Being an important social determinant of health, change of residence or migration has long received attention in both public health and the social sciences with gradually increased emphasis on the link between migration and the utilization of sexual, reproductive, maternal, and child healthcare services (1). The existing evidence indicates the intertwined and complex relationships between population movement and healthcare (2). The decision to migrate can be influenced by health considerations, while the migration experience itself can have significant impacts on the health outcomes of migrants, as well as those who remain in the place of origin and even those who host the migrants in the destination location. Studying healthcare in the context of migration provides a comprehensive understanding of the intricate and heterogeneous nature of the migration process (3). With the surge in migratory movements globally, the health needs and well-being of migrants have become a pressing issue that demands more attention (4, 5).

Moving to a new residence can have a direct impact on an individual’s health, which can either improve or deteriorate health status depending on various factors including changes in healthcare accessibility, exposure to new environments, lifestyle adjustments, and sociocultural differences. Among migrants, females are often considered more vulnerable in terms of reproductive health issues compared to their male counterparts (6, 7). The absence of sexual, reproductive, maternal, and child healthcare services has been cited as one of the top three health issues confronted by internal migrants, alongside communicable and occupational diseases (5, 8). Migrant women in Bangladesh, like many other countries, also encounter multiple challenges in accessing and utilizing adequate maternal healthcare services (3).

The access to and utilization of maternal healthcare services among Bangladeshi women differ by several factors such as place of residence, lack of women’s education, women and husband’s occupation, household poverty, higher birth order, women’s attitude toward domestic violence, geographical obstacles, climate displacement, internal migration, etc. (9–11). Although national-level data from the Bangladesh Demographic and Health Survey and Urban Health Survey indicate remarkable progress in maternal health indicators, disparities in access and utilization were found to be particularly prevalent in urban areas, characterized by high levels of poverty, inequality, and inadequate infrastructure (12, 13). With 40.5% of the total population living in urban areas—a figure expected to double by 2050, and internal migration reportedly accounting for 66% of this urban growth, the role of internal migration in urban growth and its implications for maternal healthcare service utilization (MHCSU) becomes even more critical to understand (14–16).

This issue is particularly critical given the disparities in MHCSU among different population groups, with migrants potentially facing unique challenges. Current literature indicates variations in MHCSU between migrants and non-migrants, particularly in ≥4 ANC and ID in the context of Bangladesh and India (9, 11, 12, 17–20). Nonetheless, these insights are often limited as they concentrated predominantly on urban areas and small sample sizes and were not nationally representative. Furthermore, these studies generally did not comprehensively explore how different types of internal migration streams influence MHCSU. Our study aims to fill this research gap by examining the effect of various internal migration streams on ≥4 ANC and ID service utilization, as recommended by WHO (21). After controlling for individual and household characteristics using multivariate regression analysis, we specifically aimed to ascertain whether women who participated in various migration pathways (rural-to-rural, rural-to-urban, urban-to-rural, and urban-to-urban) showed distinct patterns of maternal healthcare use in comparison to non-migrant women in both rural and urban settings.

Table 1 presents the background characteristics of female internal migrants and non-migrants aged 15–49 who had given birth in the last 3 years. The weighted sample population included a total of 3,453 (86.1%) migrants and 554 (13.8%) non-migrants. The highest proportion of women possessed secondary education (49.4%), followed by primary education (28.7%). The proportion of women with more than secondary education was notably higher than women with no education at all. Regarding working status, the majority of women were not working (57.8%), while the rest were employed (42.2%). The study population was relatively evenly distributed across the wealth quintiles. The largest proportion of women fell within the “poorest” category (21.8%), followed by the “poorer” category (20.5%). The “middle” category (19.8%) had the third-largest proportion of women, followed by the “richer” category (19.5%) and the “richest” category (18.4%).

