Frequently Asked Questions

The 6-month digital multidomain intervention improved global cognition, visuospatial memory, and loneliness compared with usual daily activities. Significant between-group differences favored the intervention on the Montreal Cognitive Assessment (MoCA; t=2.106, P=.037, d=0.327), Rey-Osterrieth Complex Figure Test (ROCFT) immediate recall (t=2.42, P=.017, d=0.376), ROCFT long-delayed recall (t=2.797, P=.006, d=0.434), and the UCLA Loneliness Scale (Z=-2.641, P=.008, r=0.237). Other measured domains, including verbal memory, language, attention, executive function, mobility, depression, anxiety, quality of life, and health-promoting behaviors, showed favorable or maintained trends but did not differ significantly between groups.

Participants were classified as high risk if they were aged 60 years or older and had a modified dementia risk score (MDRS) greater than 79. In this study, that cutoff represented the top 20% risk percentile and was associated with a hazard ratio of 3.2 (95% CI, 2.1-4.8). The MDRS incorporated multiple risk factors, including variables such as age, sex, education, hypertension, physical activity, and depression, and excluded APOE ε4 status.

The intervention was delivered through a WeChat applet over 6 months and combined 6 modules: health education, health monitoring, cognitive training, cognitive stimulation, cognitive rehabilitation, and social interaction. Health education sent weekly dementia-related articles and videos, while health monitoring included daily lifestyle records and monthly self-assessments of emotional and psychosocial status.

Cognitive training and cognitive rehabilitation were prescribed 3-5 times per week for at least 120 minutes weekly. Cognitive stimulation used weekly art-based assignments, and the social interaction module encouraged ongoing participation through sharing progress, giving feedback, and commenting within the WeChat group.

No. The clearest benefits were selective rather than broad across all domains. Significant improvements were seen in general cognition on the MoCA and in visuospatial memory on ROCFT immediate and long-delayed recall, but between-group differences were not statistically significant for verbal memory on the AVLT, language on the VFT and BNT, or attention and executive function on the STT-A and STT-B.

The study also did not find significant between-group differences for mobility measures, health-promoting behaviors, depression, anxiety, or quality of life, even though the intervention group generally showed improvement or maintenance in those areas.

Yes. The intervention group had a significantly greater reduction in loneliness than the control group on the UCLA Loneliness Scale, with Z=-2.641, P=.008, r=0.237. The authors interpreted this as a meaningful psychosocial benefit alongside the cognitive findings.

Yes. Subgroup analyses suggested that intervention effects varied by demographic, lifestyle, social, and clinical characteristics. The study reported stronger effects in some outcomes among participants with higher education, females for MoCA improvement, males for some memory and loneliness outcomes, married individuals, those not living alone, nonsmokers, nondrinkers, and those more active in physical, intellectual, or social activities.

The authors also reported stronger responses in some outcomes among participants without conditions such as stroke, diabetes, hyperlipidemia, chronic heart disease, insomnia, or depression, and among those without a family history of dementia. These subgroup findings were exploratory and are presented as differences in responsiveness rather than definitive patient-selection rules.

Yes. Participants with non-amnestic MCI showed significant benefits across multiple outcomes, whereas no significant between-group differences were detected in the amnestic MCI subgroup. In the non-amnestic MCI subgroup, significant effects were reported for MoCA (t=2.420, P=.017), AVLT long-delayed recall (t=2.120, P=.036), ROCFT immediate recall (t=3.138, P=.002), ROCFT long-delayed recall (t=3.912, P<.001), STT-A (Z=-2.029, P=.042), and the UCLA Loneliness Scale (Z=-2.021, P=.043).

No. The study found that overall adherence level was not significantly associated with intervention outcomes across cognitive, mobility, psychosocial, or health behavior measures (all P>.05). This means participants appeared to derive benefit even with varying levels of total engagement.

However, adherence to specific modules mattered. Higher participation in daily health lifestyle records was associated with better AVLT long-delayed recall (t=-2.277, P=.025, Cohen d=0.607), and higher participation in cognitive rehabilitation was associated with better AVLT recognition recall (H=8.557, P=.014, η²=0.104) and STT-A performance (H=14.48, P<.001, η²=0.177).

The intervention appeared feasible in this study. Of 166 randomized participants, 154 completed the post-intervention assessment, and 69 of 83 participants in the intervention group met the study's high-adherence threshold. The authors described this as strong engagement and sustained participation over 6 months.

The platform was designed with age-friendly adaptations, and adherence support combined passive system tracking with active researcher follow-up during the first 3 months, including reminders by phone or WeChat when participants missed tasks or became inactive for more than 1 week.

This was a 2-arm, parallel randomized clinical trial conducted in China with 166 participants, using intention-to-treat analysis and a 6-month intervention period. Participants were randomized 1:1 to either the WeChat-based multidomain intervention or a control group that continued usual daily activities without the applet, and outcome assessors and statistical analysts were blinded.

Because this was a nonpharmacologic intervention, participants and intervention staff could not be blinded. The design supports a comparison between the digital intervention and usual activity, but the absence of an active control group limits conclusions about which individual components of the program caused the observed benefits.

The main limitations were that it was a single-center study, the intervention lasted only 6 months, and outcomes were based mainly on neuropsychological testing without neuroimaging or biomarker data. The authors also noted that the digital format could introduce variability related to participants' technology adaptability and device access.

Additional limitations included the absence of an active control group, lack of stratification by specific cognitive impairment subtypes at enrollment, limited detail on control of chronic conditions such as glycemic or blood pressure control, and incomplete assessment of all modifiable dementia risk factors. These factors may affect generalizability and limit causal inference about specific intervention components.