Our summary
Real-world data (RWD), including sources such as electronic health records (EHRs) and billing claims, is increasingly being leveraged to generate valuable real-world evidence (RWE) across the oncology landscape. As use of RWD/E has expanded, so has regulatory guidance around its use, and it is clear that data quality matters. It is also clear that quality is not a single concept, and that multiple dimensions must be addressed in order to determine whether RWD is fit for use.
Assessing quality in EHR-based RWD is challenging for many reasons, including the variety of formats, fragmentation of patient information, and unstructured data sources that require specialized processes for curation. In this paper, researchers summarize the key dimensions of data quality that should be addressed when making a fit-for-use assessment. They then demonstrate how clinically informed processes and scientific methods can be used to comprehensively address data quality for EHR-derived data, including considerations in balancing feasibility, robustness, and scalability.
Why this matters
Development of high-quality, scaled EHR-based RWD requires integration of systematic processes assessing quality across the data lifecycle. This paper shows clinically informed processes, assessments, and scientific methods can be used to generate RWD. This framework and approaches to data quality can be applied by other initiatives scaling real-world data, or in assessing the quality of different real-world data sources.