About

Dr. Michael Emch is the lead of the Spatial Health Research Group (SHRG) at the University of North Carolina at Chapel Hill. His research focuses on exploring spatio-temporal patterns of disease, with a primary emphasis on infectious diseases in the developing world. Dr. Emch and his group study disease patterns using a holistic approach that investigates the roles of social, natural, and built environments in disease occurrence across different places and populations. Their work employs diverse statistical and spatial analytical methods informed by theories from medical geography, epidemiology, ecology, and related fields. These theories and methods are applied to examine topics such as the contribution of social connectivity to disease incidence, the influence of population-environment drivers on viral evolution, and the use of environmental indicators to predict disease outbreaks. Additionally, Dr. Emch is a member of the UNC Infectious Diseases, Epidemiology, and Ecology Lab, which collaborates closely with the SHRG.

Research topics

• Medicine
• Demography
• Geography
• Biology
• Environmental health
• Sociology
• Immunology
• Pathology
• Virology
• Internal medicine
• Ecology
• Gerontology

Selected publications

• Application of Machine Learning to Identify Influential Factors for Fecal Contamination of Shallow Groundwater

  Water · 2025-01-09 · 4 citations

  articleOpen accessSenior author

  Understanding influential factors for fecal contamination in groundwater is critical for ensuring water safety and public health. The objective of this study is to identify key factors for fecal contamination of shallow tubewells using machine learning methods. Three methods, including recursive feature elimination (RFE) with XGBoost, Random Forest, and mutual information, were implemented to examine E. coli presence and concentration in 1495 tubewell water samples in Matlab, Bangladesh. For E. coli presence, climatic variables, including average rainfall and temperature over the 30, 15, and 7 days preceding sampling, as well as ambient temperature and rainfall on the sampling day, emerged as critical predictors. Land cover characteristics, such as the percentages of urban and agricultural areas within 100 m of a tubewell, were also significant. For E. coli concentration, land cover characteristics within 100 m, the number of hot and heavy-rain days in the 30 days preceding sampling, average rainfall and temperature in the 3 days preceding sampling, and ambient temperature on the sampling day were identified as key drivers. Random Forest and mutual information yielded results that were more similar to each other than to those of RFE with XGBoost. The findings highlight the interplay between climatic factors, land use, and population density in determining fecal contamination in shallow well water and demonstrate the power of machine learning algorithms in ranking these factors.

  PublisherOA PDFDOI

• Estimating exposure to neighborhood crime by race and ethnicity for public health research

  UNC Libraries · 2025-06-11

  articleOpen accessSenior author
  PublisherDOI

• Children with hepatitis B virus infections, Democratic Republic of the Congo

  Bulletin of the World Health Organization · 2025-06-01

  articleOpen access

  Objective: To characterize childhood hepatitis B virus (HBV) epidemiology to inform elimination efforts in the Democratic Republic of the Congo, one of the most populous African countries. Methods: Using the most recent (2013-2014) nationally representative Demographic and Health Survey, we analysed hepatitis B surface antigen (HBsAg) on dried blood spots and associated survey data from children aged 6-59 months. We estimated HBsAg-positivity prevalence nationally, regionally and by potential correlates of infection. We evaluated spatial variation in HBsAg-positivity prevalence overall, and by age, sex and vaccination status. Findings: Using data representing 5773 children, we observed a national HBsAg-positivity prevalence of 1.3% (73/5773; 95% confidence interval, CI: 0.9 to 1.7), ranging from 0.0% in Kinshasa to 5.6% in Sud-Ubangi. Prevalence among boys (1.8%; 95% CI: 1.2 to 2.7) was double that among girls (0.7%; 95% CI: 0.4 to 1.3). Testing negative for tetanus antibodies, rural residence and poorer household were associated with higher HBsAg-positivity prevalence. We observed no difference in prevalence by age. Children had higher HBsAg-positivity odds if living with one or more HBsAg-positive adult household member (odds ratio, OR: 2.3; 95% CI: 0.7 to 7.8), particularly an HBsAg-positive mother (OR: 7.2; 95% CI: 1.6 to 32.3). Notably, nearly two thirds (36/51) of HBsAg-positive children had a HBsAg-negative mother. Conclusion: Our investigation highlights the importance of subnational prevalence estimates in large countries such as the Democratic Republic of the Congo, and we have identified regions that may benefit from improved childhood vaccination delivery strategies and community HBV prevention efforts.

