Multiple models for outbreak decision support in the face of uncertainty.

Policymakers must make management decisions despite incomplete knowledge and conflicting model projections. Little guidance exists for the rapid, representative, and unbiased collection of policy-relevant scientific input from independent modeling teams. Integrating approaches from decision analysis, expert judgment, and model aggregation, we convened multiple modeling teams to evaluate COVID-19 reopening strategies for a mid-sized United States county early in the pandemic. Projections from seventeen distinct models were inconsistent in magnitude but highly consistent in ranking interventions. The 6-mo-ahead aggregate projections were well in line with observed outbreaks in mid-sized US counties. The aggregate results showed that up to half the population could be infected with full workplace reopening, while workplace restrictions reduced median cumulative infections by 82%. Rankings of interventions were consistent across public health objectives, but there was a strong trade-off between public health outcomes and duration of workplace closures, and no win-win intermediate reopening strategies were identified. Between-model variation was high; the aggregate results thus provide valuable risk quantification for decision making. This approach can be applied to the evaluation of management interventions in any setting where models are used to inform decision making. This case study demonstrated the utility of our approach and was one of several multimodel efforts that laid the groundwork for the COVID-19 Scenario Modeling Hub, which has provided multiple rounds of real-time scenario projections for situational awareness and decision making to the Centers for Disease Control and Prevention since December 2020.

Cost effectiveness of preemptive school closures to mitigate pandemic influenza outbreaks of differing severity in the United States.

BACKGROUND: Nonpharmaceutical interventions (NPIs) may be considered as part of national pandemic preparedness as a first line defense against influenza pandemics. Preemptive school closures (PSCs) are an NPI reserved for severe pandemics and are highly effective in slowing influenza spread but have unintended consequences. METHODS: We used results of simulated PSC impacts for a 1957-like pandemic (i.e., an influenza pandemic with a high case fatality rate) to estimate population health impacts and quantify PSC costs at the national level using three geographical scales, four closure durations, and three dismissal decision criteria (i.e., the number of cases detected to trigger closures). At the Chicago regional level, we also used results from simulated 1957-like, 1968-like, and 2009-like pandemics. Our net estimated economic impacts resulted from educational productivity costs plus loss of income associated with providing childcare during closures after netting out productivity gains from averted influenza illness based on the number of cases and deaths for each mitigation strategy. RESULTS: For the 1957-like, national-level model, estimated net PSC costs and averted cases ranged from $7.5 billion (2016 USD) averting 14.5 million cases for two-week, community-level closures to $97 billion averting 47 million cases for 12-week, county-level closures. We found that 2-week school-by-school PSCs had the lowest cost per discounted life-year gained compared to county-wide or school district-wide closures for both the national and Chicago regional-level analyses of all pandemics. The feasibility of spatiotemporally precise triggering is questionable for most locales. Theoretically, this would be an attractive early option to allow more time to assess transmissibility and severity of a novel influenza virus. However, we also found that county-wide PSCs of longer durations (8 to 12 weeks) could avert the most cases (31-47 million) and deaths (105,000-156,000); however, the net cost would be considerably greater ($88-$103 billion net of averted illness costs) for the national-level, 1957-like analysis. CONCLUSIONS: We found that the net costs per death averted ($180,000-$4.2 million) for the national-level, 1957-like scenarios were generally less than the range of values recommended for regulatory impact analyses ($4.6 to 15.0 million). This suggests that the economic benefits of national-level PSC strategies could exceed the costs of these interventions during future pandemics with highly transmissible strains with high case fatality rates. In contrast, the PSC outcomes for regional models of the 1968-like and 2009-like pandemics were less likely to be cost effective; more targeted and shorter duration closures would be recommended for these pandemics.

Correction: Recommended reporting items for epidemic forecasting and prediction research: The EPIFORGE 2020 guidelines.

[This corrects the article DOI: 10.1371/journal.pmed.1003793.].

Addressing delayed case reporting in infectious disease forecast modeling.

Infectious disease forecasting is of great interest to the public health community and policymakers, since forecasts can provide insight into disease dynamics in the near future and inform interventions. Due to delays in case reporting, however, forecasting models may often underestimate the current and future disease burden. In this paper, we propose a general framework for addressing reporting delay in disease forecasting efforts with the goal of improving forecasts. We propose strategies for leveraging either historical data on case reporting or external internet-based data to estimate the amount of reporting error. We then describe several approaches for adapting general forecasting pipelines to account for under- or over-reporting of cases. We apply these methods to address reporting delay in data on dengue fever cases in Puerto Rico from 1990 to 2009 and to reports of influenza-like illness (ILI) in the United States between 2010 and 2019. Through a simulation study, we compare method performance and evaluate robustness to assumption violations. Our results show that forecasting accuracy and prediction coverage almost always increase when correction methods are implemented to address reporting delay. Some of these methods required knowledge about the reporting error or high quality external data, which may not always be available. Provided alternatives include excluding recently-reported data and performing sensitivity analysis. This work provides intuition and guidance for handling delay in disease case reporting and may serve as a useful resource to inform practical infectious disease forecasting efforts.

