Patient-Reported Markers of Delayed Physical Function Recovery After Kidney Transplantation: Supervised Cluster Analysis of PROMIS Physical Function Scores

Background: Kidney transplant recipients (KTRs) with delayed physical function (PF) recovery post-transplant may benefit from early rehabilitation support. We aimed to identify patterns of PF recovery using supervised cluster analysis of longitudinally obtained Patient Reported Outcome Measurement System (PROMIS) PF scores. Methods: Longitudinal convenience sample of adult KTRs who completed PROMIS-PF Computer Adaptive Test (CAT) (higher=better PF) within ~1 week post-transplant and biweekly over 2 months. We stratified participants by baseline PROMIS-PF T-score (>30 vs. ≤30) and by T-score change between baseline and week 2 (≥2 vs <2; improved vs. non-responder). Four clusters (EXPOSURE) were identified (Fig. 1): CL1(poor baseline– non-responder), CL2(poor baseline–improved), CL3(better baseline–non-responder), CL4 (better baseline–improved). We used linear mixed-effects models with participant random effects and time-by-cluster interactions to assess trajectories, and interval regression to estimate time-to-recovery (T-score ≥45-OUTCOME), followed by Wald tests for group comparisons. Results: Of 63 participants, 3(5%) were in CL1, 15(24%) in CL2, 37(59%) in CL3, and 8(12%) in CL4. Age, organ type, socioeconomic status, and ethnicity were similar across clusters. CL1 were all females, and had highest (most fatigue) baseline PROMIS fatigue scores; other baseline patient-reported outcomes characteristics were similar across clusters. At 2 months post-transplant, mean(95% CI) PROMIS-PF scores were: CL1: 32(26–38) vs CL2: 43(40–46)(p=0.001); CL3: 46(44–48) vs CL4: 51(47–55)(p=0.01). The proportion who reached outcome event were 0%, 54%, 54%, and 94% in CL1,CL2,CL3 andCL4, respectively(p=0.01). Conclusion: Baseline and early PROMIS-PF change identified four distinct recovery clusters among KTRs. Funding: Government Support – Non-U.S.