  PublisherDOI

• Geospatial and phylogenetic clustering of acute and recent HIV infections in Lilongwe, Malawi

  UNC Libraries · 2025-12-02

  articleOpen access

  HIV transmission during early HIV infection impedes efforts to end HIV as a public health threat, as diagnosis typically occurs after this period of elevated transmission risk. To guide diagnosis and prevention strategies, we evaluated the geospatial and phylogenetic clustering of acute and recent HIV infection in Lilongwe, Malawi. We identified people with acute (pre-seroconversion) HIV infection (AHI) and a random sample of people with post-acute HIV infection who presented to a sexually transmitted infections (STI) clinic in Lilongwe, Malawi between 2015 and 2019. We evaluated infection recency in people with post-acute HIV using a LAg-Avidity assay. We mapped the household locations of people with AHI and identified geospatial clusters using a flexible scan statistic. We constructed consensus sequences from deep sequencing reads to identify phylogenetic clusters through genetic distance thresholds and maximum likelihood trees. We identified 141 people with AHI, 30 people with recent HIV, and 652 people with chronic (non-recent) HIV. We identified four geospatial clusters that contained the residences of 30% of clinic attendees with AHI, despite comprising just 0.8% of the populated land area and 3.5% of the population. We also identified fourteen distinct two-person phylogenetic clusters. Ten of the fourteen were male-female pairs, nine of which were clinic referral pairs. The remaining four were same-sex pairs who had not referred each other to the clinic and may have been missing network intermediaries. Three of the fourteen phylogenetic pairs consisted of only acute/recent members, and zero phylogenetic linkages were located within geospatial clusters. AHI detection programs anchored in STI clinic populations and their neighborhoods could facilitate identification of early HIV infection, enabling treatment initiation and transmission prevention efforts during this most infectious period. Future studies of intervention packages and deployment approaches can help inform the optimal design and implementation of AHI-focused strategies for reducing HIV incidence.

  PublisherDOI

• Sexually Transmitted Infection Rates and Closure of Family Planning Clinics Because of Abortion Restrictions in Iowa

  UNC Libraries · 2025-05-03

  articleOpen access

  This cohort study assesses rates of sexually transmitted infections in Iowa counties before and after closure of family planning health centers and compared with national rates.

  PublisherDOI

• Geographic dimensions of gastric cancer risk in western Honduras: A spatial ecological analysis

  Health & Place · 2025-12-02

  articleOpen accessSenior author

  Gastric cancer is fifth most common cancer globally, with a complex and multifactorial etiology, including human and bacterial genetics, health behaviors, and environmental exposures. Honduras has among the highest rates of gastric cancer in the western hemisphere, and previous research indicates that gastric cancer risk varies geographically at a sub-national scale. We analyzed characteristics of all 788 incident cases diagnosed in western Honduras between 2002 and 2015 for which residence information was available to assess geographic patterns of excess risk and population-level risk factors. We built hierarchal Bayesian Poisson models to implement spatial ecological analysis stratified by sex and gastric cancer sub-type (diffuse, intestinal, mixed/indeterminate) to explore sex and sub-type specific etiologies. We identified differential geospatial patterns of high observed relative risk (i.e., greater than 1.0) for across male/female and diffuse/intestinal strata. Unexplained variance was more spatially structured in models for males, but less so among females, suggesting that unobserved, spatially autocorrelated factors contributing to geographic risk patterns among males in this area. Additionally, ecological associations varied by sex/sub-type strata. Among males, risk of intestinal gastric cancer was elevated with an increase in the percentage of indigenous Maya/Chortí in the population, but this association was not detected in female strata. Our findings demonstrated substantial differences in geographic patterns of risk across sex and ancestry, which implicated that unobserved, sex-stratified factors, such as germline mutations and behaviors, play a role in shaping geographic patterns of cancer risk at the subnational level in western Honduras.

