Background The quantity and quality of an individual’s social relationships have been associated with health behaviours, healthcare utilisation and health outcomes. The Public Health and Health and Social Care Outcomes Frameworks include social isolation indicators but do not explicitly consider socioeconomic patterning.

Observational studies suggest that social network characteristics are socioeconomically patterned, with those with lower educational attainment, income or occupational grade having more frequent contact with family but smaller friendship networks and less frequent participation in groups. Results, however, are not consistent and studies fail to consider changes over time.

Methods Using data from the MRC National Survey of Health and Development (also known as the 1946 British birth cohort), this study takes a lifecourse approach to exploring the relationship between socio-economic position on trajectories of social integration between age 36 and 60–64 (n = 1487). Trajectories of social integration were developed from observed variables (frequency of social group participation, frequency of informal social contact and marital status) using Latent Class Analysis. Socio-economic position in childhood (father’s occupation) and early adulthood (educational attainment and head of household occupation) were modelled as covariates to explore how the odds of following each profile varied by socioeconomic position.

Results Four profiles of social integration were identified for men, one increasing and three decreasing. Higher levels of education were associated with reduced odds of following a profile of decreasing social integration (ORs 0.61 to 0.73) compared to the increasing social integration profile. Four profiles were identified for women: one stably high, one increasing and two declining. The relationship with socioeconomic position was stronger and consistent across measures. Those with higher education or higher head of household occupation were more likely to maintain high levels of social integration (each one level increase in educational attainment was associated with an OR of maintaining high social integration of 2.89). An association between childhood socioeconomic position and head of household was also found.

Conclusion These findings suggest that those with lower socioeconomic position are more at risk of declining social integration up to age 60–64, particularly amongst women. Previous research suggests that low levels of social integration are related to both morbidity and mortality. Further research is required to explore the health impact of profile of social integration and the potential benefit of intervention to support maintenance of social networks at older ages, particularly for those with lower educational attainment or lower occupational grade.