Harmonising electronic health records for reproducible research: challenges, solutions and recommendations from a UK-wide COVID-19 research collaboration

Abstract

Background: The CVD-COVID-UK consortium was formed to understand the relationship between COVID-19 and cardiovascular diseases through analyses of harmonised electronic health records (EHRs) across the four UK nations. Beyond COVID-19, data harmonisation and common approaches enable analysis within and across independent Trusted Research Environments. Here we describe the reproducible harmonisation method developed using large-scale EHRs in Wales to accommodate the fast and efficient implementation of cross-nation analysis in England and Wales as part of the CVD-COVID-UK programme. We characterise current challenges and share lessons learnt.

Methods: Serving the scope and scalability of multiple study protocols, we used linked, anonymised individual-level EHR, demographic and administrative data held within the SAIL Databank for the population of Wales. The harmonisation method was implemented as a four-layer reproducible process, starting from raw data in the first layer. Then each of the layers two to four is framed by, but not limited to, the characterised challenges and lessons learnt. We achieved curated data as part of our second layer, followed by extracting phenotyped data in the third layer. We captured any project-specific requirements in the fourth layer.

Results: Using the implemented four-layer harmonisation method, we retrieved approximately 100 health-related variables for the 3.2 million individuals in Wales, which are harmonised with corresponding variables for > 56 million individuals in England. We processed 13 data sources into the first layer of our harmonisation method: five of these are updated daily or weekly, and the rest at various frequencies providing sufficient data flow updates for frequent capturing of up-to-date demographic, administrative and clinical information.

Conclusions: We implemented an efficient, transparent, scalable, and reproducible harmonisation method that enables multi-nation collaborative research. With a current focus on COVID-19 and its relationship with cardiovascular outcomes, the harmonised data has supported a wide range of research activities across the UK.

Keywords: COVID-19; Common data model; Data harmonisation; Electronic health record; NHS digital TRE for England; Population health; Reproducible research; SAIL databank; Trusted Research Environments.

Fig. 1

A simplified example of the four-layer data preparation process used to harmonise data within SAIL with data for England (within the NHS Digital TRE for England). Layer 1 consists of raw data sources in SAIL (e.g., primary care and secondary care data sources). Layer 2 includes Research Ready Data Assets (RRDAs) and generated curated version of raw data sources. Examples of RRDAs are the COVID-19 C20 cohort, combined mortality data for COVID-19 C20 cohort [47] and RRDA version of dispensing data [45]. In Layer 3, phenotypes related data are generated using Layer 2 data and phenotype code-lists. Many phenotype code-lists in the HDR UK Phenotype Library [12] have already been imported into SAIL (only a subset of phenotypes has been displayed for illustrative purposes). Finally, in Layer 4 fully harmonised project-specific data tables are derived from Layer 2 and 3 data

Fig. 2

The four-layer data harmonisation process for Welsh data analysis used in [43]. See the GitHub repository [62] for the scripts

Fig. 3

The process for combining analyses results from SAIL (Wales) and NHS Digital TRE for England. The SAIL Databank for Wales and the NHS Digital TRE for England provide a secure remote data access system and analysis environment. Many phenotype code-lists in the HDR UK Phenotype Library [12] have already been uploaded/imported in these TREs. Approved researchers within each TRE can access data and phenotype code-lists, and perform analyses in the TRE. Then the results of analyses from these TREs can be combined (using meta-analysis) outside of the TREs once approved through the TREs disclosure control process, with phenotype code-lists and code accessible outside the TREs, and a copy of what is needed imported/exported from the TREs as required through standard disclosure control processes

Fig. 4

Categories of harmonised variables for population of Wales and England