Recommended reporting items for epidemic forecasting and prediction research: The EPIFORGE 2020 guidelines.

BACKGROUND: The importance of infectious disease epidemic forecasting and prediction research is underscored by decades of communicable disease outbreaks, including COVID-19. Unlike other fields of medical research, such as clinical trials and systematic reviews, no reporting guidelines exist for reporting epidemic forecasting and prediction research despite their utility. We therefore developed the EPIFORGE checklist, a guideline for standardized reporting of epidemic forecasting research. METHODS AND FINDINGS: We developed this checklist using a best-practice process for development of reporting guidelines, involving a Delphi process and broad consultation with an international panel of infectious disease modelers and model end users. The objectives of these guidelines are to improve the consistency, reproducibility, comparability, and quality of epidemic forecasting reporting. The guidelines are not designed to advise scientists on how to perform epidemic forecasting and prediction research, but rather to serve as a standard for reporting critical methodological details of such studies. CONCLUSIONS: These guidelines have been submitted to the EQUATOR network, in addition to hosting by other dedicated webpages to facilitate feedback and journal endorsement.

Surveilling Influenza Incidence With Centers for Disease Control and Prevention Web Traffic Data: Demonstration Using a Novel Dataset.

BACKGROUND: Influenza epidemics result in a public health and economic burden worldwide. Traditional surveillance techniques, which rely on doctor visits, provide data with a delay of 1 to 2 weeks. A means of obtaining real-time data and forecasting future outbreaks is desirable to provide more timely responses to influenza epidemics. OBJECTIVE: This study aimed to present the first implementation of a novel dataset by demonstrating its ability to supplement traditional disease surveillance at multiple spatial resolutions. METHODS: We used internet traffic data from the Centers for Disease Control and Prevention (CDC) website to determine the potential usability of this data source. We tested the traffic generated by 10 influenza-related pages in 8 states and 9 census divisions within the United States and compared it against clinical surveillance data. RESULTS: Our results yielded an r2 value of 0.955 in the most successful case, promising results for some cases, and unsuccessful results for other cases. In the interest of scientific transparency to further the understanding of when internet data streams are an appropriate supplemental data source, we also included negative results (ie, unsuccessful models). Models that focused on a single influenza season were more successful than those that attempted to model multiple influenza seasons. Geographic resolution appeared to play a key role, with national and regional models being more successful, overall, than models at the state level. CONCLUSIONS: These results demonstrate that internet data may be able to complement traditional influenza surveillance in some cases but not in others. Specifically, our results show that the CDC website traffic may inform national- and division-level models but not models for each individual state. In addition, our results show better agreement when the data were broken up by seasons instead of aggregated over several years. We anticipate that this work will lead to more complex nowcasting and forecasting models using this data stream.

Summary results of the 2014-2015 DARPA Chikungunya challenge.

BACKGROUND: Emerging pathogens such as Zika, chikungunya, Ebola, and dengue viruses are serious threats to national and global health security. Accurate forecasts of emerging epidemics and their severity are critical to minimizing subsequent mortality, morbidity, and economic loss. The recent introduction of chikungunya and Zika virus to the Americas underscores the need for better methods for disease surveillance and forecasting. METHODS: To explore the suitability of current approaches to forecasting emerging diseases, the Defense Advanced Research Projects Agency (DARPA) launched the 2014-2015 DARPA Chikungunya Challenge to forecast the number of cases and spread of chikungunya disease in the Americas. Challenge participants (n=38 during final evaluation) provided predictions of chikungunya epidemics across the Americas for a six-month period, from September 1, 2014 to February 16, 2015, to be evaluated by comparison with incidence data reported to the Pan American Health Organization (PAHO). This manuscript presents an overview of the challenge and a summary of the approaches used by the winners. RESULTS: Participant submissions were evaluated by a team of non-competing government subject matter experts based on numerical accuracy and methodology. Although this manuscript does not include in-depth analyses of the results, cursory analyses suggest that simpler models appear to outperform more complex approaches that included, for example, demographic information and transportation dynamics, due to the reporting biases, which can be implicitly captured in statistical models. Mosquito-dynamics, population specific information, and dengue-specific information correlated best with prediction accuracy. CONCLUSION: We conclude that with careful consideration and understanding of the relative advantages and disadvantages of particular methods, implementation of an effective prediction system is feasible. However, there is a need to improve the quality of the data in order to more accurately predict the course of epidemics.

