1. Target populations|
• A risk reduction intervention should not be applied unselectively (focus on various at-risk groups).
• At-risk groups should be preferably selected using validated risk scores or algorithms.
• The most suitable risk score or algorithm should be carefully chosen to fit the purpose, e.g., stage of the risk/disease continuum, age group, level of cognitive performance, and type of intervention to be applied. For example, for multidomain lifestyle interventions, the risk score/algorithm should select individuals with the type of risk profile that the intervention aims to modify.
• Risk reduction interventions should preferably start early, before substantial brain pathology and cognitive/functional impairment have already occurred.
• People with genetic susceptibility for dementia (e.g., based on APOE ε4 genotype) may also benefit from early risk reduction interventions. This should be further investigated in intervention studies.
• Multidomain interventions (targeting several risk factors and disease mechanisms simultaneously) may be needed for an optimal dementia risk reduction.
• Interventions should (i) do the right things and (ii) do enough for them, i.e., target an individual’s overall risk profile with sufficient intensity to produce an effect. Only general healthy lifestyle advice may not be enough, and a more structured intervention program should be proposed.
• Intervention content should be adapted to local/national risk context (e.g., some risk factors may be more prevalent/severe in some countries than others) and various settings and integrated with other chronic non-communicable diseases risk reduction programs when feasible.
• Radical lifestyle changes may be difficult to both initiate and maintain longer term. Smaller changes gradually introduced across multiple lifestyle domains may facilitate long-term adherence.
• As the social component is important, group sessions and/or group activities should be facilitated when feasible.
• New technology may facilitate effective, personalized, and feasible interventions and implementation (eHealth and mHealth).
• Intervention effects should be monitored. Risk scores could be useful for this purpose as well, if they include modifiable factors and are sufficiently sensitive to change over time.