  PublisherDOI

• Expanding disease ecology even more: One Health, politics and place. Response to commentaries on Michael Emch and Varun Goel’s ‘Disease ecology in health and medical geography: History, progress, and innovations’

  Singapore Journal of Tropical Geography · 2025-11-27

  articleOpen access1st authorCorresponding

  Our paper describes the development of the disease ecology tradition of health and medical geography including some key themes and a few recent innovations. The four commentaries by a distinguished set of scholars add more background to this tradition and offer many useful opinions about productive areas of inquiry that this tradition has and should focus on. Our paper does not provide history, themes and areas of innovation to the entire subdiscipline of health and medical geography but is specifically devoted to disease ecology; and because of our expertise, it focuses mainly on infectious disease ecology. This tradition started with studies of infectious disease ecology almost a century ago, but the breadth of the field has expanded substantially in both infectious and non-communicable disease research as well as health and healthcare. We focus on disease ecology and not health, although one could argue that these cannot be separated. One of us (Michael Emch) writes this while on sabbatical studying infectious diseases at the Faculty of Public Health at Mahidol University in Bangkok, Thailand just six kilometres from the hospital where Jacques May began his explorations in disease ecology in 1932. As Mayer points out in his commentary, ‘his basic concepts added much at the time’ but ‘research has far transcended May and it was – and continues to be – important to progress beyond the past intellectual history of the field’ (Mayer, 2025). This paper and these fine commentaries attempt to bring this scholarly tradition forward. In his commentary, Mayer asks whether the field of health and medical geography should extend beyond the discipline of geography. He advocates for a more inclusive approach to which we wholeheartedly agree. John Hunter championed this inclusive approach in the 1960s and 1970s as the field of medical geography and the tradition of disease ecology within the subdiscipline developed in the United States (US). He believed that the integrative perspective of geography and all the questions and methodologies of geography's various traditions are useful for studying health and disease (Hunter, 1974). Even though our brief review of the disease ecology literature mostly focuses on the work of geographers, our field cannot be separated from cognate disciplines that focus on health. Disciplinary lines are always arbitrary, and debates about the delineation of our discipline and subdiscipline will continue as they have with these commentaries. One of the strengths of geography is that it fosters interdisciplinary and inter-subdisciplinary scholarship. As Mayer states, our subdiscipline is ‘firmly within the health and public health sciences’, which is why many health and medical geographers hold faculty appointments in health sciences. Mayer suggests that our subdiscipline is particularly well-suited to conduct studies that fall into the interdisciplinary fields of One Health and EcoHealth. Oppong and Messenger's commentary also mentions One Health and EcoHealth as areas of growth for geographic disease ecology, as well as the new interdisciplinary field of Planetary Health (Oppong & Messenger, 2025). We agree; and given how complementary these fields are to the disease ecology tradition of health and medical geography, we have actively undertaken research and collaborations that span across these interdisciplinary fields. For example, although we framed our swine influenza case study as a landscape genetics disease ecology study in our paper, one of the study's major goals, as framed to the US National Science Foundation (NSF) is to help build and expand One Health theory. Mayer highlights that while fields such as One Health have created new avenues of research in the health sciences, factors such as the role of social structure, human behaviour, culture and political and economic power are often ignored. In our previous study on H5N1 avian influenza in Vietnam, our current swine flu study and our new study on H5N1 in the US, we examine those factors by collaborating with an interdisciplinary team of veterinarians, virologists and computational biologists. We echo Mayer's call for more involvement in these fields, and our own experience suggests that health and medical geographers can uniquely leverage their holistic interdisciplinary lens to contribute to theoretical and methodological advancements in these fields. In her commentary, McLafferty states that place and politics should have a more prominent role in contemporary disease ecology studies and adds that most quantitative and empirical disease ecology studies often do not acknowledge political ecology of health and disease literature that relies strongly on qualitative and critical approaches (McLafferty, 2025). As Hunter wrote so many decades ago, all approaches are needed to fully understand the geographies of health and disease. McLafferty gives a contemporary example of how politics can influence health as new leaders in the US are presently changing health policy, restricting access to public health data (e.g., Demographic and Health Surveys), and cutting funding for research and programmes. Policy shifts and funding cuts have been dramatic as we write this in 2025 and have affected populations in the US and globally such as a 20 per cent US cut to future federal government allocations to the Supplemental Nutrition Assistance Program which provides monthly funding to low-income households for groceries. Another example is the abrupt 90-day pause of funding on 25 January 2025, for the President's Emergency Plan for AIDS Relief (PEPFAR), which provides services for more than 20 million people worldwide. While the funding was eventually restored in July 2025, the disruption has already led to an estimated 118 960 adult and 12 663 child deaths as of October 2025 (Nichols et al., 2025; PEPFAR Program Impact Tracker, 2025). McLaffterty's point that politics