Causal machine learning analysis identifies subpopulations most likely to benefit from healthy lifestyles for cognitive outcomes

Background: With dementia posing an escalating public health threat, adopting healthy lifestyles is increasingly recognised as a key preventive measure for preserving cognitive function. However, the effect of multidomain lifestyle intervention on cognitive health remains inconclusive and likely heterogeneous. Aims: We aimed to investigate the heterogeneity in the associations between healthier lifestyles and cognitive outcomes (cognitive performance and incident dementia) across identifiable subpopulations and to use causal machine learning to identify participant characteristics predictive of greater benefit. Methods: Causal machine learning analyses were performed using two prospective cohorts. The English Longitudinal Study of Ageing (ELSA) (waves 4–9, 2008–2018) served as the discovery cohort and the Health and Retirement Study (HRS) (waves 9–14, 2008–2018) served as the validation cohort. Healthy lifestyles included smoking status, alcohol consumption, physical activity and social contact. Global cognitive z-scores were computed based on standardised tests of memory, executive function and orientation. Dementia diagnoses were ascertained by integrating physician evaluations with measures of cognitive and functional performance. Results: We analysed 8771 participants from ELSA and 8531 participants from HRS. High-benefit groups displayed substantially improved cognitive performance (ELSA: mean: 0.27, 95% confidence interval [CI] 0.24–0.30; HRS: mean: 0.51, 95% CI 0.48–0.54) and reduced risk of dementia (ELSA: hazard ratio [HR]: 0.28, 95% CI 0.21–0.34; HRS: HR: 0.22, 95% CI 0.17–0.27). Subgroups deriving greater benefits in cognitive performance were younger and had higher baseline cognitive reserve, superior physical and respiratory function, and better cardiometabolic profiles with particularly lower blood pressure. Those deriving greater benefits in incident dementia were likewise younger, with higher baseline cognition and superior vascular metrics, including lower blood pressure. Conclusions: The cognitive benefits of healthy lifestyles vary by individual characteristics, underscoring the potential of personalised prevention strategies. Future randomised clinical trials are warranted to validate these findings and refine intervention approaches.