Thomas Rawson
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Thomas Rawson has a total of 158 co-authors. Here are the top co-authors by number of shared publications:
- Christian Morgenstern
Imperial College London
11 shared publications - Neil Morris Ferguson
MRC Centre for Global Infectious Disease Analysis, School of Public Health, Imperial College London, London, UK
9 shared publications - Michael B. Bonsall
Department of Biology, University of Oxford, Oxford, United Kingdom
7 shared publications - Natsuko Imai
Imperial College London
6 shared publications Anne Cori
Imperial College London
5 shared publications- Anne Cori
MRC Centre for Global Infectious Disease Analysis, School of Public Health, Imperial College London, London, UK
5 shared publications
Latest Publications
- Open Access
Privacy-Preserving, Decentralised Inference of Epidemic Transmission Dynamics from Digital Exposure Histories
Background: Exposure-history data collected during epidemics or pandemics within digital contact-tracing applications hold immense potential for rapid, scalable epidemiological inference. The sensitivity of these data makes privacy-preserving designs essential to secure public trust and app uptake. However, during the COVID-19 pandemic, such approaches were not yet available at the scale or maturity required for deployment. We present a Bayesian decentralised framework to infer transmission dynamics from exposure-history data while protecting user privacy. Methods: Our approach is based on iterative communication between user devices and a central server, computing log-likelihood gradients on device from exposure-history data and centrally updating parameter values using only aggregated gradients. Applying this framework to synthetic exposure-history data from a hypothetical respiratory pathogen epidemic simulated on dynamic contact networks, we assessed its ability to estimate the underlying transmission parameters, perform infector attribution, and reconstruct serial-interval (SI) distributions. We systematically explored how inference was affected by incomplete exposure-history data arising from partial app adoption, case under-ascertainment, and incomplete exposure logging, and by transmission scenarios. Findings: Our decentralised framework accurately recovered transmission parameters, closely matching estimates derived from centralised data with no systematic bias in posterior means. Parameter recovery remained robust when hazards from unobserved exposures were known as a daily average across the population. Although incomplete exposure-history data and specific transmission scenarios introduced compensatory parameter biases, the inferred host infectiousness profile retained sufficient accuracy, and the weekly SI distributions reconstructed from infector attributions closely resembled the ground-truth distributions. Interpretation: Our findings highlight that, beyond its primary role as a control measure, digital contact tracing could help characterise transmission dynamics almost in real time and inform public health interventions while preserving user privacy. Funding: Oxford Martin School Programme on Digital Pandemic Preparedness; National Institute for Health and Care Research (NIHR) Health Protection Research Unit in Emerging and Zoonotic Infections. Evidence before this study: We searched PubMed for studies published from database inception to Sept 30, 2026, using the keywords (“digital contact tracing” OR “exposure notification” OR “contact tracing app”) AND (“federated learning” OR “decentralised” OR “privacy-preserving” OR “statistical inference” OR “Bayesian inference”) AND (“transmission” OR “serial interval” OR “infectiousness”). During the COVID-19 pandemic, digital contact-tracing (DCT) apps were deployed at an unprecedented scale: 224 COVID-19 apps were launched across 127 countries, 128 of which supported exposure notification. In England and Wales, the NHS COVID-19 app had 16.5 million regular users (28% of the total population) and averted an estimated 1 million cases and 9,600 deaths in its first year. To protect privacy, however, most apps used decentralised architectures that kept exposure-history data on users’ devices. This privacy-preserving principle has largely prevented these data from being used for epidemiological inference to understand transmission dynamics, such as estimating the serial-interval (SI) distribution and host infectiousness profile over time. While existing decentralised algorithms estimate each user’s infection risk, we found no study that developed and evaluated a framework for estimating population-level epidemiological parameters from exposure-history data that remain on users’ devices. Added value of this study: To our knowledge, this is the first such framework. It estimates key quantities required for understanding transmission dynamics, including how an infected individual’s infectiousness changes over time relative to their symptom onset date, how much close contacts increase the risk of transmission, and the SI distribution. The framework learns an exposure hazard function from data that never leave users’ devices, then attributes infections to likely infectors, and reconstructs the SI distribution. We applied the framework to synthetic exposure-history data from epidemics simulated on dynamic contact networks of 166,034 individuals, matching the 2024 population of Oxford, UK. It produced estimates comparable to those from centralised analysis, identified the timing of peak host infectiousness relative to symptom onset to within one day, and reconstructed SI distributions that closely tracked the ground truth. We also systematically assessed how participation levels (app adoption and case ascertainment), technical limitations (incomplete exposure logging), and surveillance (knowledge of the background hazard from unobserved exposures) affected inference, and showed that daily population-level knowledge of the background hazard was sufficient for reliable inference. Implications of all the available evidence: Beyond their role as a control measure, DCT apps could provide timely, population-representative estimates of key transmission parameters, such as the host infectiousness profile and SI distribution. Our findings show that such inference is feasible within the privacy-preserving architectures likely to underpin future DCT systems. Greater app adoption, case ascertainment, and exposure logging would improve inference accuracy. To realise this potential, the framework will need to be embedded in DCT systems, supported by concurrent epidemiological research, real-world validation, and early collaboration with public health authorities and platform providers.
