Persistence of Enrollment Among Commercial Enrollee Populations with Rare Hematologic Diseases: Implications for Long-Term Outcome Assessment and Value-Based Agreements

Background: Innovative therapies, including gene therapies for incurable hematologic conditions, have highlighted the need for long-term assessment of outcomes. Administrative claims are an important source for observing patients over time, but enrollment "churn" across health plans can limit the length of meaningful follow-up analysis. This analysis examined a large, United States (US) dataset with complete enrollment data for health plan members in order to assess continuous enrollment (CE) among commercial populations with and without rare hematologic diseases.

Methods: From January 1, 2016 to December 31, 2021 (identification period), a matched-cohort analysis was used to compare enrollment data for commercial members using the Optum Research Database. Members were grouped into the following disease cohorts based on diagnosis codes on non-diagnostic medical claims and/or treatment during the identification period: sickle cell anemia (SCA), beta thalassemia (beta-thal), and hemophilia (HA or HB). SCA was defined using ≥2 claims, on different dates, with an SCA diagnosis code. Beta-thal was defined using ≥1 claim with a beta-thal diagnosis code and ≥2 claims, on different dates, for blood transfusion or luspatercept. Hemophilia was defined using ≥1 claim with a HA/HB diagnosis code and ≥2 claims, on different dates, for on-demand or prophylaxis treatment specific to HA/HB (factor VIII and IX concentrates and emicizumab-kxwh). Members without these rare diseases were eligible for the control cohort if they had at least 1 medical claim during the identification period. The date of the first identifying claim defined the index date. Persistence of CE (medical and pharmacy) was measured in months (30-days), from the index date to the earlier of the disenrollment date or end of data. Rare disease and control cohorts were matched 1:3 without replacement on age, gender, index year, geographic region, fully-insured/administrative services only coverage, and subscriber/dependent status. Member characteristics were descriptively summarized overall, and results were stratified by cohort, index year, and age group (at index).

Results: Overall, among ~25.7 million members, 3,459 were identified with a rare hematologic disease of interest (Table 1). Per rare disease cohort, 2,168 with SCA, 422 with beta-thal, and 869 with HA/HB were identified; the average age (standard deviation, SD) was 29.0 (15.1), 36.8 (20.7), and 24.5 (16.0) years in each cohort, respectively. The average (standard deviation [SD]) CE length in months among those with rare disease compared with respective controls was shorter in SCA (19.2 [17.8] vs. 20.4 [18.4]), longer in beta-thal (22.5 [18.1] vs. 20.6 [18.1]), and longer in HA/HB (23.7 [19.4] vs. 21.0 [19.2]) (Table 2). This pattern was consistent across all index years stratifications and among pediatric patients and adults <45 years old.

Conclusions/Implications: Continuous enrollment time was overall shorter among members with SCA compared with their controls but longer for those with beta-thal and hemophilia compared with their controls. Even among those with the earliest index year (thus maximum possible follow-up time of ~72 months), average persistence of enrollment ranged from 25-31 months across disease cohorts and their controls. Notably, these findings reflect average enrollment for commercial members, in the setting of no minimum enrollment requirements. By contrast, claims-based research often requires a minimum duration of CE for a particular study population, which often increases the average CE observed. CE varied not only from matched controls, but also across SCA, beta-thal, and HA/HB cohorts, supporting the need to specifically evaluate enrollment in eligible treatment populations.

Value-based agreements traditionally have involved commercial populations and have contracted on short-term (6-12 month) outcomes. Providers, payers, and manufactures will need to assess current data assets, available follow-up, and innovative data-sharing solutions that allow for patient portability (e.g., multi-payer platforms) to measure long-term outcomes. These methodological considerations will provide the foundation for aligning risk-sharing interests so that patients may access and benefit from these innovative therapies.

Vu:Optum: Current Employment, Current equity holder in private company. Song:Optum: Current Employment, Current equity holder in private company. Engel-Nitz:Optum: Current Employment, Current equity holder in private company. Hulbert:Optum: Current Employment, Current equity holder in private company.