and economics are essential to understanding disease ecologies is painfully obvious as we write this. You cannot understand the ecology of HIV-AIDS at present without considering these funding and programme cuts in the ecology of this disease. Funding cuts to research programmes on diseases will also influence the ecology of diseases. McLafferty emphasizes that this present political climate is a stark reminder of the embeddedness or our work within social and political systems. Our research is rooted in these systems just as disease ecologists in Nazi Germany were embedded in the state goals of empire-building and German racialized nationalism. Our paper mentioned ‘environmental justice’ as an area of innovation and expansion for disease ecology studies that focus on political and economic factors that comprise the underlying structural determinants of health and disease and unequal distributions of environmental burdens. An NSF Directive on 18 April 2025 led to the termination of hundreds of awards that had already been funded that are ‘not aligned with program goals or agency priorities’ including research grants that are focused on ‘environmental justice’ (NSF Directive, 2025). Additionally, the US Environmental Protection Agency Environmental Justice Office was recently shut down and many grants to communities aimed to mitigate unequal distributions of environmental burdens have been cut. As these large-scale cuts and changing political climate shape data reliability, access and availability, we echo McLafferty's call for disease ecologists to apply critical lenses to studying health outcomes. We can also use contemporary approaches such as big data and associated methods to substitute or complement previously available data. Oppong and Messenger mention more historical literature on early disease mapping endeavours that we did not include in the review that are an important part of the history of the disease ecology tradition. They also argue that the disease ecology tradition should describe how environment constrains human behaviours and determines relationships between composition and context, something that Melinda Meade discussed in her seminal paper on human ecology. This theoretical area is also part of the interdisciplinary neighbourhoods and health tradition, which many health and medical geographers have contributed and is described elsewhere (e.g., Emch, Root & Carrel, 2017). Oppong and Messenger also critique existing disease ecology studies, including our Bangladesh case study, and describe challenges with obtaining individual-level data, incorporating structural and political processes, and explaining complex overlapping causes. However, we believe that many of these obstacles can be overcome when geographers engage in interdisciplinary collaborations with other health-related disciplines. Our Bangladesh study – which is part of a long-standing collaboration – is a case in point. By supporting and working with a large network of community health workers and leveraging an existing health and demographic surveillance system, we utilize an existing infrastructure to collect rich individual-level data. Additionally, although spatial Bayesian models, as used in our case study, cannot explain all underlying structural, political and economic factors driving the effectiveness of the intervention, they can be used as a starting point to further investigate where intervention effectiveness may potentially be limited by those structural, political and economic factors. In fact, as more sophisticated quantitative models and rich individual and ecological data sources become available, health and medical geographers are uniquely positioned to leverage their diverse theoretical and methodological expertise and build interdisciplinary teams to address limitations such as incorporating place dynamics and politics as mentioned by both McLafferty, and Oppong and Messenger. Additionally, as echoed by Oppong and Messenger, disease ecologies are complex and cannot be fully explained by linear, causal narratives. To explain the complex causal linkages between several factors and effectiveness of arsenic mitigation interventions, it would be impossible to include every variable related to the effectiveness of the intervention. Hence, our approach has been to incrementally understand the full story of the ecology of the drinking water and the groundwater arsenic problem in Bangladesh through multiple analyses over many years with teams of epidemiologists, environmental microbiologists, geologists and economists, including the narrow example we offer as part of our review paper (e.g., Emch, 1999; Escamilla et al., 2013, Goel et al., 2019, 2023, 2025; Wu et al., 2016). In Paul's commentary, (he has also studied the groundwater arsenic problem in Bangladesh – e.g., Paul & De, 2000; Paul, 2004) he nicely describes more details about the context of this problem in rural Bangladesh and highlights some of our case study findings, namely that deep tubewells are associated with lower diarrheal disease incidence and that the odds vary over space (Paul, 2025). Oppong and Messenger offer another critique of the approach of using landscape genetics to understand geographic drivers of the evolution of pathogens and question whether this approach ‘could help prevent or help limit its spread particularly given the long complex relationship of humans and viruses.’ Very few studies investigate how non-individual level contextual variables such as farming practices or environmental context are involved with evolution and transmission of influenza viruses. This methodological approach is just one that geographers use to conduct disease ecology studies; there are many others that geographic disease ecologists can use. The common theme of all the commentaries is that there are many theoretical and methodological approaches that the tradition of disease ecology—and health and medical geography in general—should include. We concur that a wide range of geographic theories and approaches should be embraced, as well as engagements with interdisciplinary fields such as One Health and collaborations with scholars in other disciplines who also contribute to the larger field of public health.

