Use of a Composite Metric to Assess Hospital Performance after Hip Fracture

To the Editor: I read with great interest the paper by Elkassabany et al.1 that identifies hospital-related predictors of 30-day readmission and mortality following hip fracture surgery in Medicare beneficiaries. The authors conclude that higher volume hospitals have a lower risk of readmission or death. However, the magnitude of this effect differs considerably between the two outcomes, with an adjusted odds ratio of 0.83 for death but 0.94 for readmission, with the latter failing to reach statistical significance. Furthermore, facility size and hospital teaching status had divergent effects on the two outcome measures. There is increasing interest in defining the best metric(s) to gauge hospital performance, and 30-day readmission rate is currently used by the Centers for Medicare and Medicaid for reimbursement following hip and knee arthroplasty.2 This study highlights that the effect of a particular hospital-related factor is variable depending upon the outcome measure being assessed, and thus the use of composite outcome measures to judge performance could be misleading. Indeed, many potentially modifiable hospital-related factors that affect mortality after hip fracture surgery, such as time to surgery, have been shown to have little effect on readmission risk; in contrast, patient co-morbidities that are more difficult for hospitals to mitigate against are highly significant predictors of readmission.3-5 Taking mortality and readmission as a composite measure may therefore both miss potential opportunities to mitigate against risk and lead to an unfair assessment of hospital performance. As the drive to measure and report hospital performance intensifies, it is imperative that we recognise that whilst outcome measures complement one another, each is telling us something different about the quality of care being delivered. Conflict of Interest: I have no conflicts of interest to declare. Financial Disclosure: No funding was required. Author contributions: Adam M Ali was the sole author. Sponsors: No external funding or support was received in preparing this manuscript.