Public Health.

Background: This study aimed to examine the influence of adherence on intervention effect in the Japan-multimodal intervention trial for the prevention of dementia (J-MINT).

Method: J-MINT was an 18-month randomized controlled trial, and participants aged 65-85 years with mild cognitive deficits were randomized into multidomain intervention (physical exercise, nutritional counseling, cognitive training, and vascular risk factor management) and control groups. This study included the participants in J-MINT who had undergone the intervention program and follow-up evaluation at least once. Adherence in each intervention component was classified as adherent (1 point) or non-adherent (0 point), and the total score was calculated (score range: 0-4 points). Adherent was defined as follows: the adherence rate to group-based physical exercise ≥ 70%, the number of nutritional counseling ≥ 14 times, the number of days of cognitive training ≥ 156 days, and those who have blood pressure, glucose levels, and cholesterol within the target range. Three groups were created based on the total score: adherent group (4 points), partially adherent group (2-3 points), and non-adherent group (0-1 points). The outcome was the changes in composite score calculated by eight neuropsychological tests from baseline to 18 months, including tests of global cognitive function, memory, attention, and executive function/processing speed. Mixed models for repeated measures were used to calculate mean differences (MD) of outcome between each adherence and control group.

Result: Out of 531 participants enrolled in J-MINT, 421 participants (mean age: 73.8 years, 52% female) were included. The number of participants in the adherence group was 9 in the adherent group, 163 in the partially-adherent group, and 31 in the non-adherent group. Statistical analysis showed that the adherent group exhibited a significant interventional effect compared to the control group (MD = 0.266, p = 0.025), while no significant mean difference was observed between the other adherence group and the control group.

Conclusion: Our findings suggest that supporting adherence is important to achieve the sufficient effect of multidomain interventions.