Preliminary assessment on data appropriateness of electronic health records for real world studies: a survey on hospitals in a developing region in China (Preprint)

Résumé fourni par la source

BACKGROUND Assessment of data appropriateness is a top priority before conducting real world studies (RWS). However, this evaluation on electronic health records (EHRs) from hospitals in developing regions has been long-term overlooked. OBJECTIVE This study aimed to preliminarily assess whether EHRs from hospitals in a developing region in China were appropriate enough to meet RWS purposes. METHODS All secondary and tertiary hospitals (n = 28) in Fuzhou, Jiangxi province, China, were recruited. Through face-to-face interviews with chief administrators of hospital EHR systems, the assessment was conducted via semi-quantitative questionnaire to evaluate three dimensions of EHR appropriateness (coverage, consistency, accessibility). Coverage and consistency were measured using 80 core variables compared with two well-established common data models (CDMs), the former focusing on the existence of relevant variables in targeted databases and the latter on data structuring, data types and standardized coding systems. Accessibility was measured by willingness to data sharing. RESULTS Structured EHRs restoration was accomplished in at least 15 hospitals. Variables on drug exposure were the least covered (of 75.0% databases), structured (of 53.6% databases), relatively inconsistent in data types (of 10.7-75.0% databases) and standardized coding systems (of 7.1% databases). Low willingness to data sharing was responded among 64.3% of local hospitals, over 50.0% due to administrative concerns. CONCLUSIONS EHR databases from hospitals in the developing region showed favorable coverage of potentially relevant variables for RWS. However, poor consistency, including data structuring, data type and data coding on the preliminary level of variables, and low willingness to share data challenged RWD application.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.