As far as awareness regarding community clinics was concerned, 61.5% of the individuals were aware of any community clinic in their area. Besides, the table indicated that accessibility to healthcare was problem for some individuals in the study population. Approximately 26.0% of the women reported ‘yes’ to distance and money for treatment being an issue to access healthcare. Furthermore, the data showed that 63.3% of the women were exposed to mass media. The table also presents information on the level of women’s autonomy in household decision-making. The table indicates that 64.5% of women reported moderate autonomy in household decision-making, while 28.1% reported low autonomy, and 7.20% reported high autonomy. In terms of place of delivery, 51.7% of women had delivered birth at home, whereas 48.3% delivered at a health center. Besides, 54.2% of women had utilized less than 4 ANC and 45.8% have utilized four or more ANC.

This is one of the early studies in the context of Bangladesh that has operationalized multiple migration-related variables aimed at an extensive comprehension of the relationship between migration and MHCSU using a nationally representative cross-sectional dataset. The findings revealed that the likelihood of delivering birth at a health center was higher among the types of migration streams in comparison to the rural non-migrants, net of all control variables. The study revealed broadly comparable results concerning the utilization of at least four antenatal care (ANC) services, with two distinct trends. Firstly, the positive effect of all migration streams on four-plus ANC utilization was generally smaller in magnitude than institutional delivery care, and secondly, individuals who migrated from rural-to-rural areas exhibited lower likelihoods of seeking ≥4 ANC services than the rural non-migrants although this finding is not statistically significant in the fully adjusted model.

These findings align with conclusions from Thapa et al. (7) in Nepal, who also found that migration increased the probability of utilizing maternal healthcare services compared to rural non-migrants, and that the magnitude of this positive effect was higher for institutional delivery than for ≥4 ANC utilization (7). Our study also suggested that migration involving urban areas (to, from, or within) offered distinct advantages over purely rural migration streams. There appeared to be a notable advantage for women who migrated to or reside in urban areas, as well as those who move from urban areas, as opposed to their rural counterparts. This finding also found consistent in line with the study from Nepal (7).

After accounting for potential indicators of migrant selectivity in Model 2, the findings indicated that the estimated effect of migration streams on institutional delivery was reduced across all categories compared to Model 1. Furthermore, the effects became statistically insignificant for migration originating from rural areas (both rural-to-rural and rural-to-urban streams). A similar pattern was also observed for ≥4 ANC utilization, where after adjusting for selection factors, only urban-to-urban migrants were still significantly more likely to utilize services than rural non-migrants; the associations for all the other categories of migrants become insignificant.

This observed reduction in effect size after controlling for selection factors suggested that migrant selectivity likely played a significant role in the initial association observed between migration and MHCSU in Model 1. Thus, one reason for the higher likelihood of service utilization initially observed across migrant streams compared to rural non-migrants could be inferred: the migrant population may have possessed certain attributes (e.g., higher education, greater wealth, more autonomy) that made them more predisposed to utilizing maternal healthcare services, independent of the migration experience itself. However, it is important to note that even after adjustments, urban non-migrants and urban-to-urban migrants often remained significantly more likely to utilize services compared to rural non-migrants, aligning with results reported by Cotton in an African country context (22).

In the fully adjusted Model 3, the effect of the migration stream was further reduced for all of the categories of migration streams compared to Model 2 and becomes insignificant for migrants from urban to rural areas. Almost no difference was observed in the findings of ANC utilization between the second and third models. Even after controlling for the factors that affect maternal healthcare services utilization, urban-to-urban and urban non-migrants still had more likelihood of utilizing ≥4 ANC and institutional delivery compared to rural non-migrants.

Thus, following a meticulous inquiry into the effect of migration on MHCSU, while controlling for various confounding variables through sequential regression models, the findings of this study established a significant association between migration and the MHCSU. However, there was a remarkable trend within categories of migration streams while exerting its effect on maternal healthcare services utilization. Migration had a significant and positive association with MHCSU only in two scenarios: when migrating from an urban to a rural area (as in model 2), and when relocating between urban areas (as in model 3).

