The numbers behind that headline are stark. Drawing on the American Hospital Association’s annual survey, the research team tracked every hospital closure and opening over the fifteen-year window and found that closures increased by roughly 4 percent year over year while openings declined by about 3 percent annually. In total, 432 hospitals closed during the study period, but only 216 new ones opened, meaning that for every two institutions that shut their doors, just one emerged to take their place. The arithmetic of that imbalance produced the net loss of 216 hospitals, and the accompanying loss of more than 24,000 staffed beds represents a substantial reduction in the country’s capacity to deliver acute care at precisely the moment when demand for hospital services continues to grow.

Perhaps the most consequential finding is what the data did not show. Despite a national policy conversation that has treated hospital closure as an overwhelmingly rural phenomenon, the researchers found no statistically significant difference between rural and urban closure rates. Both categories of hospitals experienced rising closure rates over the study period, and both exceeded the rate at which new hospitals were being established. That symmetry challenges a core assumption embedded in recent federal legislation, including the Rural Health Transformation Program, which directs billions of dollars in state grants specifically toward rural hospitals while leaving urban institutions serving similarly strained populations without comparable support.

The timing of the study gives its findings particular urgency. Since the passage of the One Big Beautiful Bill Act, which included significant changes to Medicaid policy, hospitals across the country have been preparing for the possibility of losing billions of dollars in funding, and anxiety about potential closures has intensified. Much of that anxiety has been channeled toward rural facilities, prompting Congress to create the Rural Health Transformation Program as a lifeline for small-town medical centers. Yet the Harvard analysis suggests that the financial stress now rippling through the hospital sector is not geographically selective, and that urban safety-net hospitals, which disproportionately care for Medicaid patients and uninsured patients, may be just as exposed to the coming fiscal pressure.

When the researchers examined which hospitals actually closed, a consistent profile emerged. For-profit hospitals, small hospitals, safety-net hospitals, and facilities located in the counties with the highest levels of social vulnerability were the most likely to shut down, and that pattern held in both rural and urban settings. Safety-net hospitals and those serving a high volume of Medicaid patients were also rarely the sites of new openings, meaning that the communities most dependent on these institutions were least likely to see replacement capacity emerge. The result is a compounding dynamic in which the places with the greatest health and social needs lose hospital beds fastest and gain them slowest.

Corresponding author Thomas Tsai, associate professor of health policy and management at Harvard Chan School and co-director of the Healthcare Quality and Outcomes Lab, framed the findings as a direct challenge to the prevailing policy narrative. Many policymakers have been treating hospital closures as a rural problem, he observed, but the data show it is a national problem. In his view, the country must shift away from asking only how to save rural hospitals and begin asking how to protect access to hospital care for vulnerable communities wherever they are located. That reframing, if it takes hold in Washington and in state capitols, could reshape how federal and state dollars are targeted in the years ahead.

From a methodological standpoint, the study is notable for its comprehensiveness and its recency. By characterizing closures and openings from 2010 through 2025 using the American Hospital Association annual survey, the team captured the full arc of a turbulent era that included the Affordable Care Act’s coverage expansions, the COVID-19 pandemic, pandemic-era relief funding, and the recent Medicaid policy changes. Observational analyses of this kind cannot by themselves establish why particular hospitals close, but the year-over-year trend lines and the consistent characteristics of closing institutions provide a robust descriptive foundation for policy debate, and the publication in JAMA places the evidence before the clinical and policy communities in a form designed to influence practice and legislation alike.

The financial mechanics behind the pattern are not difficult to reconstruct, even though the study itself does not model them. Small hospitals operate with thin margins and limited ability to absorb reimbursement cuts. For-profit facilities face investor pressure that can accelerate decisions to exit unprofitable markets. Safety-net hospitals, by design, serve patients whose care is reimbursed at low rates or not at all, making them structurally dependent on public funding streams such as Medicaid supplemental payments and disproportionate share allocations. When any of those streams tighten, the institutions at the bottom of the revenue hierarchy are the first to reach the point of insolvency, and the communities they serve, which already face elevated rates of chronic disease and reduced access to primary care, bear the consequences in longer travel times, crowded emergency departments, and delayed treatment.

The loss of more than 24,000 beds over fifteen years is more than an accounting figure. Bed capacity functions as a ceiling on a health system’s ability to respond to surges, whether from pandemics, natural disasters, or seasonal respiratory waves, and the pandemic demonstrated how quickly even well-resourced systems can be pushed past their limits. A national bed supply that is both shrinking and concentrating away from vulnerable communities raises the prospect that the next public health emergency will land hardest on the very populations that lost their local hospitals. The study’s finding that closures outpaced openings in both rural and urban areas suggests this is a systemic contraction rather than a regional reallocation of capacity.

The research, authored by Joshua E. Calianos, Julia H. Song, E. John Orav, Jose F. Figueroa, and Thomas T. Tsai, was supported by the Commonwealth Fund, and its authors’ disclosed funding relationships span federal agencies and major health philanthropies. Its central message, however, is simple enough to survive the disclosures: hospital closure in America is neither a rural story nor an urban one, but a national one that tracks social vulnerability. As Medicaid policy changes begin to flow through hospital balance sheets, the evidence suggests that policymakers who want to preserve access to care will need to look beyond geography and toward the financial fragility of the safety-net institutions, large and small, that hold the health care system’s last line of defense.

Subject of Research: Longitudinal trends in U.S. hospital closures and openings from 2010 to 2025

Article Title: Hospital closures on the rise in both rural and urban areas—particularly among safety-net hospitals

Article References: Hospital closures on the rise in both rural and urban areas—particularly among safety-net hospitals. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: hospital closures, rural hospitals, urban hospitals, safety-net hospitals, Medicaid, health policy, JAMA, Harvard Chan School, hospital beds, social vulnerability, healthcare access, Rural Health Transformation Program

The central insight of the study is deceptively simple but potentially transformative for disaster planning: the tools used to assess flood risk inequity are themselves biased. When the researchers compared their modeled estimates against observed data from the 2016 flood, they discovered that the models captured only 35.8 percent of the actual risk inequity between low-income and high-income communities. In other words, nearly two-thirds of the real disparity in emergency service disruption was invisible to the standard estimation approaches. A city planner using these models would conclude that flooding affects rich and poor neighborhoods in roughly similar ways, and would allocate resources accordingly, unaware that the true burden on disadvantaged areas was far heavier than the numbers suggested.

Why do these models fail so badly, and why do they fail disproportionately for the poor? The Georgia Tech team traced the bias to two mechanisms that existing approaches simply do not account for. The first is traffic congestion. During a flood, residents evacuate, roads clog, and vehicles cluster on the few elevated or unblocked routes that remain passable. Lower-income communities, the study found, experienced greater traffic congestion during the disaster, which slowed emergency vehicles far more than the models predicted. The second mechanism is the failure of the emergency infrastructure itself. More ambulance stations serving lower-income communities were closed or ran out of capacity during the flood, forcing vehicles to travel from farther away and adding delays that no static distance-based model could anticipate.

