Influxes of migrant women of childbearing age to Western receiving countries have made their reproductive health a priority in those countries. Yet, providing optimal care to these women may be hampered by an inadequate volume or quality of research to inform practice. We reviewed reports of studies recently published to assess the extent to which current research is able to inform reproductive health care practices for migrant women (i.e., those born in countries other than the receiving country) – in so doing, we sought to offer a view of the landscape from which clinicians may interpret relevant publications. Additionally, we sought to identify topics for which clinicians may choose to advocate for additional research to be performed.
Highlights
- •
Many migrant reproductive health reports are inadequate to inform clinical practice.
- •
Comparative migrant reproductive health reports are concentrated in the USA, Spain, and Italy.
- •
Recommended migration indicators related to reproductive health remain underreported.
- •
Quality studies to inform clinical practice are available for some pregnancy-related care.
- •
Non-pregnancy-related reproductive health outcomes in migrant women warrant further investigation.
Background
By the end of 2013, Western nations received over 130 million international migrants with women constituting more than half (56.1%) of this group and care for migrants has recently been declared to be a priority for research and action . Studies in the industrialized West have recorded health inequalities between migrants (including immigrants, refugees, asylum-seekers and undocumented persons) and receiving-country nationals . These inequalities encompass disparities in perinatal health outcomes and care and to a lesser extent, other reproductive health and care outcomes . In light of the increasing proportion of births to migrant women , these disparities and the reproductive health status of migrant women at large take on added importance.
Although the health of migrants has arrived on the policy agenda across a number of settings, the research available to support effective clinical decision-making is by no means conclusive . One of the key reasons for this has to do with methodological differences in choices made by investigators when developing their research studies . Each design decision can affect the results obtained, making it difficult for both the direction and estimates of effects to be similar across studies. The consequent variation in results may lead to confusion or often a perception that results are too complicated to be understood and put into practice.
Some of the broad design issues which could affect results are those included in assessments of study quality such as: group comparability, adequacy of follow-up, appropriate variable measurement, ‘exposure’ definition, and control for potential confounders.
There exist key methodological issues specific to the examination of the reproductive health of migrant women and these are driven by the study question. The basis of all scientific inquiry is conceptual – the scientific issue in question. All study design issues are driven by this question and require selection of the most rigorous approaches to answering it. Once the basic study design is chosen, the underlying hypothesis being tested will direct how the sample should be created. For example, is it a difference in genetic make-up that is suspected as the mechanism through which the outcomes occur, in which case ‘ethnicity’ as defined by lineage might be the appropriate way to select the sample, or are cultural practices thought to be the mechanism, in which case degree of “connection” to the culture of origin may be the most appropriate way to select the sample? Is eligibility for governmental subsidy of needed services thought to be the mechanism? If so, defining study groups based on legal immigration class (e.g., undocumented/documented) may be most useful. Variations in hypotheses such as these can be extended to other migration-related variables such as length of time in receiving country, documentation of health history, and language fluency. Prior to changing clinical practice, repeatedly similar results testing the same hypothesis in a range of settings obtained through high quality inquiry is reassuring, if not essential.
In some studies, migrant populations are characterized by cross-border movement, while in others, migrants are characterized by their reported ethnicity. It is also possible that studies that control for social factors attain different outcomes from those that do not. Lower socio-economic status, for example, is more common among migrant families and is correlated with a higher prevalence of maternal and child health complications .
Differences in study outcomes may also signify heterogeneity within migrant groups. For instance, many individual risk profiles can be merged into the single groupings, ‘migrant’ or ‘receiving country-born’ . For example, refugee and asylum-seekers often face greater risk of poor outcome because of a history of gender-based violence , post-traumatic stress disorder and official language difficulties . Increased risk of infectious diseases and poor maternal nutritional status are also commonly observed in women who have resided in refugee camps and war-torn areas . In contrast, economic migrants may be richer and in better health than other migrants, and their data may well outshine those of higher-risk and less numerically visible migrants, such as refugees and asylum-seekers . The ‘healthy immigrant effect’, in which newcomers exhibit better health outcomes than their receiving-country counterparts and subsequently worsen over time , is cited to account for cases where migrants have better or similar health outcomes.