The superiority of urban migrants in terms of MHCSU is also reflected in studies in other country contexts (5, 23). This may be explained partly by an independent ‘urban effect’ on MHCSU, potentially related to the greater availability and accessibility of healthcare facilities and skilled medical professionals in urban destinations, irrespective of migrant selectivity. Migrants residing in or moving between urban areas may be more likely to access and utilize available healthcare services compared to those in rural areas. This suggests that the migrants in urban areas enjoy the ‘urban advantage’ in healthcare provision, which may drive the differential utilization of healthcare between urban and rural migrants.

The finding that urban-to-urban migrants had consistently high odds of MHCSU in both Model 2 and Model 3 may indicate that this group performed particularly well compared to rural non-migrants. However, it should be noted that these odds were calculated with rural non-migrants as the reference group. When comparing descriptive statistics or considering the odds ratios relative to urban non-migrants (who often served as the group with the highest utilization), the utilization level for urban-to-urban migrants typically fell short of that observed among urban non-migrants.

These findings align with previous research on migration and MHCSU, which often finds that among migrant groups, those migrating between urban areas tend to have high odds of utilization. However, similar to the findings of this study, these studies also reported that urban non-migrants have the highest likelihood of utilizing maternal healthcare services (5, 22–25).

The urban advantage in healthcare access is the likely underlying reason behind this. In addition, our bivariate analysis revealed that, within our sample, urban non-migrants exhibited the highest proportions with secondary or higher education, the greatest exposure to mass media, the highest levels of decision-making autonomy, and faced the fewest reported barriers (cost and distance) to accessing healthcare. These advantageous characteristics almost certainly benefited urban non-migrants, contributing to their higher maternal healthcare utilization levels.

There are some limitations exist in this study which need to be taken into consideration when interpreting the results. This study used data from BDHS, 2017–18 which adopts a cross-sectional design that may not be capable of establishing a causal relationship between migration and MHCSU. Consequently, formal testing of migrant selectivity, disruption, and adaptability processes over time was also not feasible. Additionally, reliably sequencing events like migration relative to specific pregnancies was challenging, as the BDHS did not inquire about the precise timing (year or month) of migration events.

This study provided an analysis of the implications of internal migration on maternal healthcare services utilization in Bangladesh. The research addressed a significant gap by identifying diverse migration streams and applying a rigorous analytical methodology to discern their independent effects on MHCSU. This approach yields a comprehensive understanding of the interplay between migration and maternal healthcare service utilization, thereby contributing substantially to the extant literature on this subject. The findings indicated that MHCSU varied notably between migrant and non-migrant groups, with the lowest access observed among rural non-migrant women. In contrast, migration to urban areas appeared to confer certain advantages in accessing and using maternal healthcare services. However, it was noteworthy that all migrant categories, regardless of their origin, exhibited lower use of maternal healthcare services compared to their urban non-migrant counterparts.

These insights underscore the necessity for future research to establish a causal link between migration and health outcomes and to delve deeper insights into the migration-health nexus. Based on our findings, a key policy recommendation is the development of public health strategies that explicitly consider migration status. Enhancing health information systems to capture more nuanced data on migrants’ needs and challenges is crucial. Such data can inform targeted, demand-specific programs. The outcomes of this study have significant implications for policy formulation and program development aimed at improving MHCSU for both migrant and non-migrant women in Bangladesh. Given the study’s finding that rural-to-rural migrants and rural non-migrants have lower utilization rates of maternal healthcare services compared to their urban counterparts, there is a pressing need to improve healthcare availability and accessibility, particularly in rural areas. Furthermore, the distinct challenges potentially confronted by migrants may necessitate healthcare programs tailored to their unique situations (e.g., considering integration challenges, portability of services). Such specialized programs would not only address the immediate healthcare needs of these populations but also contribute to the overarching aim of equitable healthcare access, thus supporting the broader goals of national development and the realization of sustainable development objectives.