Diabetes mellitus is widely recognised as a global epidemic, with the World Health Organization repeatedly emphasising its growing prevalence and impact on public health systems. Yet the disease is not distributed evenly. Social deprivation—limited access to education, food, housing, and healthcare—fundamentally shapes who bears the greatest burden. In Mexico, these disparities are especially visible, making the country a critical site for studying the intersection of health and inequality.

Introduction

Our paper, “Diabetes Disparities in Mexico: A Spatio-Temporal and Marginalization Index Analysis,” presented by Carlos Hernández Nava at the 22nd International Symposium on Web and Wireless Geographical Information Systems (W2GIS 2025), takes up this challenge. Working with colleagues Sergio Flores and Miguel Mata, we examined diabetes-related avoidable hospitalisations (DRAH) and diabetes-related deaths (DRD) across eight regions of Mexico between 2005 and 2022. By combining national health records with the marginalization index—a composite measure reflecting inequalities in education, housing, income, and access to basic services—we sought to reveal how geography and deprivation interact in shaping health outcomes.

The motivation for this work comes from recognising diabetes not only as a medical problem but also as a social one. Mexico records millions of hospitalisations every year linked to diabetes, and nearly a million deaths over the period studied. These raw figures are alarming, but the deeper patterns reveal an even starker story: where you live, what services you can access, and how much social deprivation your community faces significantly affect your chances of being hospitalised or dying from diabetes.

Results

Our findings show the persistent and uneven geography of diabetes across Mexico. Regions with higher marginalisation scores, such as the Pacífico Sur and Península, consistently record higher rates of hospitalisation and mortality. The story is not one-dimensional: factors such as sex, age group, and healthcare affiliation intersect with marginalisation to create layered vulnerabilities. Adults aged 45–64, for instance, emerged as a particularly at-risk group, especially those affiliated with IMSS and living in areas of high deprivation.

We also observed temporal shifts. While hospitalisation rates have decreased since 2017, likely due to structural changes and the disruptive effects of COVID-19 on health-seeking behaviours, deaths remain high. The strong statistical association between diabetes-related avoidable hospitalisations and diabetes-related deaths suggests that many lives could be saved through better preventative care. Timely and accessible ambulatory treatment can prevent hospitalisation, which in turn reduces the likelihood of progression to fatal outcomes.

One striking observation was how sharply hospitalisation numbers fell during the first months of 2020, coinciding with the onset of the COVID-19 pandemic. This highlights the extent to which health crises interact, with one epidemic influencing the visibility and management of another. Yet the broader trend persists: marginalised communities consistently shoulder the heaviest burden.

Why Spatio-Temporal Analysis Matters

By applying a spatio-temporal framework, this research demonstrates how health outcomes are never static. Geography is not just a backdrop; it is an active determinant. Patterns shift over time, revealing how interventions, social programmes, or crises shape the trajectories of disease. Using the marginalisation index alongside longitudinal data allows us to go beyond national averages and instead highlight regional disparities and temporal dynamics.

This approach also foregrounds the idea of diabetes as an ambulatory care sensitive condition. In other words, diabetes-related hospitalisations should be largely avoidable with timely and adequate outpatient care. Where such care is lacking, hospitalisation rates rise, and so too do deaths. Mapping this spatially, and tracking it over time, provides evidence of where healthcare systems are failing to meet needs—and who is left behind.

Takeaways and Future Work

Looking forward, there are several promising avenues for extending this research. First, we plan to incorporate environmental variables such as air quality, water contamination, and urban infrastructure, which may compound existing vulnerabilities. Linking health records with environmental data could reveal how ecological inequalities exacerbate health disparities.

Second, predictive modelling using machine learning offers the possibility of identifying at-risk populations with greater precision. Classifiers and algorithms could supplement traditional statistical analysis, highlighting communities where targeted interventions would save the most lives. Importantly, such models must remain transparent and accountable, ensuring they do not reinforce the very inequalities they seek to address.

Third, we see real value in scaling the analysis down to state or even municipal levels. While regional analysis identifies broad patterns, localised studies could highlight micro-geographies of risk—areas within a state where deprivation and health burdens intersect most acutely. This kind of fine-grained evidence is essential for policymakers tasked with designing equitable health interventions.

Finally, this work underscores the importance of interdisciplinary collaboration. Bringing together geographers, health scientists, and data analysts enables a fuller understanding of the problem. At W2GIS 2025, we were reminded that geographic information systems are not only about roads, parcels, or coordinates—they are about people, and the uneven ways health and inequality manifest across space and time.

Conclusion

For us, the key message is clear: diabetes in Mexico cannot be understood solely as a medical condition. It is also a social condition, inseparable from the inequalities that structure everyday life. Recognising this means designing interventions that are as much about justice as they are about medicine. Our work at W2GIS 2025 is one step in that direction, using spatio-temporal analysis to uncover hidden geographies of risk and vulnerability.

This is not just about charting patterns of hospitalisation and death. It is about exposing the systemic inequalities that underlie them, and ensuring that public health responses confront those inequalities directly. Only by holding together the spatial, the temporal, and the social can we begin to address the true scale of the diabetes crisis in Mexico. That is the vision we brought to W2GIS 2025, and one we will continue to pursue in the next phase of this research.

See the full paper here: Springer