Bidirectional dynamics: A cross-lagged examination of physical exercise and loneliness in older adults

This longitudinal study employs cross-lagged panel modeling to analyze dynamic associations between physical activity and loneliness in older adults. The observed reciprocal associations provide preliminary evidence supporting the potential integration of exercise regimens into multidimensional approaches for addressing social isolation in aging populations.

This longitudinal study assessed 176 community-dwelling older adults (60-85 years) from four urban districts in Jinhua, China, using stratified sampling. Physical activity was measured with PARS-3 (assessing weekly frequency, duration, and intensity); loneliness with ULS-8 (α = 0.89). Participants completed three assessments at 1-month (T1), 2-month (T2), and 3-month (T3) follow-ups, with 92.6 % retention.

Mann-Whitney U tests indicated significant gender differences in exercise adherence (U = 2153, P < 0.001, d = 0.82), with females engaging more consistently. No gender disparities emerged in loneliness trajectories (P = 0.27). Robust inverse exercise-loneliness correlations persisted across waves: T1 (r = -0.316), T2 (-0.454), T3 (-0.491) (all P < 0.001). Cross-lagged analyses revealed bidirectional links: T1 exercise predicted lower loneliness at T2 (β = -0.409) and T3 (-0.335), while T1 loneliness predicted reduced subsequent exercise at T2 (β = -0.211) and T3 (β = -0.247) (all P < 0.001).

The study demonstrates a bidirectional relationship between physical exercise and loneliness among Chinese older adults, marked by time-sensitive asymmetry and self-reinforcing dynamics. Findings validate exercise interventions' potential to disrupt the "loneliness-sedentary loop," with older women showing higher activity engagement yet comparable loneliness levels to men. Optimized efficacy requires supervised group-based regimens (3-4 sessions/week, 60-90 min/session), balancing physiological adaptability and social connectivity in aging populations.