Data Quality- and Utility-compliant Anonymization of Electronic Health Record Data in the context of Multiple Common Data Models and Research Data Standard: Protocol for a Scoping Review (Preprint)

BACKGROUND The anonymization of EHR data is essential to ensure privacy protection in secondary use scenarios. For interoperability reasons, a range of data-driven medical research projects adopted Common Data Models (CDMs) which represent data in a quality- and utility-compliant way suitable for research. Few reviews investigate CDM-based implementations of formal data anonymization processes with a reflection of data quality and –utility issues. OBJECTIVE This scoping review will investigate the state-of-the-art regarding how formal data anonymization processes are applied on medical research CDMs and data representation standards, and to what extent strategies or gaps in dealing with quality problems of resulting anonymized datasets are observed. METHODS In developing the protocol for this review, we used the framework of Arksey and O'Malley. Based on this, we will include only articles published in English and available through the databases PubMed and Web of Sciences. The literature search will be based on a query syntax validated by a librarian, and accompanied by manual queries to include further informal sources. Eligible references will undergo a de-duplication step, followed by a screening of papers titles and abstracts. In a second phase, a full-text reading will allow the final selection of the corresponding articles, while a domain expert will support resolving literature selection conflicts. During this phase, key information will be extracted, categorized, summarized, and analyzed with reference to a template-based structure. Tabulated and graphical analyses will be included in alignment with the PRISMA-ScR model. We also performed some tentative searches on the both target literature databases for estimating the retrievability of eligible papers. RESULTS The tentative searches of the PubMed and Web of Sciences databases resulted in 119 and 296 de-duplicated matches respectively, suggesting the availability of (potentially) relevant articles. Further analysis and selection steps will allow reaching a final literature set. The completion of this scoping review project is foreseen to take place by the end of the second quarter of 2023. CONCLUSIONS Outlining approaches to deploy formal data anonymization on CDMs with a consideration of potentially associated data quality and utility issues will provide useful insights to increase the preservation of fitness-for-use of anonymized data in the scientific usage of CDMs. This protocol describes a schedule to perform a scoping review, to address the existing evidence and challenges, which will support the conduction of follow-up investigations.