A rapid review of associations between provider volume and outcome of total knee arthroplasty. Where do the magical threshold values come from?

We set out to clarify whether in hospitals with a large volume morbidity and mortality rates after total knee arthroplasty (TKA) can be improved, whether the effects are consistent, and whether minimum recommendable caseloads can be inferred. We conducted a systematic review using MEDLINE, EMBASE, CENTRAL, and CINAHL and performed a hand search without restrictions on language or publication types. We identified 1406 citations, of which 13 studies including 1,110,962 patients met our inclusion criteria. Of these, six studies explored the same administrative data source. Five studies enrolling 448,897 were eligible for quantitative analysis. All studies corresponded to evidence level 2b (prospective or retrospective cohort study with >80% follow-up). We found homogeneous results about hospital mortality. Between 2551 and 821 TKA must be performed by high-volume rather than by low-volume providers to prevent 1 extra death. Absolute event rates are notably small.