Previous studies have reported inconsistent findings regarding the association between post-traumatic stress (PTS) and post-traumatic growth (PTG). Three major issues could account for this inconsistency: (1) the lack of information about mental health problems before the disaster, (2) the concept of PTG is still under scrutiny for potentially being an illusionary perception of personal growth and (3) the overlooking of PTS comorbidities as time-dependent confounding factors. To address these issues, we explored the associations of PTS and PTG with trauma-related diseases and examined the association between PTS and PTG using marginal structural models to address time-dependent confounding, considering pre-disaster covariates, among older survivors of the 2011 Japan Earthquake and Tsunami.

Seven months before the disaster, the baseline survey was implemented to ask older adults about their health in a city located 80 km west of the epicentre. After the disaster, we implemented follow-up surveys approximately every 3 years to collect information about PTS and comorbidities (depressive symptoms, smoking and drinking). We asked respondents about their PTG in the 2022 survey (n = 1,489 in the five-wave panel data).

PTG was protectively associated with functional disability (coefficient −0.47, 95% confidence interval (CI) −0.82, −0.12, P < 0.01) and cognitive decline assessed by trained investigators (coefficient −0.07, 95% CI −0.11, −0.03, P < 0.01) and physicians (coefficient −0.06, 95% CI −0.11, −0.02, P < 0.01), while PTS was not significantly associated with them. Severely affected PTS (binary variable) was associated with higher PTG scores, even after adjusting for depressive symptoms, smoking and drinking as time-dependent confounders (coefficient 0.35, 95% CI 0.24, 0.46, P < 0.01). We also found that an ordinal variable of the PTS score had an inverse U-shaped association with PTG.

PTG and PTS were differentially associated with functional and cognitive disabilities. Thus, PTG might not simply be a cognitive bias among survivors with severe PTS. The results also indicated that the number of symptoms in PTS had an inverse U-shaped association with PTG. Our findings provided robust support for the theory of PTG, suggesting that moderate levels of psychological struggles (i.e., PTS) are essential for achieving PTG, whereas intense PTS may hinder the attainment of PTG. From a clinical perspective, interventions that encourage social support could be beneficial in achieving PTG by facilitating deliberate rumination.