Enhancing CF registry data quality: A collaborative endeavour

Abstract

Introduction/Aim: The Australian Cystic Fibrosis Data Registry (ACFDR) has collected and reported data in people with cystic fibrosis (CF) since 1998. Regular data collection is, however, a significant burden, potentially limiting data completeness. In 2019 Monash University, CFA and Vertex Pharmaceuticals entered into a collaborative agreement to improve the timeliness, completeness and standardisation (i.e., quality) of ACFDR data. In 2018 lumacaftor/ivacaftor (LUM/IVA) was reimbursed in Australia for people aged ≥6 years homozygous for F508del, conditional on data reporting to the Pharmaceutical Benefits Advisory Committee. We describe the improvement in quality of key ACFDR data items before and three years into the collaboration. We also demonstrate enhanced usability of the data via an analysis of registry LUM/IVA data incorporating this period. Methods: The collaborative agreement supported registry database enhancements and quarterly site data collection. The quality of core data items including genotype and CFTR modulator (CFTRm) use were compared for 2017 and 2021. To demonstrate impact of the collaboration, registry data were extracted for F508del homozygous persons aged ≥12 years initiated on LUM/IVA, with annualised change in ppFEV 1 calculated using each individual's sequential ppFEV 1 measurements throughout the follow-up period. To focus on change in ppFEV 1 after acute improvement, the first post-initiation measurement was excluded. Results: Data quality of core items improved from 2017 to 2021. In particular, data quality relating to CFTRm use improved from ad hoc free text collection in 2017 to include start and stop dates and reasons for stopping for patients in 2021. Additionally, genotyping data completeness to 98.4% of participants with at least one known allele. The enhanced data quality enabled credible analyses of the change in ppFEV 1 for patients initiated on LUM/IVA to be conducted (Table 1). These analyses demonstrated maintenance of ppFEV 1 following LUM/IVA initiation. Conclusion: ACFDR data quality improved during implementation of the collaboration; specifically, this enabled analysis of the effect of LUM/IVA on ppFEV 1 over time. This is an example of successful collaboration between a clinical registry and the pharmaceutical industry to support real-world data reporting to government agencies.