These findings arrive at a moment when cities around the world are investing heavily in digital twins, flood simulators, and equity dashboards intended to guide climate adaptation. The implicit promise of such tools is that if we can model the hazard precisely enough, we can protect everyone fairly. The Virginia Beach case study punctures that promise. A model can be technically accurate about water depth, road network topology, and travel distances, and still be profoundly wrong about who suffers, because it omits the dynamic, human, and infrastructural realities of a disaster as it unfolds. Congestion is a behavioral phenomenon; station closures are an operational one. Neither appears in the static network calculations that underpin most accessibility analyses.

The methodological approach of the study deserves attention because it offers a template other cities can follow. Rather than building a new simulation from scratch, the researchers performed a validation exercise: they took modeled travel time estimates for the Hurricane Matthew flood and compared them, neighborhood by neighborhood, with the actual ambulance response times recorded during the event. This comparison between modeled and observed data is the gold standard for assessing whether a risk estimation tool reflects reality, yet it is rarely done, particularly with an equity lens. By stratifying the comparison by community income level, Pan and colleagues were able to reveal not just that the models were biased, but that the bias itself was unequal, larger for the communities least able to absorb delayed medical care.

The implications for health outcomes are stark. Ambulance response time is one of the most tightly studied variables in emergency medicine, and the difference between a response within a few minutes and one delayed by five or more can determine whether a patient survives cardiac arrest, stroke, severe trauma, or childbirth complications. During a flood, when call volumes spike and normal hospital capacity is degraded, every additional minute of travel time compounds the strain on the system. If planners believe low-income areas face a two-minute underestimation when the real figure approaches six, they will size their ambulance fleets, pre-position their units, and design their evacuation protocols for a disaster that is milder than the one that actually arrives in those neighborhoods.

There is also a deeper equity dimension. The study’s authors frame their work around the question of whether socially vulnerable neighborhoods face disproportionate travel delays during urban floods, and their answer is an emphatic yes, compounded by the fact that the disproportion is systematically hidden. Risk inequity assessments are supposed to be the corrective lens that lets policymakers see and remedy these disparities. When the lens itself distorts, the harm is doubled: vulnerable communities suffer worse outcomes during the flood, and then suffer again in the planning process, because the data used to justify investments understates their need. Resources flow to where the models say the risk is, and the models say the risk is lower in the places where it is actually highest.

What would better estimation look like? The study points toward models that incorporate real-time traffic dynamics during flood events, rather than assuming free-flow conditions on the surviving road network. It points toward models that track the operational status of emergency facilities, including closures and capacity exhaustion, as part of the disruption itself rather than as an afterthought. And it points toward a culture of validation, in which cities routinely compare their modeled disaster scenarios against observed response data from past events, disaggregated by demographic characteristics, to detect and correct bias before it shapes policy. None of these requirements is conceptually exotic; the novelty of the study lies in demonstrating how much inequity goes undetected without them.

The research was supported by the Georgia Partnership for Innovation Community Research grant program and the National Science Foundation, and it arrives with an urgency amplified by climate change. Urban flooding is intensifying in frequency and severity across coastal and inland cities alike, and the populations most exposed, often lower-income residents in flood-prone districts, are precisely those with the fewest private resources to compensate for public failures. An ambulance that arrives six minutes late in a wealthy suburb with a private car in the driveway is a different catastrophe than one that arrives six minutes late in a neighborhood where that ambulance was the only realistic route to emergency care. The bias in the models, in this sense, is not merely a technical flaw but a distributional one, determining whose emergencies the system is calibrated to meet.

For the field of urban sustainability and disaster resilience, the Virginia Beach study is likely to become a canonical example of what researchers call estimation bias in risk assessment, and a reminder that equity analysis is only as good as the models on which it rests. The authors’ finding that conventional approaches captured barely a third of the true risk inequity should prompt every city that maintains a flood response plan to ask an uncomfortable question: if our models missed 64 percent of the disparity in Virginia Beach, how much are they missing here? The answer will not be found in more sophisticated water simulations or finer-grained road networks alone, but in grounding those models in the messy, observed reality of how floods actually paralyze traffic, close stations, and stretch emergency capacity, and in checking, neighborhood by neighborhood, whether the numbers we plan with match the minutes patients actually wait.

Subject of Research: Biased estimation of flood-related emergency service disruptions and ambulance response inequity in urban areas

Article Title: Exposing biased risk estimation of emergency service disruptions during urban floods

Article References: Pan, X., Mohammadi, N., & Taylor, J. E. (2026). Exposing biased risk estimation of emergency service disruptions during urban floods. npj Urban Sustainability. https://doi.org/10.1038/s42949-026-00473-3

Image Credits: AI Generated

DOI: 10.1038/s42949-026-00473-3

Keywords: urban flooding, ambulance response time, risk estimation bias, emergency services, social vulnerability, Hurricane Matthew, Virginia Beach, traffic congestion, health equity, disaster modeling, infrastructure resilience, npj Urban Sustainability

The urgency of the review is grounded in the changing character of viral threats. Roughly 60% of emerging infectious diseases originate as zoonoses, and 72% of those derive from wildlife reservoirs. The past two decades delivered SARS-CoV, H1N1 influenza, MERS-CoV, Ebola, Zika and SARS-CoV-2 in rapid succession, and recent events, including a Bundibugyo virus outbreak in the Democratic Republic of the Congo that spread to Uganda and prompted a World Health Organization emergency declaration, together with an Andes virus cluster aboard the cruise vessel MV Hondius, illustrate how mobility, urban density and humanitarian crises accelerate dissemination. The authors emphasise that these events are syndemics rather than simple epidemics: pathogens interact synergistically with poverty, structural inequality and chronic disease burden, compounding health and economic damage in ways that purely biological models fail to capture.

A central vulnerability the review identifies is the countermeasure gap during the earliest phase of any outbreak. Traditional vaccine development takes 10 to 15 years, and even the accelerated COVID-19 effort required roughly 11 months from genome sequencing to emergency authorisation. Infectious disease therapeutics have a median clinical development time of 7.3 years. Diagnostics pose an equally severe problem, because targeted assays such as multiplex PCR depend on prior knowledge of a pathogen’s genome or antigens, meaning a genuinely novel virus goes undetected. Metagenomic next-generation sequencing offers an agnostic alternative that can sequence all nucleic acids in a sample without prior assumptions, but it suffers from lower sensitivity at low viral loads, complex bioinformatics, high costs and limited availability precisely in the low- and middle-income regions where most spillovers originate.

