Stability, Asymmetric Recovery, and Sociodemographic Heterogeneity in Social Health Trajectories Across Social Disruption: A Four-Wave Hidden Markov Model

Social health is a key pillar of overall health, yet knowledge of distinct social health states, how these change during long-term social disruption, who is at risk of social ill-health, and how social health states affect mental and physical illness remains limited. We used Hidden Markov models to analyse prospective data from 954 adults aged 40 and over who were assessed before, during, and after the onset of the 2019–2025 pandemic. This enabled us to identify latent social health states based on network size, density, loneliness, and emotional support, and to estimate transition probabilities between states over time. We examined the impact of sex, education, and age on state membership and trajectories, as well as the association between states and depressive symptoms and physical health. Three states were identified: socially isolated (15.9%), moderately connected (55.5%), and socially thriving (28.6%). Most participants (72.5%) remained in one state throughout the observation period; 14.4% deteriorated, 6.5% recovered, and the rest fluctuated. Being male and having completed a practical education were associated with the socially isolated state, while being female was associated with the thriving state, as well as state deterioration. Worse states showed a graded association with depressive symptoms. Social health is characterised by three relatively stable states, even in the face of persistent social disruption. Nevertheless, many individuals exhibited persistent deterioration even after social restrictions were lifted. The differing risk across sociodemographic groups highlights the need for targeted strategies to promote resilient social health.