Yoeri Bemelmans

Safety and efficacy of outpatient hip and knee arthroplasty 115 7 Discussion Current literature is divided on the safety (in terms of complications and readmissions) of OJA pathways. Several studies reported an increased risk for (S)AEs if patients were discharged on the same day of surgery [32, 39, 50], while others stated that there were no safety issues [9, 24, 29, 34, 53]. The most important finding of this study was that the implementation of OJA pathways in a selective group of patients resulted in acceptable clinical outcomes regarding complications and readmission rates while reducing costs and preserving the patient-reported outcome compared to standard inpatient pathways. Subanalyses by type of arthroplasty (e.g. THA, TKA, UKA) found that THAs in OS pathways were associated with fewer AEs compared to inpatient pathways. For SAEs and readmissions, subanalyses found no significant differences between outpatient and inpatient pathways, suggesting that OJA following hip or knee arthroplasty ismost likely safe in a selected patient population. Thecurrent literatureonOJA, however, consistsof lowtomoderatequalitywithmoderate to high risk of selection bias. Only one RCT was included in the present systematic review. Most included studies, however, had a retrospective and observational design, with high risk of selection bias [5, 6, 20, 21, 25,26,27, 29, 30, 35,36,37,38,39,40,41, 43, 44, 47, 51,52,53]. Many studies on outpatient joint arthroplasty used large (national claims and administrative) databases [2,3,4, 19, 21, 28, 31,32,33, 39, 40, 42, 43, 46, 47, 49,50,51]. These databases provide high power to outcome measures and although results can provide a raw estimate of certain relationships, heterogeneity in the sample (as variations in practice patterns are not accounted for can lead to potential recording bias and/or confounding) may worsen the accuracy that comes with interpreting the results. Bonvonratwet et al. [4] also found that definitional differences in “outpatient” status were present in the American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP®) – a national (USA) database which is used in many studies on OJA [4]. The same study also described the possible inf luence of definitional differences on outcomes. An explanation for this could be a more stringent patient selection for patients discharged on the day of surgery compared to patients discharged within 24 h (i.e. the first postoperative day). Besides, there is substantial potential of confounding by selection bias in these database and observational studies, as OJA is often reserved for the more “active” and “healthy” patients. As expected, we found that patients included in the OJA pathways were overall younger and less infirm (in terms of ASA class), and had a lower BMI compared to patients in the inpatient pathways [20, 24, 28, 34, 43, 45, 48]. This should