Open AccessPrivacy-Preserving, Decentralised Inference of Epidemic Transmission Dynamics from Digital Exposure Histories
Background: Exposure-history data collected during epidemics or pandemics within digital contact-tracing applications hold immense potential for rapid, scalable epidemiological inference. The sensitivity of these data makes privacy-preserving designs essential to secure public trust and app uptake. However, during the COVID-19 pandemic, such approaches were not yet available at the scale or maturity required for deployment. We present a Bayesian decentralised framework to infer transmission dynamics from exposure-history data while protecting user privacy. Methods: Our approach is based on iterative communication between user devices and a central server, computing log-likelihood gradients on device from exposure-history data and centrally updating parameter values using only aggregated gradients. Applying this framework to synthetic exposure-history data from a hypothetical respiratory pathogen epidemic simulated on dynamic contact networks, we assessed its ability to estimate the underlying transmission parameters, perform infector attribution, and reconstruct serial-interval (SI) distributions. We systematically explored how inference was affected by incomplete exposure-history data arising from partial app adoption, case under-ascertainment, and incomplete exposure logging, and by transmission scenarios. Findings: Our decentralised framework accurately recovered transmission parameters, closely matching estimates derived from centralised data with no systematic bias in posterior means. Parameter recovery remained robust when hazards from unobserved exposures were known as a daily average across the population. Although incomplete exposure-history data and specific transmission scenarios introduced compensatory parameter biases, the inferred host infectiousness profile retained sufficient accuracy, and the weekly SI distributions reconstructed from infector attributions closely resembled the ground-truth distributions. Interpretation: Our findings highlight that, beyond its primary role as a control measure, digital contact tracing could help characterise transmission dynamics almost in real time and inform public health interventions while preserving user privacy. Funding: Oxford Martin School Programme on Digital Pandemic Preparedness; National Institute for Health and Care Research (NIHR) Health Protection Research Unit in Emerging and Zoonotic Infections. Evidence before this study: We searched PubMed for studies published from database inception to Sept 30, 2026, using the keywords (“digital contact tracing” OR “exposure notification” OR “contact tracing app”) AND (“federated learning” OR “decentralised” OR “privacy-preserving” OR “statistical inference” OR “Bayesian inference”) AND (“transmission” OR “serial interval” OR “infectiousness”). During the COVID-19 pandemic, digital contact-tracing (DCT) apps were deployed at an unprecedented scale: 224 COVID-19 apps were launched across 127 countries, 128 of which supported exposure notification. In England and Wales, the NHS COVID-19 app had 16.5 million regular users (28% of the total population) and averted an estimated 1 million cases and 9,600 deaths in its first year. To protect privacy, however, most apps used decentralised architectures that kept exposure-history data on users’ devices. This privacy-preserving principle has largely prevented these data from being used for epidemiological inference to understand transmission dynamics, such as estimating the serial-interval (SI) distribution and host infectiousness profile over time. While existing decentralised algorithms estimate each user’s infection risk, we found no study that developed and evaluated a framework for estimating population-level epidemiological parameters from exposure-history data that remain on users’ devices. Added value of this study: To our knowledge, this is the first such framework. It estimates key quantities required for understanding transmission dynamics, including how an infected individual’s infectiousness changes over time relative to their symptom onset date, how much close contacts increase the risk of transmission, and the SI distribution. The framework learns an exposure hazard function from data that never leave users’ devices, then attributes infections to likely infectors, and reconstructs the SI distribution. We applied the framework to synthetic exposure-history data from epidemics simulated on dynamic contact networks of 166,034 individuals, matching the 2024 population of Oxford, UK. It produced estimates comparable to those from centralised analysis, identified the timing of peak host infectiousness relative to symptom onset to within one day, and reconstructed SI distributions that closely tracked the ground truth. We also systematically assessed how participation levels (app adoption and case ascertainment), technical limitations (incomplete exposure logging), and surveillance (knowledge of the background hazard from unobserved exposures) affected inference, and showed that daily population-level knowledge of the background hazard was sufficient for reliable inference. Implications of all the available evidence: Beyond their role as a control measure, DCT apps could provide timely, population-representative estimates of key transmission parameters, such as the host infectiousness profile and SI distribution. Our findings show that such inference is feasible within the privacy-preserving architectures likely to underpin future DCT systems. Greater app adoption, case ascertainment, and exposure logging would improve inference accuracy. To realise this potential, the framework will need to be embedded in DCT systems, supported by concurrent epidemiological research, real-world validation, and early collaboration with public health authorities and platform providers.
