The Impact of Human Social Networks and Pets on Cognitive Outcomes and Mortality: A Longitudinal Analysis of the German National Cohort (2014–2019)

Jonas Mauersberger , University of Rostock
Constantin Reinke, University Medical Center Hamburg-Eppendorf - Center for Psychosocial Medicine - Department of Medical Sociology
Gabriele Doblhammer, University of Rostock

Background: Social embeddedness—the web of relationships in which individuals are situated—is a key determinant of cognitive functioning and survival in later life. This study assessed whether cognitive performance and mortality vary by social network type and whether pet ownership represents a modifiable component of social embeddedness. Data/Methods: We used Wave 1 of the German National Cohort (NAKO) with mortality follow-up to May 2024, restricted to adults aged =50 (N = 101,235). Factor analysis of six neuropsychological items yielded two domains (episodic memory; semantic–executive functions) which were categorized into low/medium/high (±1.5 SD) status of cognitive functioning. k-means clustering identified five network types (family-poor [ref.], close-family, multi-tie [friends], multi-tie [family], family & household rich). Logistic regressions estimated odds ratios (OR) of the association of network types and pet ownership with low cognition (vs. medium/high) and mortality, adjusted for sociodemographic and health covariates. Findings: Extensive, tightly knit networks are linked to better cognition and lower mortality relative to the vulnerable family-poor type. For episodic memory, the family-rich cluster shows ~20% lower odds of low performance (OR = 0.809, p = 0.023). Dog ownership relates to lower odds of low concentration/processing speed (OR = 0.798, p = 0.029). Mortality risks are substantially reduced in close-family networks (OR = 0.684, p = 0.009) and are lower across other clusters; pet ownership shows no significant mortality effect. Conclusion: Cognitive performance and survival in later life follow social—not solely medical—structures. Targeting network-poor constellations and leveraging modifiable resources (e.g., dog ownership) offer actionable levers for prevention and intervention.

See extended abstract

 Presented in Session P6. Health, Mortality, and Ageing 2