Data for this analysis came from the 2019–2020 GDHS,⁴⁷ conducted between November 2019 and March 2020 by The Gambia Bureau of Statistics (GBoS). The GDHS is a repeated cross-sectional study that aims to provide up-to-date estimates of basic demographic and health indicators. We specifically draw on the Woman’s Questionnaire, which provides detailed data on healthcare utilisation and perceived barriers to access among female respondents. The GBoS uses a stratified two-stage sampling design to ensure that results are representative at the national, urban and rural levels, and at the local government area levels.⁴⁷ In total, 7025 households were selected for the survey, and 11 865 female respondents aged 15–49 years were eligible for the interview. It is important to note that ‘female’ and ‘woman’ are used interchangeably in this article, as the 2019–2020 GDHS collected data on female reproductive health issues in individuals who identified as women. For our analysis, we excluded individuals who reported being visitors to the household and/or not residents of the city, town or village where the interview took place (404 observations), leading to a raw sample of 11 461 participants. However, the survey-weighted count, which accounts for the sampling design, adjusts this number to 11 455 participants.

Given that migration status is not randomly assigned, migrants usually exhibit different characteristics than non-migrants. Failing to recognise this lack of overlap in the distribution of covariates of migrants and their non-migrant counterparts may lead to extrapolation.⁵³ Our empirical approach relies on the propensity score (PS), a covariate balancing tool,⁵⁴ to draw controlled descriptive comparisons between migrants and non-migrants, as recommended by Li and Li.⁵⁵ Specifically, we make use of PS weighting with overlap weights (PSOW),⁵⁶ an innovative approach enabling us to concentrate on the subset of migrants and non-migrants that exhibit the highest amount of covariate overlap.⁵⁷ Analyses were undertaken using Stata/SE 17.0 (StataCorp LLC, College Station, Texas, Unites States [US]).

Table 1 presents survey-weighted descriptive statistics for eligible respondents, disaggregated by migration status: non-migrants (32.69%), internal migrants (57.12%) and international migrants (10.19%). Healthcare use was highest among international migrants, 72.36% of whom reported visiting a health facility in the last 12 months, followed by internal migrants (69.80%) and non-migrants (60.42%). Barriers related to geographic and financial accessibility were frequently reported across all groups. Approximately one-quarter of respondents in each group cited distance to a facility as a substantial problem: 26.02% of non-migrants, 25.83% of internal migrants and 21.47% of international migrants. Financial barriers were even more commonly reported: 28.09% of non-migrants, 26.81% of internal migrants and 26.46% of international migrants indicated that obtaining money for care was a substantial problem. The safety barrier was also reported as a substantial concern by 11.85% of international migrants, 14.09% of internal migrants and 15.37% of non-migrants. In contrast, a small proportion of women reported needing permission to seek care as problematic, with the lowest rates among international migrants (3.37%) and slightly higher among internal migrants (5.29%) and non-migrants (4.20%). Covariate profiles differed substantially between migrant and non-migrant groups, with international migrants exhibiting lower educational attainment and internal migrants displaying higher levels of economic activity and wealth. One notable pattern in the descriptive data was the higher proportion of internal migrants living in urban areas: 80.48%, compared to 62.82% of non-migrants and 71.14% of international migrants.

When looking at migrant groups broken down by time spent at destination (Panel B, Table 1), 27.79% were recent internal migrants, 29.33% settled internal migrants, 4.62% recent international migrants and 5.56% settled international migrants. Only minor differences in healthcare use were observed across migrant groups, ranging from 68.87% among recent internal migrants to 72.94% among recent international migrants. Geographic barriers were reported more frequently by settled than recent migrants, for both internal (27.57% vs. 24.03%) and international (25.46% vs. 16.68%) migrant groups. Financial barriers were reported at similar levels across migrant groups, ranging from 24.79% to 28.72%, and safety-related barriers also showed comparable rates, ranging from 11.27% to 14.55%. Permission barriers were reported at lower levels across all subgroups.

Descriptive statistics and standardised differences calculated in the PSOW-weighted data can be found in Appendix 1 (broad categories) and Appendix 2 (5-year threshold).

