Effects of digital health interventions in total knee arthroplasty: a systematic review and meta-analysis

Postoperative rehabilitation is essential for recovery after total knee arthroplasty (TKA), but conventional rehabilitation may be limited by unequal access and insufficient ongoing supervision. The effectiveness of digital rehabilitation relative to conventional rehabilitation remains uncertain. The primary objective was to estimate the pooled effects of digital health interventions on postoperative rehabilitation outcomes after TKA. Post hoc exploratory analyses examined potential variation according to dominant delivery modality, follow-up duration, and geographic region. PubMed, Embase, CINAHL, Web of Science, and the Cochrane Library were searched from inception to 17 December 2025, with an updated search conducted on 19 July 2026. ClinicalTrials.gov, the World Health Organization International Clinical Trials Registry Platform, and ProQuest were also searched. Randomized controlled trials comparing digital health interventions with usual care or conventional rehabilitation were included. Primary outcomes were pain, WOMAC total score, knee flexion and extension range of motion, and Timed Up and Go (TUG) performance. Risk of bias and certainty of evidence were assessed using RoB 2 and GRADE, respectively. Pooled effects were estimated using random-effects models. Subgroup and univariable meta-regression analyses were conducted as post hoc exploratory analyses. Thirty-six randomized controlled trials involving 3,811 participants were included. Statistically significant pooled SMDs favored digital health interventions for WOMAC total score (SMD = − 0.33, 95% CI − 0.59 to − 0.06) and knee flexion range of motion (SMD = 0.30, 95% CI 0.12 to 0.48), both supported by low-certainty evidence. The clinical importance of these effects remains uncertain because the pooled SMDs were not translated into absolute changes or evaluated against common clinically important thresholds. No statistically significant pooled differences were detected for pain, knee extension range of motion, or TUG performance. Substantial heterogeneity across interventions and study-specific comparator conditions limited the generalizability of the pooled estimates. Pooled estimates statistically favored digital health interventions for WOMAC total score and knee flexion after TKA, but the clinical importance of these findings remains uncertain. No statistically significant pooled differences were detected for pain, knee extension, or TUG performance. Certainty of evidence ranged from very low to moderate, with low-certainty evidence supporting the statistically significant findings. Given the heterogeneity of interventions and comparator care, digital rehabilitation should be considered an adjunct to, rather than a replacement for, conventional rehabilitation. PROSPERO, CRD420261320908, registered on 18 February 2026; Not applicable.