Process and validity of linking cystic fibrosis patient registry with national Medicaid databases

J Cyst Fibros. 2024 Nov 19:S1569-1993(24)01805-8. doi: 10.1016/j.jcf.2024.10.012. Online ahead of print.

Abstract

Background: The Cystic Fibrosis Foundation Patient Registry (CFFPR) provides valuable clinical and demographic data but includes limited information on health services and medications provided outside of CF Care Centers. Linking CFFPR to claims databases such as national Medicaid data could address these data gaps.

Methods: Linkage algorithms based on state of residence, gender, and date of birth were utilized to match individuals with CF diagnostic codes in national Medicaid databases (2016) to individuals in the CFFPR (2015-2016). Subsets of individuals with partial social security numbers or residing in the state of North Carolina were utilized to validate the accuracy of linkages and perform exploratory analyses on care utilization and costs.

Results: Of the 32,152 individuals in CFFPR, 10,616 were uniquely linked to national Medicaid databases. The 372 linked individuals within the NC extract had 8.0 ± 7.6 visits to outpatient providers, substantially higher than the 4.2 ± 2.4 CF Care Center outpatient visits documented within CFFPR. Similarly, linked individuals had 2.1 ± 1.7 oral antibiotic prescriptions within CMS pharmacy databases versus 0.5 ± 1.9 oral antibiotic prescriptions in CFFPR. Total pharmacy costs for the linked individuals in NC were $16.4 million, with pancrealipase (19 %), dornase alfa (24 %), and CFTR modulators (29 %) the largest expenditures. Total non-pharmacy costs were $7.5 million, with inpatient hospitalization representing 53 % of costs.

Conclusion: Linkage of data from Medicaid and CFFPR can produce valid comprehensive data on low-income people with CF and provide opportunities to examine utilization/adherence or comparative effectiveness and safety of medications as well as conduct economic analyses in the low-income CF population.

Keywords: CFFPR Medicaid deterministic linkage.