Blended eHealth resistance training linked to stronger gains in walking speed among older adults

An adapted version of the World Health Organization’s Integrated Care for Older People pathway, combined with blended eHealth resistance training, was associated with improved locomotor capacity in community-dwelling older adults, according to a retrospective quasi-experimental study conducted in Taiwan. The findings, published in the European Review of Aging and Physical Activity, suggest that merging screening with immediate intervention may overcome one of the persistent weaknesses of the WHO framework: the fragmented transition between assessing older people and actually treating them.

The WHO’s ICOPE pathway is designed to detect declines in six domains of healthy aging, including locomotion, and to connect those who screen positive with appropriate care. Yet in community settings, the handoff from step one, screening, to step two, further assessment and referral, often breaks down. Older adults who test poorly on a walking test may never receive the targeted exercise program that could restore their mobility. The Taiwanese team restructured this sequence for the locomotion domain by collapsing the first two steps into a single encounter, pairing a four-metre walk test with a five-times sit-to-stand test on the same visit.

Participants who failed either threshold were not simply referred elsewhere. Instead, they entered a four-movement progressive resistance-training program delivered through a hybrid model that blended digital eHealth tools with in-person community support, embedded within Taiwan’s Long-Term Care 2.0 tier-C network. A reassessment at week eight closed the loop, allowing clinicians to verify whether the intervention had produced measurable change in walking speed and sit-to-stand performance.

The study enrolled 858 community-dwelling adults aged 60 or older at baseline. After seven exclusions, 851 participants were analyzed, of whom 811 joined the intervention arm and 40 chose to serve as assessment-only comparators. Fifteen of those comparators had baseline impairment. Because allocation was by self-selection rather than randomization, the researchers stratified participants by baseline locomotor status, defining impairment as a gait speed below 1.0 metres per second or a sit-to-stand time exceeding 10.8 seconds.

Statistical modelling revealed a group-by-time interaction favoring the intervention. Participants in the program gained 0.13 metres per second more in gait speed than comparators, with a 95 percent confidence interval of 0.04 to 0.22 and a P value of 0.005. The effect size, expressed as Cohen’s d, was 0.36. Crucially, this gain exceeded the minimum clinically important difference of 0.10 metres per second, the threshold beyond which patients and clinicians typically perceive a meaningful improvement in walking ability.

The sit-to-stand results were directionally consistent but less robust. The intervention group reduced their five-times sit-to-stand time by an additional 1.13 seconds compared with comparators, with a confidence interval spanning −2.00 to −0.26 seconds and a P value of 0.011. However, the effect was significant in 35 of 37 leave-one-district-out refits, and continuous baseline moderation was significant under model-based inference but not under the wild-cluster bootstrap, which returned a P value of 0.223. The authors also noted that no population-appropriate minimum clinically important difference exists for this measure, making interpretation more cautious.

To guard against the statistical pitfalls inherent in quasi-experimental designs, the team employed a battery of sensitivity analyses. Three-level mixed-effects models with random intercepts for participant and district estimated group-by-time effects. Inverse probability of treatment weighting, cluster-robust standard errors, the wild-cluster bootstrap, and leave-one-district-out analyses all tested whether the findings depended on particular modelling choices or on the influence of any single district. The gait-speed result remained robust across all 37 leave-one-district-out refits, while the sit-to-stand finding held in 35 of 37.

One of the most serious threats to validity in non-randomized pre-post comparisons is regression to the mean, the statistical tendency for extreme baseline values to move toward the average on repeat testing regardless of any true treatment effect. The impaired comparator stratum showed essentially no change over time, whereas the impaired intervention stratum, comprising 365 participants, improved by 2.17 seconds on the sit-to-stand test and 0.21 metres per second on the walk test, both with P values below 0.001. This divergence between equally impaired groups argues against regression to the mean as an explanation. Inverse probability of treatment weighting produced concordant results, and the estimates persisted after modelling district clustering, with intraclass correlation coefficients of 0.41 and 0.19.

The authors conclude that this ICOPE-informed locomotion pathway was associated with enhanced locomotor capacity, most robustly for gait speed, with supportive but less consistent sit-to-stand findings, particularly among those with baseline impairment. They are careful to frame the results as associational rather than causal, given that allocation was by preference and the comparator group was small. The study also addresses only one ICOPE domain rather than complete implementation of the full pathway.

For public health systems grappling with aging populations, the implications are considerable. Taiwan’s tier-C network under Long-Term Care 2.0 provides community-based stations where older adults can receive preventive services. Embedding a streamlined screen-and-treat locomotion pathway within that existing infrastructure, augmented by eHealth tools that reduce the burden of in-person visits, appears feasible at scale. Whether the same model transfers to health systems with different primary care structures, digital literacy levels, or funding mechanisms remains an open question, and randomized trials would be needed to confirm the causal interpretation that this quasi-experimental design cannot deliver.

Subject of Research: An adapted WHO ICOPE locomotion pathway with blended eHealth resistance training and its association with locomotor capacity in community-dwelling older adults.

Article Title: An adaptation of the WHO 4-step Integrated Care for Older People locomotion pathway with blended eHealth resistance training associated with enhanced locomotor capacity in community-dwelling older adults: a retrospective, quasi-experimental study

Article References: Lien, W.-C., Chen, J., Wang, W.-M., Li, T.-F., & Tsai, M.-L. (2026). An adaptation of the WHO 4-step Integrated Care for Older People locomotion pathway with blended eHealth resistance training associated with enhanced locomotor capacity in community-dwelling older adults: a retrospective, quasi-experimental study. European Review of Aging and Physical Activity. https://doi.org/10.1186/s11556-026-00444-8

Image Credits: AI Generated

DOI: 10.1186/s11556-026-00444-8

Keywords: ICOPE, WHO Integrated Care for Older People, resistance training, eHealth, gait speed, sit-to-stand test, older adults, community-dwelling, quasi-experimental study, Taiwan, locomotor capacity, Long-Term Care 2.0