- Open Access
The impact of an ageing population on pandemic risk in England
Background England's population is ageing rapidly, with those aged 80 years and older projected to increase 85% by 2047. Aging populations may change the dynamics of future pandemics, since older individuals face higher risks of severe outcomes and require a disproportionate share of early vaccine supply, for certain respiratory infections, despite having fewer social contacts. The net pandemic-risk implications of these competing effects cannot be resolved without dynamic transmission modelling. Methods We used an age-structured stochastic SARS-CoV-2 transmission model, previously calibrated to the real-time COVID-19 epidemic across seven NHS regions of England. Running the fitted model under the ONS principal population projection for 2047 as a counterfactual, we held all non-pharmaceutical interventions fixed and scaled vaccine supply to overall population growth. Nine counterfactual scenarios spanned three vaccine deployment strategies and three ONS migration assumptions. Findings The basic reproduction number remained largely unchanged by the projected age distribution. However, under the central counterfactual scenario, cumulative hospital admissions in the first two years of the pandemic were 674,004 - a 40.3% increase over the historic baseline - and hospital deaths increased by 49.4%, disproportionate to the 13.9% total population increase. Trajectories were similar before the Delta variant emerged but diverged thereafter, as a larger elderly cohort delayed vaccine protection in younger, higher-contact groups. Projected burden was robust to migration assumptions (664,451--689,833 admissions). Vaccine deployment strategy had substantially greater influence, with admissions ranging from 625,859 (optimistic age-scaled deployment) to 1,193,883 (non-scaled). Interpretations Population ageing will disproportionately amplify England's pandemic burden, driven primarily by the interaction between a growing elderly cohort and fixed age-ordered vaccine allocation, rather than by population growth alone. This vulnerability is largely independent of future migration uncertainty but is strongly modifiable through demographically-adaptive vaccine deployment planning developed before, not during, the next pandemic.
Open AccessThe impact of an ageing population on pandemic risk in England
Background England's population is ageing rapidly, with those aged 80 years and older projected to increase 85% by 2047. Aging populations may change the dynamics of future pandemics, since older individuals face higher risks of severe outcomes and require a disproportionate share of early vaccine supply, for certain respiratory infections, despite having fewer social contacts. The net pandemic-risk implications of these competing effects cannot be resolved without dynamic transmission modelling. Methods We used an age-structured stochastic SARS-CoV-2 transmission model, previously calibrated to the real-time COVID-19 epidemic across seven NHS regions of England. Running the fitted model under the ONS principal population projection for 2047 as a counterfactual, we held all non-pharmaceutical interventions fixed and scaled vaccine supply to overall population growth. Nine counterfactual scenarios spanned three vaccine deployment strategies and three ONS migration assumptions. Findings The basic reproduction number remained largely unchanged by the projected age distribution. However, under the central counterfactual scenario, cumulative hospital admissions in the first two years of the pandemic were 674,004 - a 40.3% increase over the historic baseline - and hospital deaths increased by 49.4%, disproportionate to the 13.9% total population increase. Trajectories were similar before the Delta variant emerged but diverged thereafter, as a larger elderly cohort delayed vaccine protection in younger, higher-contact groups. Projected burden was robust to migration assumptions (664,451--689,833 admissions). Vaccine deployment strategy had substantially greater influence, with admissions ranging from 625,859 (optimistic age-scaled deployment) to 1,193,883 (non-scaled). Interpretations Population ageing will disproportionately amplify England's pandemic burden, driven primarily by the interaction between a growing elderly cohort and fixed age-ordered vaccine allocation, rather than by population growth alone. This vulnerability is largely independent of future migration uncertainty but is strongly modifiable through demographically-adaptive vaccine deployment planning developed before, not during, the next pandemic.