Table 3 shows results using a threshold of 3 years to differentiate between recent and settled migrants. Final models now revealed a higher frequency of healthcare use among both recent internal (aOR = 1.14, 95% CI [1.00; 1.31]) and recent international migrants (aOR = 1.42, 95% CI [1.02; 1.98]). As such, it appeared that the most recent migrants are more likely than non-migrants to use healthcare services. The findings concerning the different barriers to healthcare faced by internal migrants persisted, except in the case of getting permission, which appeared to remain a concern primarily for settled internal migrants (aOR = 1.54, 95% CI [1.23; 1.94]), though not for recent internal migrants. Notably, with the alternative threshold, recent international migrants reported this same barrier as being less problematic (aOR = 0.42, 95% CI [0.17, 1.03]) compared to non-migrants, indicating a noteworthy change in their experience regarding permissions within the shorter migration time frame. Results remained robust when utilising alternative definitions of migrants and non-migrants (Table 4), whereby the internal migrant group strictly encompassed individuals who had moved across regions, while the non-migrant group comprised both women who had never moved and those who had moved within a region of The Gambia. Internal migrants still found obtaining permission (aOR = 1.23, 95% CI [0.99, 1.53]), getting the money needed for treatment (aOR = 1.18, 95% CI [1.07, 1.31]) and the distance to a health facility (aOR = 1.21, 95% CI [1.09, 1.35]) to be problematic issues to overcome.

This study is the first to examine differences in healthcare use and barriers to healthcare between female migrants and non-migrants in The Gambia. Using PSOW to focus on the most comparable individuals across migrant statuses, we find no differences in healthcare use between non-migrants and internal and international migrants overall. However, when accounting for time spent at the destination, recent migrants, both internal and international, demonstrated higher healthcare utilisation than non-migrants. In addition, we found that internal migrants reported significantly greater barriers related to permission, safety, financial cost and geographic access compared to non-migrants, and these healthcare barriers did not diminish with time spent at the destination. By contrast, we found no significant differences in perceived barriers to healthcare between international migrants and non-migrants.

The finding that recent migrants use healthcare services more frequently than non-migrants contradicts our disruption hypothesis, which anticipated reduced use following migration because of instability and access constraints. Two main explanations may help account for this pattern. Firstly, recent migrants (whether internal or international) may exhibit distinct or heightened health requirements that motivate them to seek healthcare services more frequently upon arrival at their destination. The process of adjustment to a new country, along with cultural and environmental changes, can lead to stress, mental health challenges and increased vulnerability to various health issues.¹⁰ Pre-existing health, migration-related health risks and limited access to preventive care can further compound these needs.²³ Secondly, it is also possible that migration constitutes a strategy to access healthcare if health services are scarce or difficult to reach in one’s origin community.^16,25,60 This may be particularly the case for individuals migrating to The Gambia, where out-of-pocket healthcare costs are lower than in many neighbouring nations.⁶¹ The trend towards urban living that we observed among internal migrants may support this interpretation, as urban areas typically offer greater primary healthcare coverage in The Gambia.⁶² The higher reliance on healthcare is particularly notable for recent internal migrants, as we simultaneously find evidence that they face significant barriers to healthcare compared to non-migrants. This indicates that recent internal migrants may employ innovative strategies to surmount these obstacles, driven by the urgency of their healthcare needs and the hurdles they confront. Note that the barriers considered are perceived, meaning that individuals with more urgent and immediate health needs may regard geographical, financial, security and authorisation-related factors as less prominent concerns.

While healthcare use converged between settled migrants and non-migrants in our study, internal migrants continue to experience elevated barriers to healthcare access, indicating only partial support for our adaptation hypothesis. Notably, settled internal migrants encounter an additional barrier related to security when attempting to access healthcare facilities. The absence of adaptation aligns with Cotton,²² who did not find compelling evidence for a positive adaptation effect in maternity healthcare access across different internal migration streams (rural–urban, urban–urban and urban–rural). This pattern may be explained by a higher prevalence of health issues among settled internal migrants that are not adequately addressed by a health facility visitation analogous to that of non-migrants. Alternatively, similar use levels may reflect widespread dissatisfaction with healthcare services, which may lead both groups to seek care only when absolutely necessary. Indeed, the World Health Organization (WHO)⁶² reports that the primary health care system in The Gambia has deteriorated over time and no longer adequately serves the population. The healthcare sector still faces challenges such as limited reach in rural areas, high out-of-pocket expenses and issues related to maternal and child health, along with a shortage of skilled healthcare personnel.⁶²

Surprisingly, no significant differences in healthcare barriers were observed between international migrants and non-migrant counterparts, despite expectations that migrants might encounter greater challenges in accessing healthcare, especially if their place of origin markedly differs from their destination.²² This contrasts with findings by Arnold et al.,¹⁷ who report prominent affordability concerns among international migrants in the Kenyan context. Moreover, we found that as time passes, settled international migrants even exhibited a greater likelihood of having the financial means for medical advice or treatment, in contrast to non-migrants. This may be explained by the financial support (cash-based interventions) that international migrants receive from international organisations such as the International Organization for Migration.⁶³ As The Gambia’s healthcare system is relatively inexpensive compared to neighbouring countries,⁶¹ it is also possible that international migrants have a higher willingness to pay for health services than non-migrants and may not perceive the healthcare costs as a significant concern.

