Abstract
For self-controlled studies of medication-related effects, time-varying confounding by indication can occur if the indication varies over time. We describe how active comparators might mitigate such bias, using an empirical example. Approaches to using active comparators are described for case-crossover design, case-time-control design, self-controlled case-series, and sequence symmetry analyses. In the empirical example, we used Danish data from 1996-2018 to study the association between penicillin and venous thromboembolism (VTE), using roxithromycin, a macrolide antibiotic, as comparator. Upper respiratory infection is a transient risk factor for VTE, thus representing time-dependent confounding by indication. Odds ratios for case-crossover analysis were 3.35 (95% confidence interval: 3.23, 3.49) for penicillin and 3.56 (95% confidence interval: 3.30, 3.83) for roxithromycin. We used a Wald-based method or an interaction term to estimate the odds ratio for penicillin with roxithromycin as comparator. These 2 estimates were 0.94 (95% confidence interval: 0.87, 1.03) and 1.03 (95% confidence interval: 0.95, 1.13). Results were similar for the case-time-control analysis, but both the self-controlled case-series and sequence symmetry analysis suggested a weak protective effect of penicillin, seemingly explained by VTE affecting future exposure exclusively for penicillin. The strong association of antibiotics with VTE suggests presence of confounding by indication. Such confounding can be mitigated by using an active comparator.
Original language | English |
---|---|
Pages (from-to) | 2181-2187 |
Number of pages | 7 |
Journal | American Journal of Epidemiology |
Volume | 190 |
Issue number | 10 |
Early online date | 16 Apr 2021 |
DOIs | |
Publication status | Published - 1 Oct 2021 |
Bibliographical note
Funding Information: No funding information.Open Access: No Open Access licence
Publisher Copyright: © The Author(s) 2021. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
Citation: Jesper Hallas, Heather Whitaker, Joseph A Delaney, Suzanne M Cadarette, Nicole Pratt, Malcolm Maclure, The Use of Active Comparators in Self-Controlled Designs, American Journal of Epidemiology, Volume 190, Issue 10, October 2021, Pages 2181–2187.
DOI: https://doi.org/10.1093/aje/kwab110
Keywords
- active comparators
- case-only designs
- confounding-by-indication