Data accuracy, consistency and completeness of the national Swiss cystic fibrosis patient registry: Lessons from an ECFSPR data quality project.

Abstract

Good data quality is essential when rare disease registries are used as a data source for pharmacovigilance studies. This study investigated data quality of the Swiss cystic fibrosis (CF) registry in the frame of a European Cystic Fibrosis Society Patient Registry (ECFSPR) project aiming to implement measures to increase data reliability for registry-based research. All 20 pediatric and adult Swiss CF centers participated in a data quality audit between 2018 and 2020, and in a re-audit in 2022. Accuracy, consistency and completeness of variables and definitions were evaluated, and missing source data and informed consents (ICs) were assessed. The first audit included 601 out of 997 Swiss people with CF (60.3%). Data quality, as defined by data correctness ≥95%, was high for most of the variables. Inconsistencies of specific variables were observed because of an incorrect application of the variable definition. The proportion of missing data was low with <5% for almost all variables. A considerable number of missing source data occurred for CFTR variants. Availability of ICs varied largely between centers (10 centers had >5% of missing documents). After providing feedback to the centers, availability of genetic source data and ICs improved. Data audits demonstrated an overall good data quality in the Swiss CF registry. Specific measures such as support of the participating sites, training of data managers and centralized data collection should be implemented in rare disease registries to optimize data quality and provide robust data for registry-based scientific research.