- Open Access
Federated analysis of incubation period distributions using individual-level observed data and heterogeneous summary statistics
The incubation period, the interval between pathogen exposure and symptom onset, is a critical epidemiological parameter for follow-up policy and outbreak response, yet individual-level exposure data remain scarce, especially early in outbreaks. For most priority pathogens, only summary statistics are available because sharing of individual-level data can be sensitive. Here we introduce a Bayesian hierarchical framework that jointly models individual-level observations and published summary statistics under a unified federated analysis framework. Simulation studies demonstrate that the method accurately recovers incubation period distributions across a range of data availability scenarios, generally outperforming approaches that use published summary statistics alone. Applying the framework to 18 pathogens, including 10 priority pathogens classified to have outbreak potential by the World Health Organization, we find substantial between-study heterogeneity in incubation period estimates, including by outbreak country for SARS-CoV-1, variants of concern for COVID-19, and exposure setting for typhoid fever. These estimates, together with the curated dataset and modelling framework in our associated R package ddsynth , provide a reproducible foundation for improved incubation period estimation and synthesis across pathogens of epidemic concern. Our framework enables robust and rapid estimation of incubation periods during new outbreaks.
Open AccessFederated analysis of incubation period distributions using individual-level observed data and heterogeneous summary statistics
The incubation period, the interval between pathogen exposure and symptom onset, is a critical epidemiological parameter for follow-up policy and outbreak response, yet individual-level exposure data remain scarce, especially early in outbreaks. For most priority pathogens, only summary statistics are available because sharing of individual-level data can be sensitive. Here we introduce a Bayesian hierarchical framework that jointly models individual-level observations and published summary statistics under a unified federated analysis framework. Simulation studies demonstrate that the method accurately recovers incubation period distributions across a range of data availability scenarios, generally outperforming approaches that use published summary statistics alone. Applying the framework to 18 pathogens, including 10 priority pathogens classified to have outbreak potential by the World Health Organization, we find substantial between-study heterogeneity in incubation period estimates, including by outbreak country for SARS-CoV-1, variants of concern for COVID-19, and exposure setting for typhoid fever. These estimates, together with the curated dataset and modelling framework in our associated R package ddsynth , provide a reproducible foundation for improved incubation period estimation and synthesis across pathogens of epidemic concern. Our framework enables robust and rapid estimation of incubation periods during new outbreaks.
- Open Access
A systematic review of Nipah virus disease epidemiological parameters, outbreaks and mathematical models
We conducted a systematic review (PROSPERO CRD42023393345) characterising the epidemiology, outbreaks and mathematical models of Nipah virus (NiV), an important public health threat in South and Southeast Asia. We extracted 243 parameters, 89 risk factors, 39 models and 23 distinct outbreaks from 119 papers. IgG seroprevalence estimates ranged from 0% to 12.5%. NiV causes severe disease, with pooled case-fatality ratio estimates ranging widely from 9.1% (95%CI: 0.2%-41.3%) in Singapore to 81.9% (95%CI: 71.9%-88.9%) in Bangladesh. NiV’s natural history is poorly characterised; we estimated a median incubation period of 8.77 days (95%CI: 7.53-10.02) based on 8 studies. Transmission parameter estimates were scarce, and all but one central estimate of the basic reproduction number were below 1. NiV mathematical models (n=39) were rarely fitted to data (n=8). All extracted information is accessible via our R package, epireview , a dynamic resource for informing responses to future outbreaks of NiV and related pathogens.
Open AccessA systematic review of Nipah virus disease epidemiological parameters, outbreaks and mathematical models
We conducted a systematic review (PROSPERO CRD42023393345) characterising the epidemiology, outbreaks and mathematical models of Nipah virus (NiV), an important public health threat in South and Southeast Asia. We extracted 243 parameters, 89 risk factors, 39 models and 23 distinct outbreaks from 119 papers. IgG seroprevalence estimates ranged from 0% to 12.5%. NiV causes severe disease, with pooled case-fatality ratio estimates ranging widely from 9.1% (95%CI: 0.2%-41.3%) in Singapore to 81.9% (95%CI: 71.9%-88.9%) in Bangladesh. NiV’s natural history is poorly characterised; we estimated a median incubation period of 8.77 days (95%CI: 7.53-10.02) based on 8 studies. Transmission parameter estimates were scarce, and all but one central estimate of the basic reproduction number were below 1. NiV mathematical models (n=39) were rarely fitted to data (n=8). All extracted information is accessible via our R package, epireview , a dynamic resource for informing responses to future outbreaks of NiV and related pathogens.
- Open Access
Epidemiological characteristics and vaccination impact scenario modelling of concurrent Clade I mpox outbreaks in the Democratic Republic of the Congo and Burundi
In 2024, mpox cases surged in the Democratic Republic of the Congo (DRC) with cross-border spread to Burundi. We developed a transmission-dynamic model calibrated against surveillance data to understand drivers in enzootic (Clade Ia) and non-enzootic (Clade Ib) areas, and the potential impact of vaccination. In non-enzootic areas we estimated that 58-84% of transmission occurred within sexual networks. MVA-BN vaccination of sex workers could have averted 91% (95% CrI 81%–98%) of infections in Sud Kivu (DRC) but only 35% (95% CrI 26%–47%) in Bujumbura (Burundi), due to later outbreak detection. In historically enzootic Equateur (DRC), ongoing zoonotic spillover best explained sustained incidence. There, pledged Lc18m8 vaccines could have averted 42% (95% CrI 40%–46%) of infections; prioritising children improved impact. Across all settings, doubling vaccine coverage by using a single dose of MVA-BN outperformed two-dose strategies. Timely detection and tailored vaccination strategies are critical to reducing mpox burden.
