Out of 1,406 screened participants, 607 were eligible for this analysis and 215 (35%) of them agreed to be randomly assigned to either CALSTI (n = 113) or SEMAI (n = 102) interventions (Figure 3 and Supplementary Material S6, Supplementary Tables 6 and 7). Most loss-to-follow-up happened before Week 12. Loss-to-follow-up was associated with lower baseline SPPB-score in the CALSTI group (6.5 ± 0.5 vs. 8.3 ± 2.2, P < 0.001), but not with any descriptive variable (age, sex, education or living status, P > 0.05). Participants mean age was 80 years (range 65–93), 54% lived alone, the majority were women and had completed at least basic education. A larger share of SEMAI participants reported interference of pain in daily activities and had lower baseline SF-LLFDI function (Table 1).
Primary Outcome. Significant group × time interactions revealed that SPPB-score increased 1.16 points more in the CALSTI compared with the SEMAI group during Week 1–12 (P < 0.001), and 0.77 points more over the total duration of the 24-week intervention (P = 0.04) (Table 2). Within-group analysis indicated that participants in CALSTI increased their SPPB-score by 1.52 point (19.7% P < 0.001) after 12 weeks, and this level remained unchanged after 24 weeks (P = 0.63). The SEMAI group also improved during intervention although this did not reach significance until after 24 weeks (0.82 points, 11.4%; P > 0.001) with the largest increase taking place from Week 12 to 24 (0.46 points, 95% confidence interval [CI]: 0.06–0.86; P = 0.025, not displayed). Changes were similar across intervention sites as demonstrated by a minimal intra cluster correlation ( <0.1). Sensitivity analysis on observed data only largely replicated group x time estimates from intention-to-treat analysis (Supplementary Table 7).
Secondary Outcomes. The CALSTI group significantly improved disability limitation- and function domains of the SF-LLFDI (6 and 9%), ADL/IADL (29%) and EQ-VAS (7%) at Week 12. After Week 24, the SF-LLFDI function and ADL/IADL levels remained higher than baseline, whereas SF-LLFDI disability limitation declined from Week 12 to 24 (P < 0.01, not displayed). No changes were found in the SEMAI group at any timepoint, and interaction effects between group and time were demonstrated only for the SF-LLFDI function subscale at Week 12. Table 2 displays an overview of all secondary results.
Age Ageing. 2022;51(7):afac137 © 2022 Oxford University Press
Copyright 2007 British Geriatrics Society. Published by Oxford University Press. All rights reserved.