To quantify transmission, the review returns to the compartmental foundations laid by Ross, Hamer and the Kermack-McKendrick SIR model of 1927, extended to the SEIR framework that adds an exposed class for diseases with incubation periods. These models are systems of non-linear ordinary differential equations whose behaviour is analysed through equilibrium and stability theory: the disease-free and endemic equilibria are assessed by linearising the system and examining the eigenvalues of the Jacobian matrix, with negative real parts indicating that an outbreak will dissipate. The pivotal parameter is the basic reproduction number R0, the average number of secondary infections generated by one infectious individual in a fully susceptible population. When R0 exceeds 1, a pathogen can invade; the herd immunity threshold follows as 1 minus 1/R0. The authors caution, however, that R0 is an emergent property of the pathogen-host-environment system rather than a biological constant, and that real-time tracking requires the effective reproduction number Rt, which declines as interventions and behavioural changes take hold.

The choice between deterministic and stochastic modelling matters most where it counts: at the beginning. Deterministic differential-equation models are computationally efficient and describe average trajectories in large populations, whereas stochastic models explicitly incorporate random variation and are critical for small populations, rare events and early outbreak phases, where the two approaches can yield qualitatively different predictions about the probability of a large epidemic. During COVID-19, models that explicitly represented pre-symptomatic and asymptomatic compartments alongside quarantined individuals revealed that these overlooked groups drive substantial transmission, making widespread testing far more effective than symptom-based surveillance alone. At the applied end, agent-based models such as Covasim simulate individuals with realistic demographics, contact networks and viral-load-based transmissibility, and can run full intervention scenarios on a standard laptop in under a minute, which is why health agencies in more than a dozen countries have used them for real-time policy support.

Artificial intelligence extends this toolkit into molecular territory that sequence-homology tools such as BLAST cannot reach. Protein structure predictors like AlphaFold and ESMFold map viral glycoproteins with atomic precision, and machine learning applied to the Flaviviridae family revealed more than 100 previously unrecognised glycoproteins. Convolutional neural networks such as ViraMiner and DeePaC detect highly divergent viral genomes in metagenomic samples that standard alignment approaches label as unknown, while transformer architectures like LucaProt have identified candidate novel RNA virus species from RNA-dependent RNA polymerase sequences. Perhaps most strikingly, the EVEscape framework, trained exclusively on viral sequences available before 2020, anticipated SARS-CoV-2 variation with accuracy comparable to high-throughput experimental antibody-escape scans, and generalises to influenza, HIV, Lassa and Nipah, enabling vaccine designers to target conserved regions less susceptible to evasion.

At the population level, AI-powered surveillance systems parse open-source data to flag anomalies days to weeks before official recognition. HealthMap uses natural language processing and Bayesian machine learning to extract hyperlocal geographic information from multilingual web reports, having provided early indications of H1N1 in Mexico and the 2019 vaping-associated lung disease outbreak, while EpiWatch proved effective during the 2022 global mpox epidemic. One global disease-activity database detected 94% of WHO-identified outbreaks an average of 43.4 days earlier, and Bayesian outbreak detection algorithms have caught influenza surges with only a four-to-five-day delay and minimal false alarms. Wearable devices add another layer: smart-ring data achieved an area under the curve of 0.85 for population-level fever surveillance. In clinical diagnostics, deep learning reaches approximately 96% accuracy for COVID-19 on CT scans and up to 99% for distinguishing COVID-19 from pneumonia on chest X-rays, and multimodal frameworks that combine imaging, tabular and text data outperform single-source models by 6 to 33% across healthcare tasks.

The review is candid about the pitfalls. Hybrid architectures that embed SEIR equations directly into neural network loss functions, so-called physics-informed neural networks, have outperformed purely data-driven models for COVID-19 forecasting, and a technique called epimodulation improved hospital-admission forecast accuracy by an average of 12.3% for COVID-19 and 32.9% for influenza, with the largest gains around epidemic peaks. But deploying streaming models without formal safeguards is risky. The authors highlight uncertainty quantification through Bayesian ensembles and conformal prediction, which reduced diagnostic errors in histopathology from 2% to 0.1% while flagging unreliable predictions for expert review, alongside structural identifiability analysis, simulation-based calibration and automated detection of concept drift, the silent degradation that occurs when the statistical relationship between inputs and outcomes shifts as pathogens and clinical practices evolve. Algorithmic bias compounds these concerns: models trained predominantly on high-income-country data may misclassify patients in low- and middle-income settings, and a systematic review found 87% of machine learning prediction models carried a high risk of bias.

The proposed Silicon Shield is therefore organised as a modular reference architecture across three tiers. A data ingestion layer aggregates molecular streams from metagenomic sequencing pipelines, clinical streams through standardised FHIR APIs, and digital intelligence from NLP engines alongside environmental and mobility data. A model pipeline layer then progresses from unsupervised anomaly detection through evolutionary and immune-escape forecasting to mechanistic transmission modelling coupled with agent-based simulators. A decision interface tier translates outputs into role-specific dashboards for public health officials and explainable clinical decision support for frontline physicians, using techniques such as SHAP to open the black box. Crucially, the framework incorporates social vulnerability indices, which have consistently predicted COVID-19 incidence and mortality, and applies causal inference methods, including directed acyclic graphs and targeted maximum likelihood estimation, to move beyond correlational surveillance toward equity-informed intelligence.

The authors conclude that the technological foundations are largely in place: algorithms capable of processing heterogeneous global data streams, sequencing technologies for rapid pathogen characterisation, and computational infrastructure supporting real-time analytics. What remains is the harder work of building institutional capacity, harmonising incompatible IT systems, navigating legal barriers to data sharing, raising AI literacy in the health workforce and ensuring that surveillance capabilities are distributed equitably rather than concentrated in wealthy nations. Only with sustained, deliberate and fair investment, they argue, can computational capability be translated into genuine public health protection before the next Disease X arrives.

Subject of Research: Integration of mathematical modelling and artificial intelligence for pandemic preparedness against emerging viral syndemics

Article Title: Integrating mathematical modelling and artificial intelligence to combat emerging viral syndemics: A systematic review

Article References: Cabrera-Rodríguez, R., Reyes-Castañeda, I., Lorenzo-Sánchez, I., Valenzuela-Fernández, A., & Trujillo-González, R. (2026). Integrating mathematical modelling and artificial intelligence to combat emerging viral syndemics: A systematic review. New Microbes and New Infections, 74, Article 101843. https://doi.org/10.1016/j.nmni.2026.101843

Image Credits: AI Generated

DOI: 10.1016/j.nmni.2026.101843

Keywords: viral syndemics, mathematical modelling, artificial intelligence, pandemic preparedness, SEIR models, metagenomic sequencing, immune escape prediction, disease surveillance, machine learning diagnostics, social vulnerability, zoonotic spillover, Disease X

Triple-negative breast cancer, which lacks the three molecular targets exploited by modern targeted therapies, accounts for roughly 10 to 15 percent of all breast cancers but is diagnosed at more advanced stages and offers fewer effective treatment options than other subtypes. It disproportionately strikes women under 40, women with BRCA1 mutations, and Black women, who are diagnosed at twice the rate of women of any other racial or ethnic background. Because clinical trials have driven nearly every recent advance in the disease’s treatment, including the landmark pembrolizumab studies that reshaped care within the last five years, equitable access to those trials is not an abstract concern. It is the mechanism by which scientific progress translates into longer lives.

