A Structured Lifestyle Program Improved Cognitive Test Scores. That’s Not the Same as Preventing Dementia.

A large U.S. clinical trial offers encouraging news about lifestyle and brain health—but its conclusion is narrower than some headlines imply.

In the U.S. POINTER study, 2,111 adults ages 60 to 79 who were considered at elevated risk for cognitive decline were randomly assigned to one of two programs. Both encouraged healthier behavior. One was intensive and highly structured, with prescribed physical activity, guidance toward the MIND diet, computerized cognitive training, social and intellectual activities, health reviews and frequent peer meetings. The other was self-guided, with fewer meetings and general encouragement to make lifestyle changes.

After nearly two years, both groups improved on a composite measure of global cognition. The structured group improved by an average of 0.243 standard-deviation units per year, compared with 0.213 in the self-guided group. The difference—0.030 standard-deviation units per year—was statistically significant, but modest.

That number is not a percentage reduction in dementia risk. It does not mean that 3 out of every 100 participants avoided dementia. Instead, it describes the difference between two groups’ average trajectories on a collection of cognitive tests. The structured program produced a better result than the lower-intensity program, but the study was not designed or powered to determine whether either program reduced the number of people who eventually developed dementia.

The distinction matters because cognitive scores and dementia are related but different outcomes. A person can perform better on memory, attention or problem-solving tests without demonstrating that the biological disease processes associated with Alzheimer’s disease have been stopped. The researchers also noted that repeated testing can create practice effects: participants may improve partly because they become familiar with the tests. U.S. POINTER adjusted for that possibility, but cannot eliminate the broader limitation.

The trial also cannot identify which ingredient did the most work. Exercise, diet, cognitive training, social engagement and cardiovascular monitoring were delivered together, and the structured group received more support and accountability. The study therefore tests a package—not exercise versus diet, or cognitive training versus social activity.

A newer systematic review and network meta-analysis offers some clues. Published in The Lancet Healthy Longevity in September 2025, it combined 109 randomized trials involving 23,010 cognitively unimpaired older adults. The strongest result against health education was for physical exercise combined with cognitive training, with a standardized mean difference of 0.26. Cognitive training alone and physical exercise alone also showed statistically significant improvements, with effect sizes of 0.21 and 0.14, respectively. A broader combination involving diet, exercise, cognitive training and health education showed a smaller effect of 0.14.

Those findings do not establish a universal ranking of lifestyle choices. The review included different exercise regimens, diets, cognitive programs and control groups, and 40% of the included studies were judged to have a high risk of bias. Social activity was included among the interventions studied, but it did not emerge as the clearest standalone signal in the headline results.

Earlier evidence is even more cautious about dementia itself. A Cochrane review of nine randomized trials involving 18,452 people found a small improvement in detailed cognitive test batteries, but no evidence that multidomain programs reduced incident dementia. Only two trials in that review counted dementia cases, and the pooled result showed no meaningful difference between intervention and control groups.

Taken together, the evidence supports a measured conclusion: structured programs that combine physical activity, cognitive challenge and broader health behaviors can modestly improve cognitive-test performance in older adults at risk of decline. They may eventually prove capable of delaying dementia for some people, but that remains unproven.

For now, the most defensible message is not that one “brain-health recipe” prevents dementia. It is that cognitive health appears responsive to sustained, multiple-domain behavior change—and that structured support may make those changes more effective than advice alone. The question researchers still need to answer is whether better test scores persist long enough, and matter enough in daily life, to translate into fewer dementia diagnoses and greater independence.