Open AccessEpidemiological characteristics and vaccination impact scenario modelling of concurrent Clade I mpox outbreaks in the Democratic Republic of the Congo and Burundi
In 2024, mpox cases surged in the Democratic Republic of the Congo (DRC) with cross-border spread to Burundi. We developed a transmission-dynamic model calibrated against surveillance data to understand drivers in enzootic (Clade Ia) and non-enzootic (Clade Ib) areas, and the potential impact of vaccination. In non-enzootic areas we estimated that 58-84% of transmission occurred within sexual networks. MVA-BN vaccination of sex workers could have averted 91% (95% CrI 81%–98%) of infections in Sud Kivu (DRC) but only 35% (95% CrI 26%–47%) in Bujumbura (Burundi), due to later outbreak detection. In historically enzootic Equateur (DRC), ongoing zoonotic spillover best explained sustained incidence. There, pledged Lc18m8 vaccines could have averted 42% (95% CrI 40%–46%) of infections; prioritising children improved impact. Across all settings, doubling vaccine coverage by using a single dose of MVA-BN outperformed two-dose strategies. Timely detection and tailored vaccination strategies are critical to reducing mpox burden.
- Top Cited PapersOpen Access
- 27 January 2026
- Published by Springer Nature in Communications Earth & Environment
- Vol. 7 (1)
An outbreak of highly pathogenic avian influenza H5N1 could impact the dairy cattle sector and the broader economy in the United States
The outbreak of Highly Pathogenic Avian Influenza H5N1 in U.S. dairy cattle poses substantial risks to public health, economic sustainability of farming, and global food systems. Using a Computable General Equilibrium model, we simulate its short- to medium-term impacts on Gross Domestic Product and other macro-economic outcomes for the US and its main trading partners. We simulate impacts under the current situation and realistic and reasonable worst-case scenarios. We estimate domestic economic losses ranging between 0.06% and 0.9% of US GDP, with losses to the dairy sector ranging between 3.4% and 20.6%. Trading partners increase dairy production to compensate for the loss. Current government subsidies are about 1.2% (95% HDI: 1% to 1.4%) of output losses, and likely insufficient to incentivise farmers to step up surveillance and biosecurity for mitigating the possible emergence of H5N1 strains with pandemic potential into human populations.
- 27 January 2026
- Published by Springer Nature in Communications Earth & Environment
- Vol. 7 (1)
Top Cited PapersOpen AccessAn outbreak of highly pathogenic avian influenza H5N1 could impact the dairy cattle sector and the broader economy in the United States
The outbreak of Highly Pathogenic Avian Influenza H5N1 in U.S. dairy cattle poses substantial risks to public health, economic sustainability of farming, and global food systems. Using a Computable General Equilibrium model, we simulate its short- to medium-term impacts on Gross Domestic Product and other macro-economic outcomes for the US and its main trading partners. We simulate impacts under the current situation and realistic and reasonable worst-case scenarios. We estimate domestic economic losses ranging between 0.06% and 0.9% of US GDP, with losses to the dairy sector ranging between 3.4% and 20.6%. Trading partners increase dairy production to compensate for the loss. Current government subsidies are about 1.2% (95% HDI: 1% to 1.4%) of output losses, and likely insufficient to incentivise farmers to step up surveillance and biosecurity for mitigating the possible emergence of H5N1 strains with pandemic potential into human populations.
