OBJECTIVES : In Japan, active efforts are being made to utilize the database for post-marketing surveillance. We are considering using an electronic medical record-based medical database as a feasibility pre-review tool to consider appropriate post-marketing surveillance designs. The purpose of this study was to evaluate the validity of diagnostic coding by secondary use of IQVIA hospital information system data.

METHODS : Electronic medical record of three hospitals from IQVIA HIS database was used. The target study period is from Nov 2018 to Nov 2019, and the target outcome setting is dyslipidemia. The new diagnosis was evaluated as a potential case, and only the laboratory values were evaluated as the gold standard for validity as the real case. The clinical test values of the potential cases were confirmed, and whether the cases were true cases was determined according to the criteria, and the positive predictive value (PPV) of each outcome definition was calculated.

RESULTS : The potential case was 13,157 in total, and the true case was 9479. The PPV for the definition of a new occurrence of dyslipidemia was 72%. The facility with the lowest PPV was 65% and the facility with the highest PPV was 81%.

CONCLUSIONS : In this pilot study, only laboratory values were used to evaluate the validity of the outcome definition. Although this method is a simpler method for confirming the validity of the outcome definition, it is important to thoroughly examine beforehand whether the clinical test value alone can be used as the gold standard. The use of IQVIA HIS could be a pre-examination tool for a post-marketing surveillance designing.