Background The WHO conceptual framework for action on the social determinants of health suggests that social relationships such as social cohesion and social capital cut across structural and intermediary determinants of health to influence social inequalities in health. Whilst evidence has found that material, behavioural, physiological, and psychosocial pathways link socioeconomic position and health, few studies have explored the contribution of social relationships to social gradients in health.

This study examines the mediating role of social relationships in the association between socioeconomic position and health amongst a representative sample of UK older adults.

Methods Cross-sectional analyses were based on 8688 core cohort participants aged ≥50 years from wave 2 of ELSA. Three health outcomes health were explored: self-rated general health, systolic and diastolic blood pressure. Non-pension wealth, education, and occupational status were used to capture associations between socioeconomic position and health outcomes. Four measures of social relationships were included as potential mediators: loneliness, social isolation, positive social support, and negative social support. Analyses were performed in STATA MP 17.0 using regression models (adjusted for age and sex) after imputing the missing data.

Results Amongst this sample, 53.9% were women than 46.1% were men. 30% of older people reported as having fair to poor self-rated general health. Mean systolic and diastolic blood pressure were reported as 136.8 mm HG and 76.5 mm HG respectively. 41% of participants had systolic hypertension and 13% had diastolic hypertension.

Older people in the lowest socioeconomic groups had between 2.5- and 5.5-times greater odds of reporting poorer self-rated general health compared to their most affluent peers (OR_wealth: 5.50, 95% CI 4.59–6.60; OR_occupation: 2.40, 95% CI 2.13–2.72; OR_education: 2.69, 95% CI 2.34–3.10). Loneliness explained between 14% and 25% of the association between socioeconomic position and self-rated general health and social isolation explained between 21% and 22% of this association.

Least affluent individuals had 2.5–3 mm HG higher systolic blood pressure compared to the most affluent groups (B_wealth: 3.06 mm HG, 95% CI 1.38–4.74; B_occupation: 2.49 mm HG, 95% CI 1.37–3.60; B_education: 2.64 mm HG, 95% CI 1.46–3.83). Social isolation appeared more important in comparison to other social relationship measures and reduced systolic blood pressure between 0.31 mm HG and 0.56 mm HG. The models measuring diastolic blood pressure were largely non-significant.

Conclusion Social relationships, especially loneliness and social isolation, contribute to socioeconomic differences in self-rated general health and systolic blood pressure. The attenuation was clearest for wealth, for which social gradients appeared the steepest.