We considered death to be a competing risk because it precludes revision surgery
and recurrence of the infection10 (link),11 (link). To consider
competing risks (death), we used competing risk models from the mstate package
in the R statistical software environment (R Foundation for Statistical
Computing)12 (link). To allow
comparison with the literature, we also report the Kaplan-Meier estimate,
although it is not valid in the setting of competing risks. Sensitivity analyses
were performed on patient and surgical factors using Cox regression. According
to the AQUILA (Assessment of Quality in Lower Limb Arthroplasty) checklist, we
considered results for the competing risk (Kaplan-Meier) estimate to be valid
when at least 20 hips remained in the analysis (i.e., were considered “at
risk”)13 (link),14 (link).
and recurrence of the infection10 (link),11 (link). To consider
competing risks (death), we used competing risk models from the mstate package
in the R statistical software environment (R Foundation for Statistical
Computing)12 (link). To allow
comparison with the literature, we also report the Kaplan-Meier estimate,
although it is not valid in the setting of competing risks. Sensitivity analyses
were performed on patient and surgical factors using Cox regression. According
to the AQUILA (Assessment of Quality in Lower Limb Arthroplasty) checklist, we
considered results for the competing risk (Kaplan-Meier) estimate to be valid
when at least 20 hips remained in the analysis (i.e., were considered “at
risk”)13 (link),14 (link).
