Performance of the Self-Controlled Case Series With Active Comparators for Drug Safety Signal Detection Using the...

There is little evidence about signal detection using UK primary care electronic health records (EHRs). The self controlled case series (SCCS) is one of the most promising methods for drug safety signal detection using real world data, and incorporating active comparators could potentially improve its performance by addressing confounding by indication.

This study aims to evaluate the performance of the SCCS with and without active comparators for signal detection using the UK Clinical Practice Research Datalink (CPRD) Aurum.

We applied the SCCS to macrolide and fluoroquinolone antibiotics, using amoxicillin and cefalexin as active comparators. In total seven drugs, and 30 outcomes from all organ classes were selected. We developed a reference set of 104 positive controls and 58 negative controls, using a taxonomy framework to ensure the selected drug outcome pairs are theoretically well suited to the SCCS design. Two-year observation periods with a 30-day risk window after each dispensing were used. Diagnostic performance was measured using sensitivity and specificity with respect to the product labels.

The sensitivity and specificity of the SCCS without active comparator in the 2017/2018 observation period were 0.57 and 0.77 when limited to pairs with satisfactory power. Specificity increased up to 0.89 with active comparators, however sensitivity decreased to 0.18. Five drug-outcome pairs were signals of disproportionality before they were present on labels.

Using a carefully designed reference set of drug-outcome pairs well suited to the study design, the SCCS performed moderately well for signal detection in CPRD. Whilst active comparators effectively reduced confounding by indication, they also reduced the number of correctly identified positive controls, due to a reduction in power. We showed some evidence that SCCS is able to highlight SDRs before they were present on labels.