Abstract
PurposeDisproportionality screening analysis is acknowledged as a tool for performing signal detection in databases of adverse drug reactions (ADRs), e.g., in the European Union (EU) Drug Authority setting. The purpose of this study was to explore the possibility of decreasing false-positive signals of disproportionate reporting (SDR) by calculating the proportional reporting ratio (PRR)-by-therapeutic area (TA), while still maintaining the ability to detect relevant SDRs.MethodsIn the EudraVigilance (EV) Database, output from PRR calculated with a restricted TA comparator background was compared in detail to output from conventional authority-setting PRR calculations for four drugs: bicalutamide, abiraterone, metformin, and vildagliptin, within the TAs of prostate gland disease and type 2 diabetes mellitus.ResultsADR reports per investigated drug ranged from 2,400 to 50,000. The PRR-TA’s ability to detect true-positive SDRs (as acknowledged in approved labeling) was increased compared to the conventional PRR, and performed 8–31 % better than a recently proposed stricter EU-SDR definition. The PRR-TA removed false SDRs confounded by disease or disease spill-over by up to 63 %, while retaining/increasing the number of unclassified SDRs relevant for manual validation, and thereby improving the ratio between confounded SDRs (i.e., noise) and unclassified SDRs for all investigated drugs (possible signals).ConclusionsThe performance of the PRR was improved by background restriction with the PRR-TA method; the number of false-positive SDRs decreased, and the ability to detect true-positive SDRs increased, improving the signal-to-noise ratio. Further development and validation of the method is needed within other TAs and databases, and for disproportionality analysis methods.Electronic supplementary materialThe online version of this article (doi:10.1007/s00228-014-1658-1) contains supplementary material, which is available to authorized users.
Highlights
Screening for statistical disproportionality in databases of adverse drug reactions (ADRs) is an accepted means for signal detection
Relative frequencies of signals of disproportionate reporting (SDR) using 3 as the case count threshold (SDR3) among all reported ADR terms ranged from 10 % for bicalutamide (i.e. 95/950) to 17.9 % for vildagliptin (Supplementary Table 2)' the rest, 82–90 %, were thereby excluded from clinical evaluation
Increasing the SDRdefining case count to ≥5 (SDR5) reduced the number of SDRs for further validation and verification by between 14 % in the abiraterone analysis and 36 % for vildagliptin, removing between 33–70 % of unclassified SDRs, potentially delaying detection and validation of important signals (Supplementary Table 2)
Summary
Screening for statistical disproportionality in databases of adverse drug reactions (ADRs) is an accepted means for signal detection. Several disproportionality methods are currently in use [1,2,3,4,5], but no gold standard has been established [6, 7] They all share the ability to detect drug safety signals years earlier. Within the European Union (EU), the EudraVigilance (EV) database [11] is continuously screened using the proportional reporting ratio (PRR) method [1, 3, 12] This method delivers signals of disproportionate reporting (SDRs)— reported ADR-drug combinations that (1) appear to a disproportionately high extent in a database (have an elevated PRR value), and (2) reach a case count above a pre-specified threshold
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
