2x, 202.4x, 202.7x-202.9x, 203.8x, 204.8x-204.9x, 273.3x). Pts required ≥1 claim for systemic anti-cancer therapy (SACT); the first observed claim was the index date. Continuous enrollment (CE) in the health plan for 6 months (mths) prior to (baseline period) and ≥6 mths after index date (variable follow-up period) was required; pts with <6 mths of follow-up due to death were included. Pts with baseline SACT or additional primary malignancies were excluded. Line of therapy (LOT) periods were defined. The 1^st LOT (LOT1) started on index date; regimens included all drugs received in the first 45 days. LOT1 ended at the earliest of: start of a new drug, ≥60-day gap in initial regimen, death or end of CE or study period. LOT2 started with a SACT after LOT1 end. PD was defined as: start of LOT2, receipt of hospice care (based on procedure or revenue codes) or death (based on Social Security Administration death data).

Results: Among 667 AML, 1354 CLL and 9399 NHL pts who met study criteria, 70%, 45%, and 46% respectively had PD during the study period. Mean (median) time in mths to PD was 6.6 (4.2) for AML, 12.8 (9.2) for CLL and 10.0 (7.1) for NHL. Descriptive results are shown in the Table. Compared to pts without PD during the study period, pts who progressed were MA pts, older and had shorter initial LOTs. The most common initial therapy varied across PD cohorts. Among pts with PD, the AML cohort had the highest percentage of death and evidence of hospice care.

Conclusion: Characteristics and treatment patterns varied for pts with PD versus non-PD AML, CLL and NHL. Understanding the variability across patient groups will aid in the development of new bundled payment policies and help providers determine whether and in which payment systems to participate.