Social determinants of mental health in new refugees in the UK: cross-sectional and longitudinal analyses

Abstract

BackgroundThe mental health of resettled refugees is poorer than that of the general population. Resettlement to a new country and culture can present social challenges, such as building new social support networks and gaining suitable employment. Social factors in resettlement represent modifiable targets for public health policies and interventions to improve refugee mental health. Despite implicating individual social factors, previous research has failed to adjust for potential confounding by investigating many social determinants simultaneously. Additionally, there is a dearth of research causally and temporally linking social determinants and mental health with longitudinal analysis. This study investigated the effect of social determinants on mental health in new refugees with adjusted cross-sectional and longitudinal analysis. MethodsSecondary analysis of a dataset from the Longitudinal Survey of New Refugees study done by the UK Home Office from 2005–07 was undertaken. All new refugees identified by the UK Border Agency were sent postal questionnaires at four timepoints in 2 years. Ordered logistic regression analysis investigated associations between self-reported mental health outcomes (emotional wellbeing and psychological distress) and social determinants (ability to understand English, frequency of contact with relatives in the UK, satisfaction with accommodation, employment, suitability of employment for qualifications and skills, money management, and physical or verbal attacks in past 6 months) with questions previously validated in refugee populations when possible. Models were adjusted for demographic factors (eg, sex, age, country of origin, and educational level) in the cross-sectional analysis only, because points at which social factors were calculated varied in the original study, so adjustment in calculation of longitudinal variables would have been done with a factor measured at a different time. Findings5678 refugees responded to the survey at baseline—70% of all new refugees in the period. Most (3575 [64%]) were men and younger than 35 years (3915 [71%]). The most common countries of origin were Eritrea, Somalia, and Iraq. 939 refugees (17%) were followed up to the conclusion of the study. In the cross-sectional analysis, refugees who were unemployed in the UK (odds ratio [OR] 2·26, 95% CI 1·58–3·24; p=0·001), could not speak English well (1·75, 1·37–2·22; p=0·001), had infrequent contact with social support networks (4·66, 1·89–11·49; p=0·001), perceived their job as of a lower level than their skills and qualifications (2·22, 1·49–3·33; p=0·001), found money management difficult (5·88, 4·00–8·33; p=0·001), were unsatisfied with their accommodation (3·89, 2·81–5·37; p=0·001), or had been a victim of physical or verbal attack (2·86, 1·67–5·00; p=0·001) had significantly higher odds of poorer mental health outcomes than did others. In the longitudinal analysis, decreased satisfaction with accommodation (1·78, 1·32–2·41; p=0·001), a job with lower than skills and qualifications than the participant had (2·34, 1·15–4·75; p=0·019), and increased difficulty in money management (1·50, 1·09–2·07; p=0·012) were significantly associated with worsening mental health outcomes. Becoming employed (0·49, 0·26–0·92; p=0·026), decreased difficulty in money management (0·65, 0·48–0·88; p=0·005), and no longer being the victim of physical or verbal attack (0·58, 0·36–0·93; p=0·024) were significantly associated with improvement in mental health outcomes. InterpretationPost-displacement social factors such as employment, financial status, and satisfaction with accommodation were important determinants of refugee mental health. Changes in these social determinants could improve refugee mental health. Public health policies and interventions could target vulnerable refugee subpopulations and seek to tackle modifiable post-displacement social factors. A key limitation of this study was the self-reported nature of the variables, introducing the potential for reverse causality. Additionally, new refugees are a difficult population to retain for longitudinal follow-up; loss to follow-up was substantial but similar to that in previous studies. FundingNone.