Time-varying associations of pre-migration and post-migration stressors in refugees' mental health during resettlement: a longitudinal study in Australia

Understanding the time-varying association of pre-migration and post-migration stressors in refugees' mental health could help in designing tailored health promotion services at different resettlement stages and improving the efficiency of resource allocation. In this study, we explored these time-varying associations. We used data from the first four waves (October, 2013, to February, 2017) of a national refugee-based longitudinal study, the Building a New Life in Australia (BNLA) project. Post-traumatic stress disorder (PTSD) and high risk of severe mental illness (HR-SMI) were used to assess mental health. The independent variables included the number of potentially traumatic events experienced during the pre-migration process, and a range of post-migration stressors. We used logistic regression models to analyse the relative importance of variables and time-varying associations between the pre-migration potentially traumatic events, post-migration resettlement stressors, and refugees' mental health. Analyses were stratified by gender, and sociodemographic covariates included age, marital status, education level, country of birth, and weekly income. 2399 participants were surveyed in Wave 1 of the BNLA project in 2013–14, of whom 2009 (83·7%) responded in Wave 2 in 2014–15, 1894 (78·9%) in Wave 3 in 2015–16, and 1929 (80·4%) in Wave 4 in 2016–17. The three most important factors associated with mental health in each wave differed for male and female refugees, but the socioeconomic stressors of loneliness and adjustment to life in Australia were consistently prominent. Positive associations between socioeconomic stressors and mental ill-health were found for both genders, with a peak at Wave 2 (adjusted odds ratio [AOR] among men, 1·60 [95% CI 1·26–2·03], p=0.0001 for PTSD; AOR 1·86 [1·35–2·55], p=0·0001 for HR-SMI; and among women, AOR 1·81 [1·27–2·57], p=0·0009 for PTSD; AOR 2·24 [1·49–3·38], p=0·0001 for HR-SMI). Associations between loneliness and mental health fluctuated, but were significant for both genders in Wave 4 (among men, AOR 1·90 [1·21–2·99], p=0·0051 for PTSD; AOR 3·70 [2·18–6·27], p