Martijn J Schuemie1, David Madigan, Patrick B Ryan. 1. Department of Medical Informatics, Erasmus University Medical Center Rotterdam, P.O. Box 2040, 3000 CA, Rotterdam, The Netherlands, m.schuemie@erasmusmc.nl.
Abstract
BACKGROUND: The availability of large-scale observational healthcare data allows for the active monitoring of safety of drugs, but research is needed to determine which statistical methods are best suited for this task. Recently, the Longitudinal Gamma Poisson Shrinker (LGPS) and Longitudinal Evaluation of Observational Profiles of Adverse events Related to Drugs (LEOPARD) methods were developed specifically for this task. LGPS applies Bayesian shrinkage to an estimated incidence rate ratio, and LEOPARD aims to detect and discard associations due to protopathic bias. The operating characteristics of these methods still need to be determined. OBJECTIVE: Establish the operating characteristics of LGPS and LEOPARD for large scale observational analysis in drug safety. RESEARCH DESIGN: We empirically evaluated LGPS and LEOPARD in five real observational healthcare databases and six simulated datasets. We retrospectively studied the predictive accuracy of the methods when applied to a collection of 165 positive control and 234 negative control drug-outcome pairs across four outcomes: acute liver injury, acute myocardial infarction, acute kidney injury, and upper gastrointestinal bleeding. RESULTS: In contrast to earlier findings, we found that LGPS and LEOPARD provide weak discrimination between positive and negative controls, although the use of LEOPARD does lead to higher performance in this respect. Furthermore, the methods produce biased estimates and confidence intervals that have poor coverage properties. CONCLUSIONS: For the four outcomes we examined, LGPS and LEOPARD may not be the designs of choice for risk identification.
BACKGROUND: The availability of large-scale observational healthcare data allows for the active monitoring of safety of drugs, but research is needed to determine which statistical methods are best suited for this task. Recently, the Longitudinal Gamma Poisson Shrinker (LGPS) and Longitudinal Evaluation of Observational Profiles of Adverse events Related to Drugs (LEOPARD) methods were developed specifically for this task. LGPS applies Bayesian shrinkage to an estimated incidence rate ratio, and LEOPARD aims to detect and discard associations due to protopathic bias. The operating characteristics of these methods still need to be determined. OBJECTIVE: Establish the operating characteristics of LGPS and LEOPARD for large scale observational analysis in drug safety. RESEARCH DESIGN: We empirically evaluated LGPS and LEOPARD in five real observational healthcare databases and six simulated datasets. We retrospectively studied the predictive accuracy of the methods when applied to a collection of 165 positive control and 234 negative control drug-outcome pairs across four outcomes: acute liver injury, acute myocardial infarction, acute kidney injury, and upper gastrointestinal bleeding. RESULTS: In contrast to earlier findings, we found that LGPS and LEOPARD provide weak discrimination between positive and negative controls, although the use of LEOPARD does lead to higher performance in this respect. Furthermore, the methods produce biased estimates and confidence intervals that have poor coverage properties. CONCLUSIONS: For the four outcomes we examined, LGPS and LEOPARD may not be the designs of choice for risk identification.
Authors: R J Gralla; D Osoba; M G Kris; P Kirkbride; P J Hesketh; L W Chinnery; R Clark-Snow; D P Gill; S Groshen; S Grunberg; J M Koeller; G R Morrow; E A Perez; J H Silber; D G Pfister Journal: J Clin Oncol Date: 1999-09 Impact factor: 44.544
Authors: Patrick B Ryan; Martijn J Schuemie; Emily Welebob; Jon Duke; Sarah Valentine; Abraham G Hartzema Journal: Drug Saf Date: 2013-10 Impact factor: 5.606
Authors: Preciosa M Coloma; Paul Avillach; Francesco Salvo; Martijn J Schuemie; Carmen Ferrajolo; Antoine Pariente; Annie Fourrier-Réglat; Mariam Molokhia; Vaishali Patadia; Johan van der Lei; Miriam Sturkenboom; Gianluca Trifirò Journal: Drug Saf Date: 2013-01 Impact factor: 5.606
Authors: Martijn J Schuemie; M Soledad Cepeda; Marc A Suchard; Jianxiao Yang; Yuxi Tian; Alejandro Schuler; Patrick B Ryan; David Madigan; George Hripcsak Journal: Harv Data Sci Rev Date: 2020-01-31
Authors: Martijn J Schuemie; Rosa Gini; Preciosa M Coloma; Huub Straatman; Ron M C Herings; Lars Pedersen; Francesco Innocenti; Giampiero Mazzaglia; Gino Picelli; Johan van der Lei; Miriam C J M Sturkenboom Journal: Drug Saf Date: 2013-10 Impact factor: 5.606
Authors: Richard D Boyce; Patrick B Ryan; G Niklas Norén; Martijn J Schuemie; Christian Reich; Jon Duke; Nicholas P Tatonetti; Gianluca Trifirò; Rave Harpaz; J Marc Overhage; Abraham G Hartzema; Mark Khayter; Erica A Voss; Christophe G Lambert; Vojtech Huser; Michel Dumontier Journal: Drug Saf Date: 2014-08 Impact factor: 5.606
Authors: S Vilar; P B Ryan; D Madigan; P E Stang; M J Schuemie; C Friedman; N P Tatonetti; G Hripcsak Journal: CPT Pharmacometrics Syst Pharmacol Date: 2014-09-24