Closing the loop: Design, implementation, and evaluation of a regular-feedback network intervention for social connectedness and mental health

Abstract

Persistent increases in social isolation pose significant public health challenges, yet community interventions rarely leverage social network analysis longitudinally. Most network interventions rely on limited diagnostics. This study introduces and evaluates an adaptive intervention where longitudinal network data iteratively informed program design in real-time. Over eight weeks, we collected weekly sociocentric and ethnographic data from an intergenerational group of women participating in a creative movement program to enhance meaningful social connections and improve mental health outcomes. We used these data to guide weekly group pairings to increase cohesion, reciprocal relations, and intergenerational connections. The intervention’s effects on tie formation and maintenance were assessed using Stochastic Actor-Oriented Models (SAOMs), and mental health outcomes were compared to a control group. Results show that while adaptive group pairings facilitated the creation of new close ties, their impact on maintaining existing ties was limited. The intervention group experienced improvements in mental health compared to controls. This study demonstrates the value of a real-time, feedback-driven approach, moving network interventions from a limited diagnostic tool to an adaptive process. Our findings highlight the value of embedding longitudinal network designs into active program implementation and underscore the distinction between the mechanisms governing tie formation and maintenance.