An accurate overview of the discrepancies that exist between study methodologies is important to know to help understand conclusions to be drawn from the combined body of literature on the reproductive health of migrant women. We therefore conducted a systematic review of recent literature (2010–2014) making particular note of methodological issues that might impact conclusions drawn in individual studies and consequently, lead to an inability to draw conclusions upon which to base clinical practice.
Research questions
- 1.
To what extent is current research able to inform reproductive health care practices for migrant women (i.e., those born in countries other than the receiving country) living in Western industrialized countries?
- 2.
Which reproductive health topics require additional research to inform clinical practice with migrant women living in Western industrialized countries?
Methodology
Criteria for considering studies for review
Study outcomes
Reports were excluded if the outcome under study was not directly related to reproductive health indicators commonly cited by national or international [World Health Organization (WHO) Reproductive Health Indicators, 2006, Euro-Peristat, 2012] bodies or to reproductive health-related behaviours (e.g., contraceptive use, termination of pregnancy, sexually-transmitted infections (STI) screening, mammography) or cancers.
Type of exposure
Migration was the exposure of interest. Therefore, any report of women who had migrated was initially included. Reports were subsequently excluded if international cross-border movement was unlikely [thus women from ‘protectorates’ such as Puerto Rico and second-generation women (born inside the receiving country with at least one parent born outside the receiving country) did not meet our criteria].
Type of study designs
We excluded case studies, clinical reports, reports without a comparison group and reports in which the results of the migrant groups were not presented separately from the comparison group.
Study population
Migrant women in Organization for Economic Co-operation and Development (OECD) or European Union (EU) 27 countries were included. Those in refugee camps were excluded.
Search strategy
We located an initial set of published reports by searching electronic literature databases from January 2010 through December 2014 using Medline, Embase, PsychINFO, Global Health, Social Work Abstracts, CINAHL, Web of Science, and ProQuest Dissertations and Theses. The search strategy ( Appendix A ) was developed in conjunction with a McGill University health sciences librarian.
Searches were supplemented with bibliographic citation hand-searches of all systematic reviews which met our outcome and inclusion criteria. We were unable to review government documents or to contact authors of reports for feasibility reasons.
No language exclusions were routinely applied. Review of non-English/French language articles was restricted to their English/French abstracts.
Review methods
Study assembly and quality assessment
Duplicate reports and titles unrelated to reproductive health were eliminated. Abstracts were then reviewed and eliminated if exclusion criteria were met. Four-hundred eighty-one reports (full text articles and abstracts) were reviewed for study inclusion. Six percent (n = 15) of 253 excluded articles were blindly reviewed by a second individual who also excluded them, giving evidence for intra-rater reliability. Inclusion criteria were met in 228 reports (140 full-texts and 88 abstracts; See Appendix B for full references). All reports that met the inclusion criteria used quantitative methods, therefore no qualitative reports were included in the final sample.
Studies were assessed for quality using the US Preventive Services Task Force (USPSTF) criteria for cohort and case-control studies ( Appendix C ). One author assessed the quality of all studies, and four others conducted blinded assessments of a 2–6% sample; no discrepancies between raters were found in the overall quality scores.
Data coding
Data sources were classified into the following categories (based on frequency of occurrence in the literature) : population-based routine data registries, including linked birth/death certificates and birth/maternity service registries ; population-based surveys ; population-based hospital records (covering a large representative proportion of a population) ; research studies that were not included in any of the categories above (using questionnaires, interviews) and hospital records.
Additional data recorded were: year of publication, language of publication, receiving country, number of migrants studied, years of database (i.e., every year for which it included any data), geographic coverage of database (i.e., international, national, regional, or local), covariates considered in the analyses, and measures of association of the outcome with the migration indicator (e.g., adjusted odds ratios).
Migration indicators were classified into general categories (based on frequency of occurrence in the literature) and operationally defined as follows : country of birth/foreign-born: any category which required data on country of birth to define it (subsequently grouped into regions – e.g., ‘Asia’ includes all Asian sub-regions) ; ethnicity: an undefined term used by authors, which could include ethnicity, ethnic group, ethnic mix or race ; nationality: a commonly undefined term used by authors, which could include national origin, citizen, citizenship ; ‘foreigner’: a term used by authors, it is usually unclear if country of birth, ethnicity or nationality is used to define the term; includes an undefined ‘immigrant’ ; language: any category which requires data on language to define it ; refugee: the term as used by authors may include those who left home unwillingly or were in resettlement camps ; immigrant status: the term as used by authors may include categories such as ‘undocumented’, ‘illegal’ or ‘irregular’; and asylum-seeker.
