JMIR Publications

ABSTRACT

Background:

Most studies on internet use and health outcomes among older adults rely on cross-sectional designs and binary exposure measures. It is usually difficult for time to capture multidimensional health-related digital engagement. The high collinearity between digital engagement and socioeconomic factors makes it challenging to disentangle independent effects from marker effects. Currently, longitudinal evidence linking health-related digital engagement to incident stroke remains limited.

Objective:

This study aimed to examine the longitudinal association between a composite Health-Related Digital Engagement Index (HDEI) and incident stroke among community-dwelling older adults. Bisides, it sought to quantify the extent to which socioeconomic factors account for this association.

Methods:

This prospective cohort study used data from the National Health and Aging Trends Study (NHATS), Waves 1-10 (2011-2020). The HDEI (range 0-4) was constructed from 4 health-related internet behaviors at baseline. The primary outcome was incident stroke ascertained by self- or proxy-reported physician diagnosis. Discrete-time hazard models with a complementary log-log link were fitted with 4 nested models progressively adjusting for demographics, socioeconomic factors, chronic disease burden, disability, and social isolation.

Results:

Among 5,384 participants (81.6% HDEI=0; 10.5% HDEI=1; 7.9% HDEI≥2) followed for a median of 5 years (IQR 2-9), 470 incident stroke events occurred. In the unadjusted model, each 1-point HDEI increase was associated with 24% lower stroke risk (hazard ratio [HR] 0.76, 95% CI 0.66-0.88; P<.001). After adjustment for age and sex, the association attenuated but remained significant (HR 0.82, 95% CI 0.71-0.94; P=.006). Upon further adjustment for race or ethnicity, education, and income, the association was no longer significant (HR 0.91, 95% CI 0.78-1.05; P=.18); full adjustment yielded similar results (HR 0.91, 95% CI 0.79-1.04; P=.18). Subgroup analyses showed a stronger association among men (HR 0.70, 95% CI 0.55-0.89; P=.003), though no interaction terms reached significance. Sensitivity analyses excluding early events and substituting cellphone use as an alternative exposure yielded consistent attenuation patterns. Sensitivity analyses excluding early events and using cell phone instead as a alternative exposure variable showed a similar attenuation patterns.

Conclusions:

In unadjusted and sociodemographic-adjusted models, higher health-related digital engagement was associated with lower stroke incidence. However, after adjusting for socioeconomic factors, this relationship was reduced. The observed association between digital engagement and stroke risk seems to be predominantly confounded by socioeconomic advantage. Therefore, digital health interventions those aiming at stroke prevention should address both the digital divide and the underlying socioeconomic determinants of cerebro-cardiovascular risk.