Performance of a semi-automated approach for risk estimation using a common data model for longitudinal healthcare databases.

摘要: Different structures and coding schemes may limit rapid evaluation of a large pool potential drug safety signals using multiple longitudinal healthcare databases. To overcome this restriction, semi-automated approach utilising common data model (CDM) robust pharmacoepidemiologic methods was developed; however, its performance needed to be evaluated. Twenty-three established drug-safety associations from publications were reproduced in claims database four these also repeated electronic health records. Concordance discrepancy pairwise estimates assessed between the results derived publication approach. For all 27 pairs, an observed agreement published greater than 85% Kappa coefficient 0.61, 95% CI: 0.19-1.00. Ln(IRR) differed by less 50% for 13/27 IRR varied 2-fold 19/27 pairs. Reproducibility based on intra-class correlation 0.54. Most covariates (>90%) available inclusion models. Once study populations inclusion/exclusion criteria obtained literature, analysis able completed 2-8 h. The methodology CDM produced consistent risk compared findings most selected drug-outcome associations, regardless original designs, databases, medications outcomes. Further assessment is useful understand roles, strengths limitations rapidly evaluating signals.