With dementia cases projected to surge most dramatically in lower-income regions, the uncomfortable reality is that nearly all prevention research has been conducted in wealthy nations whose populations, healthcare systems, and cultural contexts differ substantially from those facing the greatest burden. This systematic review directly confronts that mismatch, offering the first consolidated evidence base for non-pharmacological dementia prevention in lower- and middle-income countries.
Drawing on 25 randomized controlled trials from six countries — with China contributing 17 of those trials — and covering 1,304 participants with mild cognitive impairment or subjective cognitive decline, the meta-analysis found a standardized mean difference of 1.49 favoring lifestyle interventions over controls, a large effect by conventional benchmarks. Broken down by type, exercise-based interventions yielded the strongest signal (SMD 1.67, across eight studies), while multidomain approaches — combining cognitive training, physical activity, and sometimes nutritional or social components — also demonstrated replicated benefit (SMD 1.22, five studies). Arts and creative expression interventions lacked sufficient pooled data for meta-analysis.
The effect sizes here are notably large, which should prompt measured scrutiny rather than enthusiasm. Publication bias was explicitly detected, meaning negative or null trials are likely underrepresented — a chronic problem in behavioral intervention research that inflates apparent efficacy. The geographic concentration in China limits generalizability across the diverse social, economic, and epidemiological landscapes of LMICs in South Asia, sub-Saharan Africa, and Latin America. Sample sizes across individual trials were modest, and the heterogeneity of outcome measures complicates direct comparison. That said, the directional consistency across exercise and multidomain categories aligns with established findings from high-income country trials like FINGER and MAPT, lending biological plausibility. This review is best understood as a call for methodologically rigorous, culturally tailored prevention trials designed from the ground up for LMIC contexts — an investment the global dementia burden increasingly demands.