Community-based Exercise Enhanced by a Self-management Programme to Promote Independent Living in Older Adults

A Pragmatic Randomised Controlled Trial

Pia Øllgaard Olsen; Mark A Tully; Borja Del Pozo Cruz; Manfred Wegner; Paolo Caserott


Age Ageing. 2022;51(7):afac137 

In This Article


The main finding of this pragmatic randomised study was that the 24-week intervention combining exercise with a multi-factorial self-management programme (CALSTI) led to a faster, statistically and clinically superior increase in functional capacity compared with the self-management programme alone (SEMAI). The increase in CALSTI exceeded the 1-point threshold for substantially clinical meaningful changes in SPPB[46] after 12 and 24 weeks, whereas SEMAI exceeded the 0.5-point threshold for moderate meaningful changes only after 24 weeks. The improvement in the CALSTI group led to a follow-up SPPB-score > 9 which is relevant because this cut point has previously been shown to be predictive of home care service utilisation in a similar Danish population.[38] In addition, the CALSTI group improved self-report measures of ADL and IADL disability, and self-reported functional capacity after 12 weeks combined intervention, and despite some tendency of declines, most of these self-reported improvements remained at Week 24.

The EQ-health VAS baseline level of 60.0 corresponded a 16–17-point deficit compared with the national +75-year population-norms for men and women, respectively. Approximately 25% of this gap was caught up at Week 12 in CALSTI; however, at 24-week follow-up, the improvement was no longer significant.

The main exercise component in CALSTI was a progressive power-type resistance training protocol that has proven highly effective for enhancing neuromuscular function, including gait speed, in controlled set-ups.[35,50] Despite the diversity of facilities, trainers and equipment across sites in this study, changes in functional capacity were not inferior to those observed in earlier exercise-studies in community-dwelling older adults.[43,51] The external validity was increased by the unique recruitment strategy through a well-known and well-accepted nationally regulated pathway. This possibly allowed a broader reach, and more specific recruitment of the intended at-risk target group demonstrated by the high proportion of the screened subjects meeting the eligibility criteria (44%). Loss-to-follow up was associated with lower SPPB score in the CALSTI group only, indicating that this exercise protocol may need more extensive tailoring (i.e. reduced number of exercises, intensity, volume) to lower the barriers for persons with more advanced functional decline.

Interestingly, improvements in functional capacity followed different patterns in CALSTI and SEMAI. CALSTI improved mainly during the intensive exercise phase (Week 0–12), whereas SPPB improvements in SEMAI occurred throughout the 24-week period. Possibly, the self-management programme enabled participants of both groups to initiate and maintain new PA-behaviours on their own. Several programme features may have contributed to the observed changes. First, the self-management programme used multiple behaviour-change strategies that have previously been linked to higher PA-levels in older adults.[8–11,13,14] Second, the programme was theoretically based on the HAPA framework for which the causal pathways have been empirically demonstrated,[52] and theory-based PA interventions seem in general superior to increase PA.[53] Finally, the use of motivational interviewing principles for developing group norms to elicit a safe, respectful and open-minded environment, may potentially have facilitated self-efficacy, social support, peer learning and role modelling.

The limitations in the study should be noted when interpreting the findings. Because this study was purposely embedded in nationally regulated primary care service pathways, health care providers were not allowed to offer citizens a service of lower standard than normal care. Also, the providers wished to compare CALSTI to a structured and more intensive version of the PHV (existing service) that may potentially become new regulated standard care in the future. This ruled out the inclusion of a passive control group, preventing final conclusions of whether the SMS-programme caused the improvements in the SEMAI group. The extra attention itself as well as selection bias towards those with higher motivation and readiness for change may have contributed to some of the observed effects in both groups. On the other hand, this possibly led to more conservative effect sizes that may closer reflect the actual benefits of implementing the CALSTI intervention. Another limitation is the high number of participants, predominantly from the SEMAI group, who did not provide self-reported outcomes at any time point. The baseline battery of questionnaires was filled in after group allocation, potentially affecting participants' motivation to complete the questions, increasing the risk of selection bias and low statistical power to these outcomes. Consequently, the results from self-reported outcomes in the SEMAI group should be considered suggestive only, and we were unable to establish if the increases in SPPB performance carried over to improved self-rated functioning and disability. The considerable proportion of missingness (i.e. ≈40%, Supplementary Table 3) in the primary outcome, SPPB, at follow-up in SEMAI, was taken into consideration by adding baseline-SPPB to the imputation- and analysis models. Moreover, several diagnostic procedures were undertaken to investigate the validity of the imputation approach including sensitivity analyses based on observed data only (Supplemental Material S5). The results of these analyses were highly similar to the primary analysis on imputed data, and, importantly, the two approaches did not result in any conflicting conclusions (Table 2 and Supplementary Table 8).

Our study has several strengths. We used a pragmatic setup with involvement of key stakeholders in the development and implementation processes without compromising important aspects to methodological quality and reproducibility. We applied a novel approach, allowing participants to self-determine which barriers for PA they would address (e.g. nutritional risk factors, loneliness, incontinence), thereby accounting for the multi-factorial and subjective nature of PA-participation and prevention of late life disability. Finally, we showed strong effects on the primary functional capacity outcome, SPPB.