  PublisherOA PDFDOI

• Integrating mobility, travel survey, and malaria case data to understand drivers of malaria importation to Zanzibar, 2022-2023

  medRxiv · 2025-06-05

  preprintOpen access

  Background: Zanzibar has achieved historic reductions in malaria incidence, but high connectivity to mainland Tanzania and imported cases remain a challenge to "last mile" malaria elimination. Methods: To understand factors driving malaria importation, we collected travel histories and demographics of malaria cases presenting to 94 health facilities across Zanzibar's main island, Unguja, from 2022-2023. We also analyzed population mobility data-self-reported travel at the outbound Dar es Salaam ferry terminal and Meta colocation and movement distribution-to examine movement patterns between Unguja and mainland Tanzania. We integrated these with Climate Hazards Group InfraRed Precipitation with Station (CHIRPS) rainfall data to explore the seasonality of human movement and travel-associated malaria risk. Results: Among 1,172 malaria cases reporting recent travel to mainland Tanzania, travel to Tanga (20%), Dar es Salaam (20%), and Morogoro (15%) were most common. Nearly half of travelers had spent over 14 nights away from home; the majority were visiting family. While mainlanders made up two-thirds of ferry travelers, 28% of travel-associated malaria cases in Unguja reported primary residence on the mainland. Zanzibari residents who reported travel to mainland regions with high/moderate malaria risk during the dry season composed the highest proportion of travel-related malaria cases. Long-distance travel off Unguja decreased during the rainy seasons, and imported cases correlated with rainfall at the mainland source, rather than in Zanzibar. With different biases, colocation and ferry data approximated the proportional makeup of at-risk travel relatively well. Conclusions: Movement flows and seasonal rainfall patterns drive imported malaria in predictable ways that can be harnessed to target high-risk travelers for intervention.

  PublisherOA PDFDOI

• Integrating mobility, travel survey, and malaria case data to understand drivers of malaria importation to Zanzibar, 2022–2023

  UNC Libraries · 2025-11-14

  articleOpen access

  Background Zanzibar has achieved historic reductions in malaria incidence, but high connectivity to mainland Tanzania and imported cases remain a challenge to “last mile” malaria elimination. Methods To understand factors driving malaria importation, prospective surveys were administered to patients at 94 health facilities across Zanzibar’s main island, Unguja, from May 2022 to December 2023, and to travellers at the Zanzibar-bound ferry terminal in Dar es Salaam from August to December 2024. In addition, digital mobility data—Meta colocation and movement distribution—were analysed to examine movement patterns between Unguja and mainland Tanzania. These data were integrated with rainfall data to explore the seasonality of human movement and travel-associated malaria risk. Results Of 3,875 total malaria cases presenting to health facilities in Unguja, 1,172 reported travel to mainland Tanzania within the past 30 days (283/511 (55%) in 2022, 889/3,348 (27%) in 2023). Among these, travel to Tanga (20%), Dar es Salaam (20%), and Morogoro (15%) were most common. Children and females had higher odds of having malaria associated with travel to the mainland (aOR 3.2, 95% CI: 2.2–4.8 and aOR 1.5, 95% CI: 1.3–1.8, respectively, compared to those without mainland travel), while nightwatchmen and students with malaria were more likely to report travel within Zanzibar. Twenty-eight percent of travel-associated malaria cases reported primary residence on the mainland. However, Zanzibari residents who reported travel to mainland regions with high or moderate malaria risk during the dry season made up the highest proportion of travel-related malaria cases (n = 309, 32%). Meta movement distribution data shows that long-distance travel off Unguja decreases during the rainy seasons. In addition, imported cases correlated with rainfall at their mainland destination, rather than rainfall in Zanzibar. With different biases, data from both Meta colocation and ferry surveys approximated the proportional makeup of at-risk travel from different geographies relatively well. In particular, Meta colocation data highlighted travel between Tanga and Unguja, likely via private boats, not captured in the ferry data. Conclusions Movement flows and seasonal rainfall patterns drive imported malaria in predictable ways that can be harnessed to target high-risk travellers for intervention. Supplementary Information The online version contains supplementary material available at 10.1186/s12936-025-05605-1.