To build their national picture, the researchers queried ClinicalTrials.gov as of September 30, 2024, sifting through all 510,397 registered studies to isolate active phase II and phase III interventional treatment trials specific to triple-negative breast cancer. After excluding non-cancer studies, studies without U.S. sites, and trials no longer open to enrollment, they identified 108 active trials, 58 for metastatic disease and 50 for non-metastatic disease, supported by 1,230 trial sites across the country. Each site was assigned to a U.S. county through spatial join analysis, and the resulting tabulations were overlaid with county-level data from the U.S. Census Bureau, the U.S. Department of Agriculture’s rural-urban continuum codes, the Centers for Disease Control and Prevention’s Social Vulnerability Index, and the U.S. Climate Vulnerability Index’s baseline health vulnerability scores.

The headline numbers are sobering. Of the nation’s 3,144 counties, 76.2 percent had no available triple-negative breast cancer trial whatsoever. Only 12.4 percent had exclusively federally sponsored trials, 9.7 percent had both federally and non-federally sponsored trials, and a mere 1.7 percent relied solely on industry or other non-federal sponsors. Yet the picture is not uniformly bleak: 78 percent of the roughly 124 million U.S. women aged 18 and older lived in a county with at least one trial available, and another 17 percent had access in an adjacent county. The problem is concentrated in specific, identifiable populations, and the study’s county-level breakdown reveals exactly where the trial network frays.

Rural America sits at the sharpest edge of the disparity. Trials were available in just 9.8 percent of rural counties compared with 47.0 percent of metropolitan counties, a difference the authors report as highly significant. And among the rural counties that did host trials, 80 percent offered exclusively federally sponsored studies, compared with only 42 percent of metropolitan counties with trials. In other words, the federal clinical trials infrastructure, including the National Cancer Institute’s National Clinical Trials Network, is effectively the sole lifeline keeping rural women connected to experimental treatments. Any contraction in federally sponsored research, the authors warn, would have uniquely devastating consequences for rural patients, whose cancer mortality rates, including ten-year breast cancer mortality, are already known to exceed those of their urban counterparts.

Health vulnerability tells a parallel story. Counties in the highest quartile of the Climate Vulnerability Index’s baseline health vulnerability measure, reflecting burdens of chronic disease, limited access to care, and poorer preventive health, were significantly less likely to have any available trial: 88.3 percent had none, compared with 72.2 percent of counties in the lower three quartiles. Among counties with trials, 73 percent of the most health-vulnerable depended exclusively on federal sponsorship, versus 50 percent of less vulnerable counties. Social vulnerability showed a similar, though somewhat attenuated, gradient, with 78.6 percent of the most socially vulnerable counties lacking any trial. This mirrors a broader pattern in American oncology, where 94 percent of all cancer trial sites are located in areas more affluent than the national average.

Race and ethnicity shaped the map as well. Among counties with predominantly non-Hispanic White populations, 75.8 percent had no triple-negative breast cancer trial, compared with 79.1 percent of predominantly Hispanic counties, 81.5 percent of predominantly Black counties, and a striking 96.9 percent of predominantly American Indian and Alaska Native counties. But the researchers caution that county-level racial predominance does not necessarily reflect where members of each group actually live. When they analyzed access at the population level instead, most Black women, 82.1 percent, and Hispanic women, 86.5 percent, resided in counties with an available trial, while access dropped to 63.5 percent among American Indian and Alaska Native women and rose to 93.5 percent among Asian and Pacific Islander women. The divergence between county-level and population-level measures, the authors argue, underscores the need to evaluate trial access through both lenses.

Regional disparities compound the inequity. Ten percent of women in the South, more than five million people, lacked access to a triple-negative breast cancer trial in either their home county or any neighboring county, compared with just 1 percent of women in the Northeast. This is particularly troubling because the South carries the heaviest burden of new triple-negative breast cancer diagnoses in the country, accounting for 40 percent of cases, more than the Northeast and West combined. Two-thirds of Southern counties had a trial available, versus 87 percent of counties in the Northeast and West. Meanwhile, metastatic trials, the studies most relevant to patients with advanced disease, were scarcer than early-stage offerings: only 63 percent of trial sites had metastatic trials, and just 9.4 percent of metastatic trial sites offered four or more options, limiting both proximity and choice for the women who need trials most urgently.

The study’s findings arrive amid a broader reckoning with the representativeness of cancer research. Trial participants have historically been younger, healthier, and less racially, ethnically, and geographically diverse than the patients who ultimately receive the resulting therapies. Prior work by some of the same investigators found that 70 percent of all U.S. counties lacked any cancer treatment trial and that 51 percent had neither research nor oncology care sites. Survey data cited in the new paper show that only 37 percent of patients actively undergoing cancer treatment are willing to travel for a trial, yet more than half of women with metastatic breast cancer report driving over an hour to reach a trial site. When comparable trial options are available to rural and urban patients alike, studies suggest, the survival gap between them narrows or disappears, making geographic access a modifiable determinant of outcome.

The authors point toward concrete remedies: pragmatic and decentralized trial designs that reduce the burden of participation, expanded eligibility criteria to ensure results generalize to real-world populations, and technologies such as telemedicine and remote monitoring that can decouple enrollment from physical proximity. They also emphasize the outsized importance of sustained federal investment in programs like the NCI Community Oncology Research Program and the National Clinical Trials Network, which extend trial infrastructure into community settings. The analysis has limitations, including its reliance on manually entered ClinicalTrials.gov data and its descriptive, county-level design, which cannot establish causation or capture individual enrollment decisions. Still, as the first study to chart the geography of triple-negative breast cancer trials against the demographics of the women they are meant to serve, it offers government agencies, industry sponsors, and hospital systems a data-driven roadmap for deciding where the next trial site should open, in a disease where novel treatment strategies are urgently needed and every mile between a patient and a protocol can matter.

Subject of Research: Geographic access to triple-negative breast cancer clinical trials in the United States

Article Title: Geographic access to triple negative breast cancer clinical trials: are trials located near patients?