Reproductive health data were classified as outcomes if they were a main or secondary focus of the report or were reported as sample descriptors. Outcomes were classified into the following general categories, based on frequency of occurrence in the literature and clinical relevance : gestational age/preterm birth, including any outcome that requires gestational age to define it ; small for gestational age, including any outcome that requires birthweight combined with gestational age to define it ; mode of delivery, including caesarean birth and operative vaginal deliveries ; feto-infant mortality, including neonatal, infant, and ‘spontaneous abortion’ ; maternal or infant infection, or absence thereof, including but not limited to Human Immunodeficiency Virus (HIV), toxoplasmosis, STIs and rubella sero-negativity ; life-style factors, including cigarette/alcohol/drug use, and preventive intake of folic acid ; prenatal care defined in any way ; maternal health, including mortality, pregnancy-related morbidity and other maternal outcomes not included in the previously defined categories, such as extended length of labour and episiotomy ; congenital anomaly and infant morbidity not included in the previously defined categories, such as neonatal intensive care unit admission and low Apgar score ; family planning including contraceptive use, unintended pregnancies, termination of pregnancy ; reproductive cancers ; menopause ; sexually transmitted infections ; sexual health behaviours ; hospital and health service use for obstetrical and gynaecological reasons ; perceptions of maternity care ; maternal mental health and social support ; intimate partner violence (IPV)/sexual violence.
Analyses
Studies were described with respect to the variables defined previously.
Reproductive health results reported by authors for each study were subsequently coded as statistically worse, better, not different, or mixed, when we examined results for all subgroups of migrants compared to receiving-country women within each study. These results were coded for each of the relevant reproductive health outcome categories listed above. In the case of any discrepancies between the text and tables, the text results were coded.
Study sample
Fig. 1 shows the evolution of the study sample . Of over 14,000 reports, we included 228 (full text or abstracts) on 18 reproductive health outcomes within which 463 comparisons had been made.
Results
Fig. 2 shows the migrant-receiving countries represented in the reports and highlights an unequal distribution, with the U.S. contributing to over one-quarter of the sample; Spain, one-sixth; and Italy, one-tenth. Significantly fewer studies were reported from 19 other countries. Fig. 3 shows the database years represented in the sample. Data originated from 1958 through 2013, with the greatest proportion of data originating in the years closest to 2004. This suggests that results of studies reported in the last five years would need to be carefully assessed for application to current migrant movements in the receiving-country of the clinician since large population movements are associated with political turmoil and/or environmental catastrophes which erupt at different points in history.
There was a range of types of databases represented in the sample, including: population-based registries (38.0%), population-based surveys (8.4%), population-based hospital records (8.4%), other hospital records (15.2%), and research studies (27.2%), with a small number (2.8%) not describing the database employed. Taken as a whole, the vast majority of this body of literature originates from databases representative of a larger group, making extrapolating conclusions from them to other similar populations that a clinician is likely to encounter more reasonable. Geographic coverage of study sampling frames lends additional support to this inference since the majority of reports covered national (39.0%) or regional (24.1%) areas. One-third (34.2%) however, were limited to local samples with smaller proportions, either international (1.8%) or not described (0.9%). Taken together, these reports included approximately 12 million migrants. Well over half of all reports were of relatively small samples. Thirty-seven per cent of the 228 studies reported on a sample of less than 1,000, 21.1% reported on a sample of less than 250.
An international Delphi survey outlined the imperative of collecting data on five essential indicators of migration when monitoring or conducting research on perinatal health with migrants. These indicators were variously reported in our sample, together with other “migrant” definitions as can be seen in Fig. 4 . Two-thirds used “country of birth”, or the less precise term “foreign-born”. Unfortunately, 14.8% used the term “foreigner” – too imprecise to help clinicians direct their resources. The maximum number of reported recommended migration indicators in a single report was three (of five); this occurred in 2.2% of all 228 reports.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
