A network meta-analysis of 22 randomised trials compared non-pharmacological multicomponent interventions in community-dwelling older adults with frailty or prefrailty. Network meta-analysis allows interventions that were never tested head to head to be ranked against each other, which is its value and its main weakness.
For overall frailty level, exercise combined with cognitive intervention ranked highest, followed closely by exercise with cognitive intervention and social support, then exercise with nutrition. For motor ability specifically, exercise plus nutrition plus psychological intervention ranked first.
The consistent element is that exercise alone did not top either ranking. Every leading combination pairs it with something else — cognitive training, nutrition, psychological or social support. That is a useful corrective to frailty pathways built around an exercise class and little else.
The caution is in how these ranks are produced. SUCRA values order interventions by the probability of being best; they do not say the top-ranked one is significantly better than the second, and with 22 trials spread across many combinations several comparisons rest on very little data.
- Build frailty interventions as combinations; exercise alone ranked below every pairing.
- Cognitive training was in both top-ranked combinations for frailty level — worth including where a service can.
- Where motor ability is the goal, nutrition and psychological support were the additions that ranked highest.
- Read the rankings as ordering, not as proof that first beat second.
The statistics, in plain English
SUCRA is the probability that an intervention is among the better options across the whole network, given as a percentage. 84.7% against 83% for the top two is not a meaningful gap: those two are effectively tied, and treating the ranking as a league table over-reads it. Network meta-analysis infers comparisons that were never made directly, which is powerful when the trials are similar and misleading when they are not. Twenty-two trials across this many combinations means some estimates rest on one or two studies. The finding worth carrying is the pattern — combinations beat exercise alone — rather than the specific order.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for geriatrics, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free