Analysis of the application effect of new rehabilitation nursing methods in orthopedic postoperative rehabilitation: A systematic review and meta-analysis of randomized controlled trials

ObjectiveTo quantify the impact of technology-enabled rehabilitation nursing on patient-reported function after lower-limb arthroplasty and to explore effect modification by surgical procedure and by technological modality.MethodsTen databases were searched from inception to 2 May 2025. Randomised controlled trials (RCTs) comparing an innovative digital or electromechanical rehabilitation intervention with usual postoperative care and reporting WOMAC, KOOS or HOOS outcomes were eligible. Risk of bias was assessed with Cochrane RoB 2.0. Standardised mean differences (Hedges g) were pooled using a Hartung-Knapp REML random-effects model; heterogeneity was quantified with I². Sub-group analyses were prespecified for surgery type (TKA vs THA) and technology class (virtual reality VR, web/app telerehabilitation WB, robot/sensor RB). Publication bias was evaluated with funnel-plot inspection and Egger's regression. The certainty of evidence was assessed with the GRADE framework.ResultsFifteen RCTs (1012 experimental, 954 control participants; 11 TKA, 3 THA, 1 mixed) met the criteria and were all rated overall "low risk" by RoB 2.0. Across trials, technology-enabled care conferred a small but significant improvement in patient-reported function (g = 0.28; 95% CI 0.00 to 0.56; p = 0.049; I² = 86%). VR produced the largest point estimate (g = 0.62; 95% CI -0.18 to 1.41; 4 trials); WB yielded a modest, non-significant benefit (g = 0.18; 95% CI -0.24 to 0.59; 8 trials); RB showed a comparable, non-significant effect (g = 0.14; 95% CI -0.23 to 0.50; 3 trials). The χ² test for subgroup differences was not significant (p = 0.16). Egger's test revealed no evidence of small-study effects (p = 0.73). Leave-one-out and influence analyses confirmed robustness of the pooled estimate. The certainty of evidence was rated as moderate (GRADE).ConclusionsNext-generation digital and electromechanical rehabilitation programmes achieve at least non-inferior- and potentially clinically relevant-improvements in self-reported function after lower-limb arthroplasty while reducing in-person therapist time. Virtual-reality platforms appear most promising, but heterogeneity suggests that dose, feedback fidelity and sensor precision are key effect drivers. Large, standardised multicentre trials with cost-utility endpoints are needed to clarify which technological components add value for which patients.