- Open Access
A systematic review of Zika virus disease: epidemiological parameters, mathematical models, and outbreaks
Background: Zika virus (ZIKV) is an Aedes-borne arbovirus that can cause serious risks of neurological complications and birth defects in newborns of mothers infected during pregnancy. ZIKV is classified as a priority pathogen by the World Health Organization. Methods: We conducted a systematic review of peer-reviewed studies reporting ZIKV epidemiological parameters, transmission models, and outbreaks (PROSPERO #CRD42023393345) to characterise its transmissibility, seroprevalence, risk factors, disease sequelae, and natural history. We performed a meta-analysis of Congenital Zika Syndrome (CZS), pregnancy loss probabilities amongst ZIKV-infected mothers and the proportion of symptomatic cases. Findings: We extracted information from 574 of 27,491 identified studies. Across 418 included studies assigned a high-quality score, we extracted 969 parameters, 127 outbreak records, and 154 models. We estimated a pooled total random effect of CZS probability (4.65%), pregnancy loss probability (2.48%), and proportion of symptomatic cases (51.20%). Seroprevalence estimates (n=354) were retrieved beyond South America and French Polynesia. The basic reproduction number estimates (n=77) ranged between 1.12 and 7.4. We found 66 human epidemiological delay estimates, ranging 4-12.1 days for the intrinsic incubation period (n=11), 3-50 days for the infectious period (n=15), 5.1-24.2 days for the extrinsic incubation period (n=22), and 7.4-32.9 days for the serial interval (n=27). These data are available in the R package epireview . Interpretation: This study provides the most comprehensive systematic summary of the ZIKV epidemiological information currently available in the literature. Large heterogeneities and inconsistencies in the reporting of parameter estimates, study designs, and parameter definitions were found, underscoring the need for standardised epidemiological definitions in future publications. Funding: Academy of Medical Sciences, British Heart Foundation, Community Jameel, Diabetes UK, Imperial College London, National Institute for Health and Care Research, Royal Society, Schmidt Sciences, UK Department for Business, Energy, and Industrial Strategy, UK Medical Research Council, United Nations Foundation, Wellcome Trust. Research in Context: Evidence before this study: We searched Web of Science and PubMed from database inception to October 31, 2024 to collate epidemiological parameters, mathematical models, and outbreaks of Zika virus (ZIKV). We found 66 systematic reviews on ZIKV since 2016: of these, 29 focussed on Congenital Zika Syndrome (CZS) and/or Guillain-Barré Syndrome, 19 on ZIKV transmission, 16 on the distribution and burden of ZIKV infection and disease and the remaining on clinical presentation, severity, forecasting, time-to-event, risk mapping and vaccines and therapeutics. The published systematic reviews included between 4 to 204 publications, and most (75%) had a global scope. Only 5 systematic reviews conducted a meta-analysis assessing the overall probability of CZS upon infection in the mother, and only 1 review investigated the probability of CZS by trimester of pregnancy (using 2 articles). Added value of this study: In this systematic review, we extracted information from 574 relevant studies, which makes this the most comprehensive synthesis of Zika parameters for outbreak response to date. From 418 studies with a high-quality assessment score, we extracted 969 epidemiological parameters, 127 outbreak records and 154 transmission models, most of which were published after the 2016 outbreak. We extracted and analysed 354 geolocated and time-stamped seroprevalence estimates; 100 basic and 15 effective reproduction number estimates from 21 countries; 66 human epidemiological delays and 22 extrinsic incubation periods. Our meta-analysis provides robust pooled estimates of the proportion of symptomatic cases (51.20%) and CZS probability (4.65%), also stratified by trimester based on maternal ZIKV infection (based on 7 publications), addressing a gap present in most previous reviews. Implications of all the available evidence: This study provides the most comprehensive overview of ZIKV epidemiology to date, serving as a valuable resource for future modelling studies and stakeholders developing preparedness and response plans to control future outbreaks.
Open AccessA systematic review of Zika virus disease: epidemiological parameters, mathematical models, and outbreaks
Background: Zika virus (ZIKV) is an Aedes-borne arbovirus that can cause serious risks of neurological complications and birth defects in newborns of mothers infected during pregnancy. ZIKV is classified as a priority pathogen by the World Health Organization. Methods: We conducted a systematic review of peer-reviewed studies reporting ZIKV epidemiological parameters, transmission models, and outbreaks (PROSPERO #CRD42023393345) to characterise its transmissibility, seroprevalence, risk factors, disease sequelae, and natural history. We performed a meta-analysis of Congenital Zika Syndrome (CZS), pregnancy loss probabilities amongst ZIKV-infected mothers and the proportion of symptomatic cases. Findings: We extracted information from 574 of 27,491 identified studies. Across 418 included studies assigned a high-quality score, we extracted 969 parameters, 127 outbreak records, and 154 models. We estimated a pooled total random effect of CZS probability (4.65%), pregnancy loss probability (2.48%), and proportion of symptomatic cases (51.20%). Seroprevalence estimates (n=354) were retrieved beyond South America and French Polynesia. The basic reproduction