Article References: Amin, L. B., Kirkwood, M. K., Levit, L. A., Balogh, E. P., Waterhouse, D. M., Unger, J. M., Garrett-Mayer, E., & Chavez-MacGregor, M. (2026). Geographic access to triple negative breast cancer clinical trials: are trials located near patients?. Breast Cancer Research and Treatment, 219(3), Article 22. https://doi.org/10.1007/s10549-026-08084-3

Image Credits: AI Generated

DOI: 10.1007/s10549-026-08084-3

Keywords: triple-negative breast cancer, clinical trials, health disparities, geographic access, rural health, health equity, ClinicalTrials.gov, National Cancer Institute, social vulnerability, breast cancer mortality, trial sponsorship, United States

The numbers behind the finding are striking in their consistency. Above a heat island intensity threshold of 2.19 degrees Celsius, or about 3.9 degrees Fahrenheit, every additional 1 degree Celsius of neighborhood heat excess was associated with 8 percent higher odds of failing to reach that textbook outcome. In a cohort of 697,552 patients with an average age of 76, only 50.2 percent achieved the composite benchmark. The remainder experienced complications in 25.2 percent of cases, prolonged hospital stays in 22.9 percent, readmission within the study window in 14.8 percent, and death within 30 days in 7.1 percent.

An urban heat island is the well-documented phenomenon by which developed areas run warmer than their rural or suburban surroundings. Asphalt, concrete, brick, and dark roofing materials absorb solar radiation during the day and re-emit it slowly at night, while the scarcity of trees and vegetation removes the shade and evaporative cooling that natural landscapes provide. The result is a temperature differential that can vary substantially from one neighborhood to the next within the same city, meaning two patients discharged from the same hospital on the same day may return to environments that differ by several degrees during a heat wave.

To capture that variation at scale, the research team, led by senior author Dr. Timothy M. Pawlik, professor and chair of the Department of Surgery at The Ohio State University Wexner Medical Center, developed a method they call the Adaptive Spatiotemporal Environmental Thermal Health Exposure Resolution framework. The approach uses satellite-derived temperature data to connect each patient’s residential area with urban heat island intensity, postoperative outcomes, and community characteristics. Rather than relying on a single weather station reading for an entire metropolitan area, the framework resolves thermal exposure at the neighborhood level, which is precisely the scale at which heat risk diverges between tree-lined streets and heat-trapping industrial corridors.

One of the most consequential aspects of the analysis is that the heat-recovery association was not distributed evenly across the population. The relationship between neighborhood heat and suboptimal surgical recovery was stronger among Black patients and among residents of communities with high scores on the Social Vulnerability Index, a composite measure developed by public health agencies to quantify community-level social and economic disadvantage, including factors such as poverty, housing quality, and access to transportation. This pattern echoes a broader body of environmental health research showing that the burdens of extreme heat fall hardest on communities with the fewest resources to adapt.

The effect also varied by the type of operation. When the researchers restricted their analysis to heat island exposure above the 2.19 degree Celsius threshold, the association was particularly pronounced among patients undergoing cancer-related procedures, emergency surgeries, and operations carrying high cardiac risk. Dr. Pawlik offered a physiological rationale for the cardiac signal: dehydration and shifts in the body’s fluid balance during high heat place additional stress on the cardiovascular system, and cardiac patients may have less reserve to tolerate those stresses while a surgical wound heals and the body mounts its recovery.

The proposed mechanisms are biologically plausible. Postoperative recovery demands stable hydration, adequate perfusion of healing tissues, and the ability to mount an immune response without additional physiological strain. Sustained heat exposure increases insensible fluid loss through sweating, can thicken blood and strain the heart, and disrupts sleep, all of which are known to complicate convalescence. Older adults, who made up the bulk of this cohort, are especially vulnerable because the body’s thirst response and sweating efficiency decline with age, and because many take medications such as diuretics or beta blockers that further impair thermoregulation and fluid balance.

Because the study is observational, it identifies an association rather than proving that residential heat exposure caused the worse outcomes. Patients in hotter neighborhoods may differ from those in cooler ones in ways the analysis could not fully disentangle, including housing quality, air conditioning access, baseline health, and proximity to follow-up care. Still, the dose-response pattern, in which risk climbs steadily with each additional degree of heat island intensity above the threshold, and the concentration of the effect among physiologically vulnerable patients lend weight to the argument that heat itself is a meaningful variable in surgical recovery.

The practical implications reach into the everyday mechanics of discharge planning. Dr. Pawlik framed the issue in bluntly practical terms: does the patient have air conditioning in the house, does the family have a fan, and will the patient be able to stay hydrated at home? He suggested that questions about heat exposure may need to be increasingly incorporated into discharge planning, particularly during hot weather, and that surgical teams should screen for heat exposure and connect at-risk patients with community resources before they leave the hospital. In practice, that could mean asking about home cooling the way clinicians already ask about home oxygen or caregiver support.

Access to cooling, however, is far from equal. Public health guidance identifies older adults, people with chronic health conditions or disabilities, people with limited incomes, people experiencing homelessness, and those without safe, adequately cooled housing as being at highest risk during extreme heat. Many communities open cooling centers during periods of dangerous temperatures, and hospitals, patients, and caregivers can consult state-by-state directories to locate local assistance. As climate change intensifies heat waves and lengthens the warm season across much of the United States, the study’s authors argue that the environment a patient returns to must be treated as part of the clinical picture, not as background noise. Heat, the data suggest, is not merely uncomfortable; for hundreds of thousands of surgical patients, it may be a measurable determinant of whether they heal.

Subject of Research: The association between urban heat island intensity and postoperative surgical recovery outcomes across the United States

Article Title: Urban heat islands linked to worse recovery after major surgery

Article References: Urban heat islands linked to worse recovery after major surgery. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: urban heat island, surgery, postoperative outcomes, textbook outcome, health disparities, social vulnerability, climate change, discharge planning, dehydration, cardiac risk, satellite temperature data, American College of Surgeons

The case for this reframing is written in recent disaster statistics. When severe floods hit Indonesia in late 2025, the death toll passed 700 and roughly one million people were evacuated, according to reporting cited by the authors. Disasters on this scale are rarely products of water alone. They emerge from the interaction between a hazard and a population whose housing is fragile, whose warning systems fail, whose savings are thin, and whose political voice is muted. Risk analysts have formalised this insight for decades: risk is conventionally understood as the product of hazard, exposure, and vulnerability. Yet investment and policy attention remain overwhelmingly concentrated on the first of those three terms, because hazard modelling is tractable, fundable, and spectacularly visualisable, while vulnerability is messy, contested, and slow to measure.

The technical capabilities now exist to change this. Recent work in the flood modelling literature, including studies by Seenath and colleagues published in Risk Analysis in 2025, has pushed toward integrating social dimensions into flood risk assessment rather than treating hydraulic output as the end product. Research published in Nature Communications in 2024 by Fox and collaborators demonstrated that large-scale flood modelling can now capture inundation dynamics with enough fidelity to support decision-making at national and global scales. A 2025 study in npj Natural Hazards extended similar approaches to vulnerability-aware impact estimation. The computational machinery for vulnerability-informed flood governance is, in other words, no longer speculative. What remains missing, Stambekova and Seenath contend, is the governance architecture, and the political will, to place social vulnerability at the centre of flood risk management rather than at its margins.

