Relative risks from an unadjusted new-user cohort design



A data frame with 174 rows and 4 variables:


Name of the outcome


The log of the true effect size. Only provided for negative and positive controls, is NA for the outcome of interest (GI bleeding).


The log of the incidence rate ratio


The standard error of the log of the incidence rate ratio


A dataset containing the incidence rate ratios (and standard errors) produced using a new-user cohort design that compares new-users of dabigatran to new-users of warfarin for the outcome of GI hemorrhage. The dataset includes estimates both for the outcome of interest as well as negative and positive control outcomes. Subjects are required to have 183 days of continuous observation prior to initiating treatment, a prior diagnosis of atrial fibrillation, and are required to have no prior exposure to either dabigatran or warfarin. The study computes an incidence rate ratio without any adjustment for confounders. Time at risk is defined as the time on the drug. The original study (Southworth 2013) uses the 'Mini-Sentinel Database'. For our replication, we use the Optum databases since both databases are US insurance claims databases. We analyzed 5,982 dabigatran-exposed and 19,155 warfarin-exposed subjects. For more information on this set see Schuemie et al (2017).


Schuemie MJ, Hripcsak G, Ryan PB, Madigan D, Suchard MA. Empirical confidence interval calibration for population-level effect estimation studies in observational healthcare data. Proc Natl Acad Sci U S A. 2018 Mar 13;115(11):2571-2577

Southworth MR, Reichman ME, Unger EF. Dabigatran and postmarketing reports of bleeding. N Engl J Med 368(14):1272-1274, 2013