The Use of Active Comparators in Self-Controlled Designs

Area of Science:

• Pharmacoepidemiology
• Biostatistics
• Health Services Research

Background:

• Time-varying confounding by indication poses a challenge in self-controlled studies of medication effects.
• This bias arises when the reason for prescribing a medication changes over time.
• Active comparators offer a potential solution to mitigate this bias.

Purpose of the Study:

• To describe methods for using active comparators to address time-varying confounding by indication.
• To empirically demonstrate the utility of active comparators in pharmacoepidemiologic research.
• To investigate the association between penicillin and venous thromboembolism (VTE) using an active comparator.

Main Methods:

• Employed case-crossover, case-time-control, self-controlled case-series, and sequence symmetry analyses.
• Utilized Danish registry data (1996-2018) for the empirical example.
• Roxithromycin, a macrolide antibiotic, served as the active comparator for penicillin.

Main Results:

• Initial analyses showed a strong association between penicillin/roxithromycin and VTE, indicating confounding by indication.
• Adjusted analyses using roxithromycin as a comparator yielded odds ratios near 1 for penicillin (0.94-1.03).
• Self-controlled case-series and sequence symmetry analyses suggested a spurious protective effect of penicillin.

Conclusions:

• Time-varying confounding by indication is a significant issue in self-controlled medication effect studies.
• Active comparators, such as roxithromycin, can effectively mitigate this bias.
• The choice of analytical method is crucial, as some may still yield biased results.