The ConcePTION ecosystem uses real-world data to generate real-world evidence about medicine safety for women of childbearing age, women who are pregnant or breastfeeding, and their children. Researchers in the project have developed an open-source tool to assess the quality of real-world health data. By evaluating the quality of health data obtained outside of clinical trials, INSIGHT can help assess whether data sources can be used reliably for the generation of evidence about outcomes.
The INSIGHT tool is meant to check if data collected in the real world is good enough to include in research. It does these checks in three stages: first, it looks at whether the data is formatted correctly, then it checks if the timing of events makes sense and if each piece of data is unique, and finally, it gives an overall view of data patterns and if there are any strange points or trends over time, using a large variety of different tests (over 588 of them) to see if the data meets certain standards.
The INSIGHT quality checks of real-world data align perfectly with the real-world data quality frameworks set forth by the European Medicines Agency (EMA) emphasising five crucial dimensions of data quality: reliability, extensiveness, coherence, timeliness and relevance. This alignment underscores INSIGHTS’s robustness and its potential to contribute significantly to the field of real-world health data analysis.
Assessing data quality is not a simple checkbox exercise. It involves understanding both internal and external factors that may influence the quality of data, which is collected as a product of routine care and not for research purposes.
“Assessing data quality is essential, particularly for Real World Evidence analysis, where accuracy is paramount for making informed decisions. With its adherence to industry standards and rigorous assessment processes, the INSIGHT tool that we have developed holds promise in enhancing confidence in generated evidence and potentially driving improvements in healthcare outcomes. Pregnancy pharmacovigilance is one of the areas where this is a pressing issue, making sure women can make informed decisions about their health and treatment,” says Vjola Hoxhaj, PhD student in epidemiology and senior scientific programmer from the Department of Data Science and Biostatics at the Julius Centre, University Medical Centre Utrecht, the Netherlands and one of the authors of the paper.
By Anna Holm Bodin & Märta Karlén
Hoxhaj V, Andaur Navarro CL, Riera-Arnau J, Elbers RHJH, Alsina E, Dodd C, Sturkenboom MJCM. INSIGHT: A Tool for Fit-for-Purpose Evaluation and Quality Assessment of Observational Data Sources for Real World Evidence on Medicine and Vaccine Safety, medRxiv, 2023.10.30.23297753; DOI: 10.1101/2023.10.30.23297753