The study, conducted by Qingjun Zhao of Huzhou College, addresses a gap that has persisted in migration and health research. Most previous work on disparities in maternal and child health services among migrants has focused on individual factors: whether a particular mother speaks the local language, whether a specific encounter with a provider went smoothly, or whether she possesses the personal skills to navigate an unfamiliar clinic. What has received far less attention is whether the overall linguistic structure of a destination city, treated as a contextual exposure in its own right, is associated with whether migrants actually receive preventive services. Zhao’s analysis shifts the analytical lens from the individual consultation room to the urban environment as a whole, asking whether cities where many different languages and dialects coexist systematically complicate the delivery of public health information.

To answer this question, the study linked data from the China Migrants Dynamic Survey with measures of urban linguistic environments and city-level socioeconomic characteristics. The China Migrants Dynamic Survey is a large, nationally representative survey of internal migrants, and Zhao restricted the analytical sample to female migrants aged 15 to 49 years who had lived in their destination city for more than six months and had at least one child aged 0 to 6 years. The resulting sample comprised 26,170 respondents distributed across 333 cities, providing substantial geographic and demographic variation. The outcome of interest was receipt of maternal and child health education, a preventive service delivered through China’s Basic Public Health Services program. The primary exposure was the city-level linguistic diversity index, a quantitative measure of how many distinct linguistic groups are represented in a city and how evenly the population is distributed among them.

The statistical approach was deliberately rigorous. Baseline analyses used linear probability models that adjusted for both individual-level characteristics, such as age, education, and employment, and city-level characteristics, while also incorporating province fixed effects to account for unobserved regional differences. Standard errors were clustered at the city level to reflect the fact that mothers within the same city share a common linguistic environment. The estimate that emerged was consistent and statistically robust: higher linguistic diversity was associated with a lower probability of receiving maternal and child health education, with a coefficient of negative 0.1014 and a standard error of 0.0329, significant at the one percent level. In practical terms, a one-standard-deviation increase in the linguistic diversity index corresponded to an average 2.12-percentage-point lower probability that a migrant mother received this preventive education.

What distinguishes this study from much of the observational literature on migration and health is the depth of its robustness testing. Zhao employed entropy balancing to reweight comparison groups, Oster sensitivity analysis to probe vulnerability to unobserved confounding, city-level permutation tests to rule out chance geographic patterns, and alternative model specifications and linguistic measures, including a count of dialect subgroups represented in each city and a linguistic polarization index. The analysis was repeated on restricted samples, with additional adjustment for city administrative status, and with Conley-type spatial standard errors that correct for heteroskedasticity and spatial autocorrelation among nearby cities. Across all of these checks, the negative association between linguistic diversity and receipt of health education held steady, suggesting that the finding is not an artifact of a particular modeling choice.

The study also asked whether the association varied across subgroups of migrant mothers, and the heterogeneity results are among its most striking findings. The negative association weakened as the age of the youngest child increased, reflected in an interaction coefficient of 0.0164 with a false-discovery-rate-adjusted q value of 0.0509, just at the conventional threshold. This pattern suggests that the linguistic barriers to health education are most acute in the earliest period of motherhood, when demand for information about pregnancy, delivery, and newborn care is highest. More pronounced still was the pattern among mothers who had migrated across dialect-region boundaries. For these women, who had moved from one major dialect area of China to another, the negative association was significantly stronger, with an interaction coefficient of 0.1063 and an FDR-adjusted q value of 0.0363. In other words, the mothers who had crossed the greatest linguistic distances within China were the most likely to be missed by the health education system.

To understand why linguistic diversity matters, Zhao used Karlson–Holm–Breen decomposition, a statistical technique for assessing how much of an association between an exposure and an outcome operates through intermediate variables. Four explanatory pathways were examined: awareness of Basic Public Health Services, having a local resident health record, willingness to integrate into the destination city, and affective identification with the city. The results pointed clearly to informational and institutional mechanisms rather than purely psychological ones. Awareness of Basic Public Health Services explained 20.97 percent of the total association, and possession of a local resident health record explained a further 19.46 percent. Together, these two factors accounted for roughly forty percent of the observed relationship. Willingness to integrate and affective identification with the city played smaller roles, indicating that the problem is less about whether migrant mothers feel attached to their new homes and more about whether they know the services exist and are formally registered within the local health system.

These mechanistic findings carry practical weight. A health record establishes a mother’s formal presence in the local public health infrastructure, and awareness of available services determines whether she seeks them out. In linguistically diverse cities, outreach materials, appointment reminders, and face-to-face counseling may be delivered in forms that cross-dialect migrants struggle to comprehend, and the administrative pathways to obtaining a local health record may be harder to navigate for women unfamiliar with local institutional norms and terminology. The result is a quiet erosion of access that occurs not through explicit exclusion but through the accumulated friction of incomprehensible information and unfamiliar procedures. Because the association was strongest for mothers of very young children, the consequences fall on precisely the population for which preventive health education has its greatest documented value, from safe infant feeding practices to timely immunization and developmental monitoring.

The broader implication is a reframing of how universal health provision should be evaluated. China’s Basic Public Health Services program is designed to reach everyone, and on paper it does; the services exist in every city in the study. Yet the research demonstrates that formal availability is not the same as equitable access, and that the characteristics of the destination city itself, independent of any individual mother’s education, income, or language proficiency, shape who is reached and who is overlooked. The author suggests that health service managers in cities with more diverse linguistic structures should pay particular attention to the comprehensibility of health information, the modes through which services are communicated, and the actual language needs of different migrant groups. As internal migration continues to redraw the demographic map of Chinese cities, the study offers a quantitative foundation for a simple but consequential idea: a public health system that speaks only some of its citizens’ languages is not yet a system that serves them all.

Subject of Research: City-level linguistic diversity and migrant mothers' access to maternal and child health education in China

Article Title: From universal provision to equitable access: city-level linguistic diversity and access to maternal and child health education among migrant mothers in China

Article References: Zhao, Q. (2026). From universal provision to equitable access: city-level linguistic diversity and access to maternal and child health education among migrant mothers in China. International Journal for Equity in Health. https://doi.org/10.1186/s12939-026-03006-7

Image Credits: AI Generated

DOI: 10.1186/s12939-026-03006-7

Keywords: linguistic diversity, migrant mothers, maternal and child health education, health equity, basic public health services, cross-dialect migration, China, internal migration, public health, linguistic diversity index, health records, preventive services