The evidence base for why this matters is now substantial. A widely cited 2022 analysis by Rentschler, Salhab, and Jafino in Nature Communications estimated that roughly 1.8 billion people worldwide are exposed to substantial flood risk, and that a disproportionate share of those exposed live in low- and middle-income countries, with poorer households within those countries frequently occupying the most dangerous ground. Work on climate-related mobility, such as Sonja Ayeb-Karlsson’s research in Climate Risk Management, has documented how environmental stress interacts with poverty and immobility, trapping the least resilient in harm’s way. Studies by Rhein and Kreibich in Natural Hazards and Earth System Sciences in 2025 examined how the drivers of flood damage change over time, underscoring that vulnerability is dynamic and must be tracked, not assumed. Each of these strands points in the same direction: the distribution of flood consequences is a social variable, not a hydraulic one.

Measuring vulnerability, however, is far harder than measuring water depth. Vulnerability operates at multiple scales simultaneously. At the macro scale, it reflects national wealth, institutional quality, and infrastructure investment. At the meso scale, it reflects the characteristics of specific communities, including housing stock, access to emergency services, and social cohesion. At the micro scale, it reflects the circumstances of individual households: age, disability, income, tenure security, insurance coverage, and social networks. Stambekova and Seenath sketch a conceptual structure for micro-scale vulnerability profiles, arguing that flood governance needs to know not merely which neighbourhoods flood, but which households within those neighbourhoods will be unable to evacuate, will lose their livelihoods, and will still be displaced months after the water recedes. Bradford and colleagues, writing in Natural Hazards and Earth System Sciences, offered one of the earlier systematic frameworks for measuring flood vulnerability across such dimensions, and subsequent work by Moreira, de Brito, and Kobiyama reviewed indicator-based approaches to vulnerability assessment, revealing both progress and persistent fragmentation in how the concept is operationalised.

The fragmentation is itself a governance problem. Different agencies hold different pieces of the vulnerability puzzle: census bureaus hold demographic data, social services hold welfare data, housing authorities hold tenure data, and meteorological and hydrological services hold hazard data. Rarely are these streams integrated into a single, actionable picture of who is most at risk before a flood arrives. Recent methodological work by Yousaf, Seenath, and Speight in the Journal of Flood Risk Management has explored how uncertainty in flood risk assessment can be handled more rigorously, and the same rigour, the authors argue, needs to be applied to the social layer of risk. A warning issued 48 hours before a flood is only as effective as the capacity of recipients to act on it, and that capacity varies enormously across a population in ways that current warning systems seldom account for.

The economic argument for vulnerability-centred governance is compelling as well. Hallegatte and colleagues, writing in Economics of Disasters and Climate Change, showed that losses from natural disasters fall disproportionately on poor people, and that well-designed resilience policies can substantially reduce this unequal burden. Investment that reduces vulnerability, through safer housing, social protection floors, accessible early warning, and inclusive planning, delivers returns that pure hazard-defence investment cannot, because it protects people rather than merely repelling water. Jonkman, Curran, and Bouwer, analysing global flood fatality patterns in Natural Hazards in 2024, found that mortality from flooding remains heavily concentrated in lower-income contexts, a pattern that hardened infrastructure alone has not reversed. Meanwhile, work by Vestby and colleagues in the Proceedings of the National Academy of Sciences linked flood exposure to broader social instability, illustrating how unmanaged flood risk radiates outward into conflict, displacement, and political strain.

Susan Cutter, one of the founding figures of vulnerability science, argued in the International Journal of Disaster Risk Reduction in 2024 that the field has matured to the point where vulnerability assessment should be embedded routinely in disaster risk governance rather than confined to academic analysis. Stambekova and Seenath’s commentary extends that argument into the specific domain of flood governance, calling for a reorientation in which vulnerability profiling becomes a first-order input to decisions about levee placement, warning system design, evacuation planning, and post-flood recovery financing. The figure accompanying their article depicts the conceptual structure of micro-scale vulnerability profiles, a template for how household-level social data could be layered onto hydraulic hazard maps to produce risk pictures that predict human outcomes rather than merely water depths.

None of this diminishes the achievements of flood modelling, which have transformed what authorities can know about an impending event. The point, the authors insist, is that knowing where water will go answers only half the question. The other half, who will be there when it arrives, and whether they can survive and recover, demands data, institutions, and political commitment that flood governance has historically delegated to chance. As climate change intensifies hydrological extremes and urbanisation pushes more people onto floodplains, the gap between the physically modelled hazard and the socially distributed consequence will widen. Closing that gap, Stambekova and Seenath argue, is no longer a research aspiration but a governance imperative: flood risk management must put the most vulnerable at its centre, or it will continue to be measured, grimly, by the losses it fails to prevent.

Subject of Research: The integration of social vulnerability assessment into flood risk governance and management

Article Title: Flood risk governance must put social vulnerability at the centre

Article References: Stambekova, A., & Seenath, A. (2026). Flood risk governance must put social vulnerability at the centre. Nature Water. https://doi.org/10.1038/s44221-026-00713-x

Image Credits: AI Generated

DOI: 10.1038/s44221-026-00713-x

Keywords: flood risk, social vulnerability, flood governance, flood modelling, disaster resilience, hazard exposure, early warning systems, climate adaptation, community resilience, natural hazards, risk assessment, Nature Water

The research team, led by Natasha Malmin of Georgia State University, surveyed 133 adults visiting a rural food pantry in Williamsburg County during July 2023 and 96 members of the general community in Horry County between October and November of the same year. The pantry serves roughly 1,500 people each month through a drive-through distribution model, and researchers approached every vehicle in the line over two days of data collection, achieving an impressive 93 percent eligibility and completion rate among those who volunteered. The Horry County comparison group was recruited at community festivals, cultural events, and shopping centers, selected deliberately because the suburban county has a considerably wealthier and more educated population.

The demographic contrasts between the two cohorts were striking. The pantry cohort had a median age of 64 years, while the community cohort skewed far younger. Some 76 percent of pantry participants identified as non-Hispanic African American, compared with just 14 percent of the community sample, which was 67 percent non-Hispanic White. Educational attainment diverged sharply as well: half of pantry respondents had only a high school diploma or less, whereas the community group contained large proportions of bachelor’s and graduate degree holders. Williamsburg County itself, where the pantry sits, has a poverty rate of 23 percent and a median household income of $43,471, against Horry County’s 12.6 percent poverty rate and $61,063 median income.

Health burdens told a similarly stark story. The pantry cohort reported hypertension at a rate of 71 percent and diabetes at 41 percent, conditions that carry serious implications for disaster resilience, since chronic disease management depends on uninterrupted access to medication, refrigeration, and care. Only 10 percent of pantry respondents rated their general health as excellent, compared with 23 percent of the community cohort. In an unexpected twist, the community group reported substantially higher rates of self-diagnosed anxiety, at 65 percent, and depression, at 52 percent. The researchers caution that this pattern should be treated as hypothesis-generating rather than conclusive, noting that Williamsburg County lies within a designated mental health care desert, where limited access to diagnosis and care may suppress reported prevalence rather than reflect true differences in mental health burden.