  PublisherDOI

• Integrating mobility, travel survey, and malaria case data to understand drivers of malaria importation to Zanzibar, 2022–2023

  Malaria Journal · 2025-11-05 · 2 citations

  articleOpen access

  BACKGROUND: Zanzibar has achieved historic reductions in malaria incidence, but high connectivity to mainland Tanzania and imported cases remain a challenge to "last mile" malaria elimination. METHODS: To understand factors driving malaria importation, prospective surveys were administered to patients at 94 health facilities across Zanzibar's main island, Unguja, from May 2022 to December 2023, and to travellers at the Zanzibar-bound ferry terminal in Dar es Salaam from August to December 2024. In addition, digital mobility data-Meta colocation and movement distribution-were analysed to examine movement patterns between Unguja and mainland Tanzania. These data were integrated with rainfall data to explore the seasonality of human movement and travel-associated malaria risk. RESULTS: Of 3,875 total malaria cases presenting to health facilities in Unguja, 1,172 reported travel to mainland Tanzania within the past 30 days (283/511 (55%) in 2022, 889/3,348 (27%) in 2023). Among these, travel to Tanga (20%), Dar es Salaam (20%), and Morogoro (15%) were most common. Children and females had higher odds of having malaria associated with travel to the mainland (aOR 3.2, 95% CI: 2.2-4.8 and aOR 1.5, 95% CI: 1.3-1.8, respectively, compared to those without mainland travel), while nightwatchmen and students with malaria were more likely to report travel within Zanzibar. Twenty-eight percent of travel-associated malaria cases reported primary residence on the mainland. However, Zanzibari residents who reported travel to mainland regions with high or moderate malaria risk during the dry season made up the highest proportion of travel-related malaria cases (n = 309, 32%). Meta movement distribution data shows that long-distance travel off Unguja decreases during the rainy seasons. In addition, imported cases correlated with rainfall at their mainland destination, rather than rainfall in Zanzibar. With different biases, data from both Meta colocation and ferry surveys approximated the proportional makeup of at-risk travel from different geographies relatively well. In particular, Meta colocation data highlighted travel between Tanga and Unguja, likely via private boats, not captured in the ferry data. CONCLUSIONS: Movement flows and seasonal rainfall patterns drive imported malaria in predictable ways that can be harnessed to target high-risk travellers for intervention.

  PublisherOA PDFDOI

Recent grants

• Impacts of Environment, Host Genetics and Antigen Diversity on Malaria Vaccine Efficacy

  NIH · $3.1M · 2018–2025

• NIH Grant R03AI076748

  NIH · $147k · 2012

• Integration of Spatial and Social Network Analysis in Vaccine Trials

  NSF · $301k · 2009–2013

• Impacts of Environment, Host Genetics and Antigen Diversity on Malaria Vaccine Efficacy

  NIH · $440k · 2018–2018

• Doctoral Dissertation Research: Socioeconomic and Environmental Determinants of Birth Defects in North Carolina

  NSF · $5k · 2008–2009

Frequent coauthors

• Mohammad Yunus

  State University of Gorontalo

  69 shared

• Mohammad Ali

  International Centre for Diarrhoeal Disease Research

  48 shared

• Jonathan J. Juliano

  41 shared

• Verónica Escamilla

  Innovative Research (United States)

  39 shared

• Margaret Carrel

  University of Iowa

  35 shared

• Steven R. Meshnick

  University of North Carolina at Chapel Hill

  34 shared

• Antoinette Tshefu

  Ministry of Public Health

  30 shared

• Mark Janko

  Duke Institute for Health Innovation

  30 shared

Education

• Ph.D., Geography and Epidemiology

  University of North Carolina at Chapel Hill

Awards & honors

• Fellow of the Carolina Population Center