While our findings highlight important disparities in healthcare access between internal migrant groups and non-migrant women, the prevalence of high barriers to healthcare across all groups underscores the need for comprehensive policy interventions to improve healthcare access for all women in The Gambia. Financial and geographic accessibility emerged as the most widespread challenges, concerning over 20% of women across groups. Acceptability-related barriers, particularly those linked to safety concerns, were also found to be problematic for around 10% – 15% of migrant and non-migrant women. Policy efforts are already underway to improve financial accessibility, as The Gambia’s 2021–2030 National Health Policy and The Gambia National Health Strategic Plan 2021–2025 support universal health coverage and aim to progressively deliver quality health services to all individuals, irrespective of factors such as age, nationality, political affiliation, ethnicity, religious beliefs, or sexual orientation.^62,64 To improve geographic accessibility, we recommend expanding Village Health Worker training programmes,⁶² deploying additional mobile clinics with optimised routes and schedules to reach underserved communities efficiently and reliably,^65,66 and leveraging mHealth solutions - such as mobile phone-based appointment systems or health information services.⁶⁷ These policies may simultaneously alleviate safety concerns by reducing the need to travel long distances alone and increasing healthcare access in familiar, trusted environments. Further research is needed to evaluate the effectiveness of ongoing reforms and emerging policies and to monitor evolving patterns in women’s health needs across migrant and non-migrant groups.

Our study is subject to limitations. Because we had limited migration-related information at our disposal, we may have inadvertently included ‘return migrants’ – individuals who moved abroad and then returned to their home country – in our categorisation of migrant groups. This could bias our results towards the null, as these individuals may have previously lived at their destination and, therefore, have an existing network and knowledge they can leverage to access local healthcare resources. In addition, we were unable to explore potentially important variations in healthcare access based on migration pathways, such as rural–urban versus urban–urban migration, or with respect to countries of origin, for example, because of differences between English-speaking and French-speaking countries, as well as differences based on the reasons for migration. Hence, it is possible that the relationship between migration and healthcare access is significant for these subgroups; however, this effect may not be readily apparent. Furthermore, we relied on subjective data regarding the accessibility of healthcare services, which could vary depending on a woman’s perception. While personal experiences and perceptions may play a role in shaping behaviour, it is important to acknowledge that this information represents a proxy for healthcare access, and objective metrics, such as the presence, facilities and availability of healthcare, are not taken into consideration. Note also that we did not differentiate between formal and informal healthcare sectors. Migrants may also face additional barriers to healthcare that were not captured in our data. In addition, PSOW-weighted results are specific to the overlap population, comprising observations with a high probability of belonging to either the migrant or non-migrant group,⁵⁷ which may limit broader generalisability.⁶⁸ Finally, this study focused on comparisons between migrant groups and non-migrants; direct comparisons between internal and international migrants were beyond its scope and warrant future investigation.

This study also has important strengths. To our knowledge, this article is the first to investigate differences in healthcare access between migrant and non-migrant women in The Gambia. Our study distinguishes itself from previous literature in the broader context of sub-Saharan Africa, as it goes beyond the scope of maternal and child healthcare by evaluating women’s access to healthcare across the broader life course. Unlike a body of literature that relied on previous rounds of the Demographic and Health Survey (DHS), which only captured data on residence type (urban or rural) to identify internal migration patterns, the dataset from 2019 to 2020 considered in this study included more precise information on the location of previous residence, enabling us to derive an international migrant group. As such, our study provides a quantitative assessment of international female migrants’ barriers to healthcare, enriching the evidence that originated from other contexts^69,70,71,72 or coming from small samples and lacking a gendered perspective.¹⁷ Furthermore, we utilise a novel PS weighting approach with OWs to compare migrants, whether internal or international, with non-migrants. This method allows us to account for the lack of overlap in the distribution of covariates of migrants and their non-migrant counterparts and thus avoid extrapolation.⁵⁵ Besides Atake⁵⁰ and Porth et al.,³⁵ who both examined the balance of covariates before and after accounting for confounding, most studies exploring associations between migration and health failed to recognise this selection bias. The literature on confounding adjustment often puts forth the inverse probability weighting (IPW) technique.^57,73 Yet, this method suffers from extreme values when the PS distributions between treatment groups have little overlap,⁷³ which can lead to biased estimates and excessive variance.⁵⁷ Because migrant cohorts exhibited a markedly different covariate distribution compared to non-migrant cohorts, as observed in our descriptive statistics, we chose the newly developed PSOW, which overcomes the issue of a lack of overlap between treatment groups.⁵⁶ Overlap weights have been shown to improve balance and precision compared to IPW⁵⁷ and the PSOW is more robust to misspecification of the PS model.⁷³ To our knowledge, this article is also the first to implement this innovative approach to analyse migration-related disparities in access to healthcare services.