The most consequential findings emerged when respondents described their experiences with federally declared disasters. Both cohorts reported similar frequencies of exposure to the string of storms that have battered the region since 2015, including the Great Flood, Hurricane Matthew, Hurricane Florence, Hurricane Dorian, Hurricane Ian, and Tropical Storm Idalia. Yet when the researchers examined the consequences of that exposure, the divergence was unmistakable. Among those reporting damage, 69 percent of pantry users suffered losses exceeding $3,000, while a majority of the community cohort reported losses below $1,000. Power outages followed the same pattern: 28 percent of pantry respondents sat in the dark for six to ten days, and 18 percent for eleven to twenty days, whereas nearly half of community respondents reported outages shorter than three days.

One structural factor looms particularly large in explaining these disparities: heirs’ property. This form of ownership, in which land passes informally through generations without formal title documentation, is disproportionately common among African American households in the South due to historic exclusion from formal legal systems. Some 57 percent of pantry participants reported heirs’ property ownership, compared with just 19 percent of the community cohort. The consequences are severe, because FEMA disaster assistance historically required documented ownership verification. After Hurricane Katrina, more than 20,000 heirs’ property owners were denied aid, and after Hurricane Maria struck Puerto Rico, 80,000 applicants were rejected over land title issues. Within South Carolina alone, researchers have estimated 162,803 acres of heirs’ property, carrying an assessed value of $34.6 million.

The study found that disaster-experienced pantry users relied on FEMA aid at a statistically significantly higher rate than their community counterparts, a seemingly paradoxical result given the documented barriers heirs’ property owners face. FEMA announced policy changes in 2021 intended to ease the application burden for residents of inherited land, but many of the storm impacts reported by pantry users date back to before those reforms, and the authors note that policy implementation often lags, particularly for communities simultaneously navigating age, disability, and educational barriers. The longer outages and larger losses in the pantry cohort also align with emerging evidence that rural and minoritized communities experience slower power restoration after disasters, and with earlier research showing that homes in lower-income areas sustain greater physical damage from comparable hazard exposure, likely reflecting aging housing stock and deferred maintenance.

Climate hazard experiences added further nuance. Fully 80 percent of pantry users reported encountering at least one climate-related hazard, compared with 67 percent of the community sample, a statistically significant difference. Flooding and heatwaves were more commonly reported by the pantry cohort, though not significantly so. Curiously, the community cohort reported significantly more experience with drought, at 15 percent versus 4 percent, and wildfire, at 13 percent versus 1 percent, even though federal wildfire risk indicators rate Williamsburg County as riskier than more than 80 percent of United States counties, against Horry County’s medium ranking above 63 percent. The authors suggest that differences in awareness, hazard salience, and media coverage may drive this perception gap. Heatwave reporting offered another curious wrinkle: pantry users were surveyed during an active National Weather Service heat advisory in July, yet only 24 percent reported having experienced a heatwave, suggesting that even direct exposure does not reliably translate into recognized risk.

The study’s limitations deserve honest mention. Both samples were non-probability convenience samples, so the findings cannot be generalized to all county residents, and the modest sample sizes restricted the analysis to descriptive comparisons using Fisher’s exact tests rather than causal inference. Food security was not measured with a validated scale; instead, pantry use served as a pragmatic proxy for concentrated structural vulnerability, an assumption grounded in prior South Carolina research showing that households with very low food security had 5.9 times higher odds of using food pantries than food-secure households. Timing differences also complicated direct comparison, since Tropical Storm Idalia struck coastal South Carolina between the two survey windows, affecting the Horry sample’s recent experience but not the pantry cohort’s.

Nevertheless, the policy implications are clear and potentially transformative. The authors argue that food pantries, as trusted anchor institutions that already reach populations traditional public health systems often miss, are ideally positioned to serve as hubs for disaster communication, preparedness outreach, early warnings, and heat-health alerts. Prior pantry-based interventions covering nutrition education and chronic disease management demonstrate that these settings can effectively engage hard-to-reach residents, although sustainable implementation will require addressing capacity constraints within pantry systems. As climate change intensifies the frequency and severity of extreme weather across the Southeast, one of the nation’s least climate-prepared regions, the study’s central message resonates beyond South Carolina: communities seeking to build genuine resilience must look not only at hazard maps, but at the food pantry lines, where the most vulnerable residents are already gathered and where recovery, too often, begins from the furthest behind.

Subject of Research: Comparative disaster and climate hazard vulnerability of rural food pantry users in South Carolina

Article Title: Compounded vulnerability: comparing disaster and climate hazard experiences of rural food pantry users with a neighboring suburban community in the Southeastern US

Article References: Malmin, N., Dzokoto, M., Height, T., Le Moal, T., Idowu, O., Driffin, A., & Pressley, T. (2026). Compounded vulnerability: comparing disaster and climate hazard experiences of rural food pantry users with a neighboring suburban community in the Southeastern US. Journal of Environmental Studies and Sciences. https://doi.org/10.1007/s13412-026-01140-w

Image Credits: AI Generated

DOI: 10.1007/s13412-026-01140-w

Keywords: food insecurity, food pantries, disaster recovery, climate hazards, rural health, heirs' property, FEMA assistance, South Carolina, social vulnerability, power outages, chronic disease, hurricanes

The research team, led by Helen Fifer and Helen Peters with colleagues from sexual health, paediatric infectious diseases and pathology services across the United Kingdom, reviewed 79 reported cases through a multidisciplinary Clinical Expert Review Panel. Using strict clinical-pathological criteria, the panel classified 67 infants as confirmed or probable cases of congenital syphilis. These infants were born to 65 women, including two sets of twins. The outcomes were stark: 54 were liveborn, of whom five died within the first month of life, and 13 were stillborn. Over half of the liveborn infants — 37 of 67 — arrived preterm, and half weighed less than 2.5 kilograms at birth. Against a background of roughly 4.96 million live births in England during the study period, the absolute numbers are small, with an incidence of 0.023 per 1,000 live births still below the World Health Organization elimination threshold of 0.5 per 1,000. But each case represents a failure of a system designed to prevent exactly this outcome.

The epidemiological backdrop explains much of the rise. England recorded 9,535 cases of early infectious syphilis in 2024, the highest annual figure since the 1940s. Although most infections occur among gay and bisexual men, diagnoses in women tripled between 2015 and 2024, climbing from 273 to 830. This resurgence in adults of reproductive age inevitably feeds into pregnancy. Yet in 1999, researchers had proposed removing syphilis from the antenatal screening panel altogether because congenital cases had all but disappeared. The new data show how quickly that calculus changed: since 2019, an average of ten congenital cases per year have been reported, compared with just one or two per year in the early 2010s.

