Global patterns of mortality in international migrants: a systematic review and meta-analysis

Background: 258 million people reside outside their country of birth; however, to date no global systematic reviews or meta-analyses of mortality data for these international migrants have been done. We aimed to review and synthesise available mortality data on international migrants. Methods: In this systematic review and meta-analysis, we searched MEDLINE, Embase, the Cochrane Library, and Google Scholar databases for observational studies, systematic reviews, and randomised controlled trials published between Jan 1, 2001, and March 31, 2017, without language restrictions. We included studies reporting mortality outcomes for international migrants of any age residing outside their country of birth. Studies that recruited participants exclusively from intensive care or high dependency hospital units, with an existing health condition or status, or a particular health exposure were excluded. We also excluded studies limited to maternal or perinatal outcomes. We screened studies using systematic review software and extracted data from published reports. The main outcomes were all-cause and International Classification of Diseases, tenth revision (ICD-10) cause-specific standardised mortality ratios (SMRs) and absolute mortality rates. We calculated summary estimates using random-effects models. This study is registered with PROSPERO, number CRD42017073608. Findings: Of the 12 480 articles identified by our search, 96 studies were eligible for inclusion. The studies were geographically diverse and included data from all global regions and for 92 countries. 5464 mortality estimates for more than 15·2 million migrants were included, of which 5327 (97%) were from high-income countries, 115 (2%) were from middle-income countries, and 22 (<1%) were from low-income countries. Few studies included mortality estimates for refugees (110 estimates), asylum seekers (144 estimates), or labour migrants (six estimates). The summary estimate of all-cause SMR for international migrants was lower than one when compared with the general population in destination countries (0·70 [95% CI 0·65–0·76]; I2=99·8%). All-cause SMR was lower in both male migrants (0·72 [0·63–0·81]; I2=99·8%) and female migrants (0·75 [0·67–0·84]; I2=99·8%) compared with the general population. A mortality advantage was evident for refugees (SMR 0·50 [0·46–0·54]; I2=89·8%), but not for asylum seekers (1·05 [0·89–1·24]; I2=54·4%), although limited data was available on these groups. SMRs for all causes of death were lower in migrants compared with the general populations in the destination country across all 13 ICD-10 categories analysed, with the exception of infectious diseases and external causes. Heterogeneity was high across the majority of analyses. Point estimates of all-cause age-standardised mortality in migrants ranged from 420 to 874 per 100 000 population. Interpretation: Our study showed that international migrants have a mortality advantage compared with general populations, and that this advantage persisted across the majority of ICD-10 disease categories. The mortality advantage identified will be representative of international migrants in high-income countries who are studying, working, or have joined family members in these countries. However, our results might not reflect the health outcomes of more marginalised groups in low-income and middle-income countries because little data were available for these groups, highlighting an important gap in existing research. Our results present an opportunity to reframe the public discourse on international migration and health in high-income countries. Funding: Wellcome Trust, National Institute for Health Research, Medical Research Council, Alliance for Health Policy and Systems Research, Department for International Development, Fogarty International Center, Grand Challenges Canada, International Development Research Centre Canada, Inter-American Institute for Global Change Research, National Cancer Institute, National Heart, Lung and Blood Institute, National Institute of Mental Health, Swiss National Science Foundation, World Diabetes Foundation, UK National Institute for Health Research Imperial Biomedical Research Centre, Imperial College Healthcare Charity, and European Society for Clinical Microbiology and Infectious Diseases (ESCMID) Study Group Research Funding for the ESCMID Study Group for Infections in Travellers and Migrants. For this systematic review and meta-analysis, we searched MEDLINE, Embase, the Cochrane Library, and Google Scholar databases for studies published between Jan 1, 2001, and March 31, 2017, reporting mortality in international migrants, without language restrictions. Full search terms are provided in the appendix. We chose to search for studies published after Jan 1, 2001, because a previous systematic review of mortality in migrants had been published by this date, but it did not contain a meta-analysis and did not assess outcomes across all ICD-10 categories.3 On Sept 3, 2018, we updated our search using the same databases, search terms, and inclusion criteria. We included observational (cohort and cross-sectional), systematic reviews, and randomised controlled trials reporting quantitative data on mortality in international migrants of any age residing outside their country of birth. We excluded studies that recruited participants exclusively from intensive care or high dependency hospital units, with an existing health condition or status (eg, myocardial infarction, HIV, tuberculosis, pregnancy), or a particular health exposure (eg, smoking, high blood pressure). We also excluded studies limited to maternal or perinatal outcomes. The study with the largest or most representative sample was included, and when these were equal, the most recent study was included. Discrepancies in the inclusion or exclusion of papers during screening were discussed until consensus was achieved, and RWA resolved any final discrepancies. This study was done in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA)14 guidelines. The study protocol is available online. Deviations from the protocol are reported in the appendix. Five reviewers (RWA, SB, ALB, LBN, and PP) screened titles, abstracts, and full texts using Covidence systematic review software. Two reviewers independently examined citations at each stage. We adapted a previously used data extraction form,15 to record study design, year or years of study, country, country of origin, number of participants, standardised mortality ratios (SMRs), absolute mortality rates, and summary descriptions of the study population. Extracted data were reviewed and checked by a second author before cleaning and analysis. Duplicate data were removed for studies reporting information from the same migrant group (by country of destination) for the same mortality outcome and time period. Outcomes of interest were all-cause and ICD-10 cause-specific SMRs and absolute mortality rates. The number of datapoints that presented cause-specific mortality according to ICD-10 groups was also calculated. We report data by ICD-10 disease category, and converted outcomes from studies reporting data using older ICD versions as necessary. Four reviewers (SB, ALB, RB, and PP) assessed the risk of bias of included papers using a piloted quality assessment form adapted from the Newcastle Ottawa Scale.16 A randomly selected sample (10%) of these assessments was corroborated by LBN. We used the metafor package (version 2.0) in the statistical software R (version 3.5.1) and random-effects models to calculate pooled estimates of mortality and corresponding 95% CIs. Heterogeneity was assessed using the I2 statistic, and assessed further in subgroup analyses wherever possible. Mortality point estimates were included in each model with corresponding SEs extracted directly or calculated using CIs for each point estimate. Subgroup analyses were done when appropriate to assess mortality by sex, migrant type (eg, refugee or asylum seeker), World Bank geographical region of origin, World Bank income level of countries of origin, and evidence quality. The study is registered with PROSPERO, number CRD42017073608. The funders of the study had no role in study design, data collection, data analysis, data interpretation, writing of the report, or the decision to submit the paper for publication. All authors had full access to all data in the study and had final responsibility for the decision to submit for publication.