De-identified data quality assessment approaches by data vendors who license data to healthcare and life sciences researchers

Objective To gain insights into how data vendor companies (DVs), an important source of de-identified/anonymized licensed patient-related data (D/ALD) used in clinical informatics research in life sciences and the pharmaceutical industry, characterize, conduct, and communicate data quality assessments to researcher purchasers of D/ALD. Materials and Methods A qualitative study with interviews of DVs executives and decision-makers in data quality assessments (n = 12) and content analysis of interviews transcripts. Results Data quality, from the perspective of DVs, is characterized by how it is defined, validated, and processed. DVs identify data quality as the main contributor to successful collaborations with life sciences/pharmaceutical research partners. Data quality feedback from clients provides the basis for DVs reviews and inspections of quality processes. DVs value customer interactions, view collaboration, shared common goals, mutual expertise, and communication related to data quality as success factors. Conclusion Data quality evaluation practices are important. However, no uniform DVs industry standards for data quality assessment were identified. DVs describe their orientation to data quality evaluation as a direct result of not only the complex nature of data sources, but also of techniques, processes, and approaches used to construct data sets. Because real-world data (RWD), eg, patient data from electronic medical records, is used for real-world evidence (RWE) generation, the use of D/ALD will expand and require refinement. The focus on (and rigor in) data quality assessment (particularly in research necessary to make regulatory decisions) will require more structure, standards, and collaboration between DVs, life sciences/pharmaceutical, informaticists, and RWD/RWE policy-making stakeholders. Lay Summary De-identified/anonymized licensed patient-related data are types of real-world data purchased from data vendor companies and used by researchers in the life sciences and pharmaceutical industries to conduct studies to generate real-world evidence necessary to inform regulations. Real-world data quality is important. High-quality real-world data facilitates trust in results of real-world evidence-based research studies. This qualitative discovery study explored how data vendor companies assess the quality of de-identified/anonymized data licensed/sold to researchers. Data quality assessment practices exist amongst data vendor companies, howe