Perhaps the most striking finding is that nearly half of the affected infants — 30 of 65 pregnancies ending in congenital syphilis — were born to mothers who screened negative for syphilis at their first antenatal appointment and then acquired the infection later in pregnancy. Because England offers only a single universal screen at booking, supplemented by risk-based repeat testing, these women were never re-tested. The risk-based strategy depends on women disclosing new sexual partners, partner sexually transmitted infection diagnoses, drug injection, or sex work, and on clinicians recognising and acting on those disclosures. Many of the mothers in this study had no identifiable risk factors at all and would only have been caught by a universal repeat screen later in gestation, as is practised in higher-prevalence regions of the United States and Europe.

Diagnostic delays compounded the harm. Clinicians repeatedly reported being falsely reassured by a negative antenatal screening result. Infants born to these mothers often arrived at or shortly after birth with non-specific, multisystem illness — irritability, respiratory compromise, thrombocytopaenia, jaundice — that was frequently mistaken for presumed neonatal sepsis. Yet almost all affected infants showed additional features uncommon in sepsis: hepatosplenomegaly, rash or skin and mucosal lesions, desquamation, and abnormalities of the long bones. In one neonatal death, an infant with respiratory symptoms, thrombocytopaenia and anaemia was never tested for syphilis during life; the diagnosis emerged only at postmortem. Several children with persistent symptoms were not tested for months or years, with some cases identified only between 12 and 24 months of age, sometimes incidentally when the mother was screened during a subsequent pregnancy.

Even among infants whose mothers were diagnosed antenatally or in labour, clinical recognition proved challenging. Of 28 such livebirths, eight infants had no symptoms at birth, and the most common signs among symptomatic babies were hepatosplenomegaly in 11, rash or mucocutaneous lesions in seven, and bone abnormalities in five, typically accompanied by thrombocytopaenia, anaemia or jaundice. More unusual complications included hydrops in four infants, hydrocephalus and nephrotic syndrome in two each, and meningitis in one. The authors note that syphilis has long been called ‘the great imitator’, and the study confirms that congenital disease lives up to the name. They endorse proposals to broaden the traditional ‘TORCH’ screen for suspected congenital infection — toxoplasmosis, rubella, cytomegalovirus and herpes simplex — to a ‘SCORTCH’ panel that explicitly includes syphilis, alongside chickenpox and blood-borne viruses.

The study also exposes weaknesses in laboratory diagnosis. Serology in newborns is inherently difficult because maternal IgG antibodies cross the placenta, and a non-treponemal titre four times higher than the mother’s is considered diagnostic — yet only 24 per cent of the 41 mother–infant pairs with comparable results met that threshold, rising to 50 per cent among infants whose mothers acquired syphilis after a negative booking screen, presumably because those mothers went untreated. By contrast, treponemal IgM, which does not cross the placenta, was positive in 86 per cent of the 35 infants tested. Polymerase chain reaction (PCR) testing for Treponema pallidum DNA proved a highly useful adjunct, with positive results from nasal and throat swabs, skin lesions, blood, cerebrospinal fluid, placenta and even bone biopsy, and all four stillbirths with PCR results testing positive. Yet fewer than a quarter of liveborn infants had any specimen sent for PCR, a gap the UK Health Security Agency is now addressing by providing free PCR testing to all UK laboratories.

Behind the clinical statistics lies a darker pattern of social vulnerability. Just over half of the women — 34 of 65 — had at least one complex social factor documented during pregnancy, and most of those experienced multiple overlapping disadvantages. These included involvement with social services in 25 cases, insecure housing in 19, mental health problems in 14, difficulties engaging with healthcare in 14, drug or alcohol misuse in 12, intimate partner violence in nine, and sex work in six. Many women booked late for antenatal care, after 12 weeks, or not at all, and treatment completed too close to delivery to cure the fetus. For cure to succeed, penicillin therapy must finish at least four weeks before birth, so late booking combined with preterm delivery leaves an impossibly narrow window. Several mothers were also treated with macrolide antibiotics, a regimen now known to fail frequently because of widespread macrolide resistance in syphilis strains; UK guidelines were amended in 2019 to remove this option and now explicitly advise against it.

The authors are careful to acknowledge the limitations of their work. Prospective surveillance only began in 2020, with data for 2015 to 2020 collected retrospectively, so early cases were probably undercounted. Reporting was voluntary until April 2025, when congenital syphilis became a notifiable disease in England. The cases captured are also likely skewed towards the severe end of the spectrum, and an unknown number of asymptomatic infants may remain undiagnosed, carrying a risk of late congenital syphilis with its devastating effects on bones, teeth, eyes and the nervous system. International comparisons suggest the problem is not unique to England: series from the United States, Argentina and Western Australia describe similar patterns of asymptomatic birth followed by early symptom onset, and diagnostic difficulty.

The message for policymakers is clear. The UK National Screening Committee, which concluded in 2019 that universal repeat screening at 28 weeks would not be cost-effective, is reviewing that position, and the new data suggest many cases would have been caught by such a policy. Alongside repeat screening, the authors call for greater awareness among clinicians that a negative early screen offers no protection for the rest of pregnancy, improved recognition of maternal symptoms such as vulval lesions and rashes, wider use of PCR and IgM testing, and closer collaboration with inclusion health services — housing, drug treatment and outreach teams — to reach socially excluded women. With the World Health Organization targeting triple elimination of syphilis, HIV and hepatitis B transmission by 2030, England’s decade of surveillance data serves as a reminder that elimination is not achieved by screening programmes alone, but by the clinical vigilance, timely treatment and social support that make those programmes work.

Subject of Research: Congenital syphilis transmission, clinical presentation and diagnostic challenges in England, 2015–2024

Article Title: Clinical characteristics and factors contributing to transmission of congenital syphilis in England, 2015–24: a national surveillance study

Article References: Fifer, H., Peters, H., Kingston, M., Francis, K., Till, R., Thorne, C., Lyall, H., Dermont, S., Cohen, M. C., Sultan, B., Oomeer, S., Bamford, A., Hoodbhoy, S., Permalloo, N., Dickins, D., Emonts, M., Jones, C. E., Andrews, S., & Elbech, A. (2026). Clinical characteristics and factors contributing to transmission of congenital syphilis in England, 2015–24: a national surveillance study. The Lancet Regional Health – Europe, 70, Article 101864. https://doi.org/10.1016/j.lanepe.2026.101864

Image Credits: AI Generated

DOI: 10.1016/j.lanepe.2026.101864

Keywords: congenital syphilis, antenatal screening, Treponema pallidum, stillbirth, neonatal death, vertical transmission, PCR diagnostics, pregnancy, public health surveillance, England, social vulnerability, Lancet Regional Health Europe