number estimates (n=77) ranged between 1.12 and 7.4. We found 66 human epidemiological delay estimates, ranging 4-12.1 days for the intrinsic incubation period (n=11), 3-50 days for the infectious period (n=15), 5.1-24.2 days for the extrinsic incubation period (n=22), and 7.4-32.9 days for the serial interval (n=27). These data are available in the R package epireview . Interpretation: This study provides the most comprehensive systematic summary of the ZIKV epidemiological information currently available in the literature. Large heterogeneities and inconsistencies in the reporting of parameter estimates, study designs, and parameter definitions were found, underscoring the need for standardised epidemiological definitions in future publications. Funding: Academy of Medical Sciences, British Heart Foundation, Community Jameel, Diabetes UK, Imperial College London, National Institute for Health and Care Research, Royal Society, Schmidt Sciences, UK Department for Business, Energy, and Industrial Strategy, UK Medical Research Council, United Nations Foundation, Wellcome Trust. Research in Context: Evidence before this study: We searched Web of Science and PubMed from database inception to October 31, 2024 to collate epidemiological parameters, mathematical models, and outbreaks of Zika virus (ZIKV). We found 66 systematic reviews on ZIKV since 2016: of these, 29 focussed on Congenital Zika Syndrome (CZS) and/or Guillain-Barré Syndrome, 19 on ZIKV transmission, 16 on the distribution and burden of ZIKV infection and disease and the remaining on clinical presentation, severity, forecasting, time-to-event, risk mapping and vaccines and therapeutics. The published systematic reviews included between 4 to 204 publications, and most (75%) had a global scope. Only 5 systematic reviews conducted a meta-analysis assessing the overall probability of CZS upon infection in the mother, and only 1 review investigated the probability of CZS by trimester of pregnancy (using 2 articles). Added value of this study: In this systematic review, we extracted information from 574 relevant studies, which makes this the most comprehensive synthesis of Zika parameters for outbreak response to date. From 418 studies with a high-quality assessment score, we extracted 969 epidemiological parameters, 127 outbreak records and 154 transmission models, most of which were published after the 2016 outbreak. We extracted and analysed 354 geolocated and time-stamped seroprevalence estimates; 100 basic and 15 effective reproduction number estimates from 21 countries; 66 human epidemiological delays and 22 extrinsic incubation periods. Our meta-analysis provides robust pooled estimates of the proportion of symptomatic cases (51.20%) and CZS probability (4.65%), also stratified by trimester based on maternal ZIKV infection (based on 7 publications), addressing a gap present in most previous reviews. Implications of all the available evidence: This study provides the most comprehensive overview of ZIKV epidemiology to date, serving as a valuable resource for future modelling studies and stakeholders developing preparedness and response plans to control future outbreaks.
- Open Access
- 23 May 2025
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 11 (21) , eadu8678
Socioeconomic and temporal heterogeneity in SARS-CoV-2 exposure and disease in England from May 2020 to February 2023
The impact of COVID-19 varied significantly by deprivation, ethnicity, and policy measures. We analyzed individual-level data on 12,310,485 first SARS-CoV-2 Pillar 2–PCR-confirmed infections, 439,083 hospitalizations, 107,823 deaths, and vaccination records in England from May 2020 to February 2022. Poisson regression models adjusted for demographic and temporal factors showed higher incidence rate ratios (IRRs) for severe outcomes in the most deprived areas compared to the least. We note higher IRRs for severe outcomes for all non-white relative to white ethnicities. The magnitude of IRRs for both deprivation and ethnicities declined from the wild-type to the omicron periods for severe outcomes. For infections, we observed IRRs above one for non-white ethnicities during the wild-type and alpha periods. Vaccination significantly reduced risks across all groups. For severe outcomes, preexisting health inequalities led to large and persistent disparities. For infections, measures must be structured with ethnicity and deprivation in mind early in a pandemic.
- 23 May 2025
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 11 (21) , eadu8678
Open AccessSocioeconomic and temporal heterogeneity in SARS-CoV-2 exposure and disease in England from May 2020 to February 2023
The impact of COVID-19 varied significantly by deprivation, ethnicity, and policy measures. We analyzed individual-level data on 12,310,485 first SARS-CoV-2 Pillar 2–PCR-confirmed infections, 439,083 hospitalizations, 107,823 deaths, and vaccination records in England from May 2020 to February 2022. Poisson regression models adjusted for demographic and temporal factors showed higher incidence rate ratios (IRRs) for severe outcomes in the most deprived areas compared to the least. We note higher IRRs for severe outcomes for all non-white relative to white ethnicities. The magnitude of IRRs for both deprivation and ethnicities declined from the wild-type to the omicron periods for severe outcomes. For infections, we observed IRRs above one for non-white ethnicities during the wild-type and alpha periods. Vaccination significantly reduced risks across all groups. For severe outcomes, preexisting health inequalities led to large and persistent disparities. For infections, measures must be structured with ethnicity and deprivation in mind early in a pandemic.
