The CLARION study, a multi-country, non-interventional cohort study, has implemented a comprehensive approach to maintain data quality across multiple sclerosis (MS) registries and data sources. This effort supports the long-term safety monitoring of cladribine tablets, a treatment for relapsing-remitting MS. By defining and assessing 28 data quality indicators (DQIs), the study ensures the reliability and consistency of real-world data used for regulatory decision-making. As of end-June 2024, an estimated 101,132 people have received cladribine tablets for the treatment of MS with 251,900 cumulative years of exposure.
Development and Implementation of Data Quality Indicators
In collaboration with participating MS registries, principal investigators, and the marketing authorization holder, DQIs were defined to assess representativeness, consistency, accuracy, and completeness of data. These categories align with the European Medicines Agency (EMA) Data Quality Framework. Remedial actions were mapped to address any data quality issues identified. The DQIs were analyzed at the start of each registry's participation and annually thereafter.
Data Sources and Patient Population
The study included data from eight MS registries/data sources across 14 countries, including the MS Documentation System 3D (Germany), Danish MS Registry, Finnish MS Registry, MS Database (MSBase), Norwegian MS Registry and Biobank (NMSRB), Swedish MS Registry, Swiss MS Cohort, and US Department of Defense (DoD). Data from 5,069 patients with confirmed eligibility were analyzed, with 2,958 in the cladribine cohort and 2,111 in the fingolimod cohort, using November 1, 2022, as the data cut-off date.
Key Findings on Data Quality
- Representativeness: The distribution of key patient characteristics, such as sex and age at MS onset, was assessed. Overall, 72.0% of patients were female. The median age at MS onset ranged from 28.0 to 32.0 years. Relapsing-remitting MS was the predominant clinical course (93.5%).
- Consistency: At least one year of follow-up was available for 4,137 patients. However, 20.8% of these patients had no recorded visit during that year. The average number of recorded lymphocyte count measurements per patient per year in the cladribine cohort was 1.0.
- Accuracy: Discrepant MS onset dates were detected for 1.1% of patients, while discrepant MS diagnosis dates were detected for 0.6%. These discrepancies decreased over time. 1.6% of patients had at least one instance of duplicate MS treatment recordings.
- Completeness: There were no missing MS treatment start dates. However, 43.9% of adverse events of special interest (AESI) records were incomplete. 3.8% of patients were classified as having been lost to follow-up.
Significance of Data Harmonization
The CLARION study demonstrates the importance of harmonizing data quality across different MS registries. By using predefined DQIs, the study identified potential data issues early on, allowing for correction and alignment before statistical analyses. This approach enhances the reliability of safety evaluations and supports regulatory decision-making. The study's findings align with the FDA's guidance on real-world data, emphasizing completeness, consistency, and accuracy.
Limitations
Not all DQIs could be evaluated in all MS registries, which affected the interpretation of findings. Definitions for cladribine tablets stop dates varied between MS registries. Data absence and inconsistencies may have originated from data manipulation at each registry.
Conclusion
The CLARION study highlights the value of systematic DQI evaluation for harmonizing data quality and consistency across MS registries. The contributing MS registries offer data quality for several metrics essential for post-approval safety studies, supporting the use of real-world data in regulatory decision-making.
