Social support and posttraumatic stress disorder: A meta-analysis of longitudinal studies

Abstract

Social support has long been associated with posttraumatic stress disorder (PTSD), but there is no consistent evidence on the strength and direction of this relationship. Whereas the social causation model claims that social support buffers against PTSD, the social selection model states that PTSD reduces social support resources. As the first meta-analysis of the prospective relationships between social support and PTSD, this study synthesized the available longitudinal data (75 samples including 32,402 participants) on these two constructs with a random-effects model. In total, three hundred and fifty-five effect sizes (including cross-sectional, prospective and cross-lagged coefficients) were included in the meta-analysis. With prior levels of the relevant outcomes controlled for, results showed that social support and PTSD reciprocally predicted each other over time with similar effect sizes: Social support predicted PTSD with β = -0.10; PTSD predicted social support with β = -0.09. Moderator analyses suggested that the effects held across most sample characteristics and research designs except for several moderators (gender, time lag, publication year, source of support). These findings provided strong evidence for both the social causation and social selection models, suggesting that the link between social support and PTSD is symmetrically reciprocal and robust.

Keywords: Longitudinal studies; Meta-analysis; PTSD; Social support.