- Open Access
- 8 May 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-12
A mathematical model of H5N1 influenza transmission in US dairy cattle
2024 saw a novel outbreak of H5N1 avian influenza in US dairy cattle. Limited surveillance data has made determining the true scale of the epidemic difficult. We present a stochastic metapopulation transmission model that simulates H5N1 influenza transmission through individual dairy cows in 35,974 herds in the continental US. Transmission is enabled through the movement of cattle between herds, as indicated from Interstate Certificates of Veterinary Inspection data. We estimate the rates of under-reporting by state and present the anticipated rates of positivity for cattle tested at the point of exportation over time. We investigate the impact of intervention methods on the underlying epidemiological dynamics, demonstrating that current interventions have had insufficient impact, preventing only a mean 175.2 reported outbreaks. Our model predicts that the majority of the disease burden is, as of January 2025, concentrated within West Coast states. We quantify the uncertainty in the scale of the epidemic, highlighting the most pressing data streams to capture, and which states are expected to see outbreaks emerge next, with Arizona and Wisconsin at greatest risk. Our model suggests that dairy outbreaks will continue to occur in 2025, and that more urgent, farm-focused, biosecurity interventions and targeted surveillance schemes are needed.
- 8 May 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-12
Open AccessA mathematical model of H5N1 influenza transmission in US dairy cattle
2024 saw a novel outbreak of H5N1 avian influenza in US dairy cattle. Limited surveillance data has made determining the true scale of the epidemic difficult. We present a stochastic metapopulation transmission model that simulates H5N1 influenza transmission through individual dairy cows in 35,974 herds in the continental US. Transmission is enabled through the movement of cattle between herds, as indicated from Interstate Certificates of Veterinary Inspection data. We estimate the rates of under-reporting by state and present the anticipated rates of positivity for cattle tested at the point of exportation over time. We investigate the impact of intervention methods on the underlying epidemiological dynamics, demonstrating that current interventions have had insufficient impact, preventing only a mean 175.2 reported outbreaks. Our model predicts that the majority of the disease burden is, as of January 2025, concentrated within West Coast states. We quantify the uncertainty in the scale of the epidemic, highlighting the most pressing data streams to capture, and which states are expected to see outbreaks emerge next, with Arizona and Wisconsin at greatest risk. Our model suggests that dairy outbreaks will continue to occur in 2025, and that more urgent, farm-focused, biosecurity interventions and targeted surveillance schemes are needed.
- Open Access
A mathematical model of H5N1 influenza transmission in US dairy cattle
We present a stochastic metapopulation transmission model that simulates the spread of H5N1 avian influenza through individual dairy cows in 35,974 dairy herds in the continental United States. Transmission is enabled through the movement of cattle between herds, as indicated from Interstate Certificates of Veterinary Inspection (ICVI) data. We estimate the rates of under-reporting by state and present the anticipated rates of positivity for cattle tested at the point of exportation over time. We investigate the likely impact of intervention methods to date on the underlying epidemiological dynamics, demonstrating that current interventions have had insufficient impact, preventing only a mean 175.2 reported outbreaks. Our model predicts that the majority of the disease burden is, as of January 2025, concentrated within West Coast states, due to the network of cattle movements and distribution of the respective dairy populations. We quantify the extent of uncertainty in the scale of the epidemic, highlighting the most pressing data streams to capture, and which states are most expected to see outbreaks emerge next, with Arizona and Wisconsin at greatest risk. Our model suggests that dairy herd outbreaks will continue to be a significant public health challenge in 2025, and that more urgent, farm-focused, biosecurity interventions and targeted surveillance schemes are sorely needed.
Open AccessA mathematical model of H5N1 influenza transmission in US dairy cattle
We present a stochastic metapopulation transmission model that simulates the spread of H5N1 avian influenza through individual dairy cows in 35,974 dairy herds in the continental United States. Transmission is enabled through the movement of cattle between herds, as indicated from Interstate Certificates of Veterinary Inspection (ICVI) data. We estimate the rates of under-reporting by state and present the anticipated rates of positivity for cattle tested at the point of exportation over time. We investigate the likely impact of intervention methods to date on the underlying epidemiological dynamics, demonstrating that current interventions have had insufficient impact, preventing only a mean 175.2 reported outbreaks. Our model predicts that the majority of the disease burden is, as of January 2025, concentrated within West Coast states, due to the network of cattle movements and distribution of the respective dairy populations. We quantify the extent of uncertainty in the scale of the epidemic, highlighting the most pressing data streams to capture, and which states are most expected to see outbreaks emerge next, with Arizona and Wisconsin at greatest risk. Our model suggests that dairy herd outbreaks will continue to be a significant public health challenge in 2025, and that more urgent, farm-focused, biosecurity interventions and targeted surveillance schemes are sorely needed.