Epidemic Forecasting is Messier Than Weather Forecasting: The Role of Human Behavior and Internet Data Streams in Epidemic Forecast.

Mathematical models, such as those that forecast the spread of epidemics or predict the weather, must overcome the challenges of integrating incomplete and inaccurate data in computer simulations, estimating the probability of multiple possible scenarios, incorporating changes in human behavior and/or the pathogen, and environmental factors. In the past 3 decades, the weather forecasting community has made significant advances in data collection, assimilating heterogeneous data steams into models and communicating the uncertainty of their predictions to the general public. Epidemic modelers are struggling with these same issues in forecasting the spread of emerging diseases, such as Zika virus infection and Ebola virus disease. While weather models rely on physical systems, data from satellites, and weather stations, epidemic models rely on human interactions, multiple data sources such as clinical surveillance and Internet data, and environmental or biological factors that can change the pathogen dynamics. We describe some of similarities and differences between these 2 fields and how the epidemic modeling community is rising to the challenges posed by forecasting to help anticipate and guide the mitigation of epidemics. We conclude that some of the fundamental differences between these 2 fields, such as human behavior, make disease forecasting more challenging than weather forecasting.

Forecasting the 2013-2014 influenza season using Wikipedia.

Infectious diseases are one of the leading causes of morbidity and mortality around the world; thus, forecasting their impact is crucial for planning an effective response strategy. According to the Centers for Disease Control and Prevention (CDC), seasonal influenza affects 5% to 20% of the U.S. population and causes major economic impacts resulting from hospitalization and absenteeism. Understanding influenza dynamics and forecasting its impact is fundamental for developing prevention and mitigation strategies. We combine modern data assimilation methods with Wikipedia access logs and CDC influenza-like illness (ILI) reports to create a weekly forecast for seasonal influenza. The methods are applied to the 2013-2014 influenza season but are sufficiently general to forecast any disease outbreak, given incidence or case count data. We adjust the initialization and parametrization of a disease model and show that this allows us to determine systematic model bias. In addition, we provide a way to determine where the model diverges from observation and evaluate forecast accuracy. Wikipedia article access logs are shown to be highly correlated with historical ILI records and allow for accurate prediction of ILI data several weeks before it becomes available. The results show that prior to the peak of the flu season, our forecasting method produced 50% and 95% credible intervals for the 2013-2014 ILI observations that contained the actual observations for most weeks in the forecast. However, since our model does not account for re-infection or multiple strains of influenza, the tail of the epidemic is not predicted well after the peak of flu season has passed.

Global disease monitoring and forecasting with Wikipedia.

Infectious disease is a leading threat to public health, economic stability, and other key social structures. Efforts to mitigate these impacts depend on accurate and timely monitoring to measure the risk and progress of disease. Traditional, biologically-focused monitoring techniques are accurate but costly and slow; in response, new techniques based on social internet data, such as social media and search queries, are emerging. These efforts are promising, but important challenges in the areas of scientific peer review, breadth of diseases and countries, and forecasting hamper their operational usefulness. We examine a freely available, open data source for this use: access logs from the online encyclopedia Wikipedia. Using linear models, language as a proxy for location, and a systematic yet simple article selection procedure, we tested 14 location-disease combinations and demonstrate that these data feasibly support an approach that overcomes these challenges. Specifically, our proof-of-concept yields models with r2 up to 0.92, forecasting value up to the 28 days tested, and several pairs of models similar enough to suggest that transferring models from one location to another without re-training is feasible. Based on these preliminary results, we close with a research agenda designed to overcome these challenges and produce a disease monitoring and forecasting system that is significantly more effective, robust, and globally comprehensive than the current state of the art.

Quantifying the impact of decay in bed-net efficacy on malaria transmission.

Insecticide-treated nets (ITNs) are at the forefront of malaria control programs and even though the percentage of households in sub-Saharan Africa that owned nets increased from 3% in 2000 to 53% in 2012, many children continue to die from malaria. The potential impact of ITNs on reducing malaria transmission is limited due to inconsistent or improper use, as well as physical decay in effectiveness. Most mathematical models for malaria transmission have assumed a fixed effectiveness rate for bed-nets, which can overestimate the impact of nets on malaria control. We develop a model for malaria spread that captures the decrease in ITN effectiveness due to physical and chemical decay, as well as human behavior as a function of time. We perform uncertainty and sensitivity analyses to identify and rank parameters that play a critical role in malaria transmission. These analyses show that the basic reproduction number R0, and the infectious human population are most sensitive to bed-net coverage and the biting rate of mosquitoes. Our results show the existence of a backward bifurcation for the case in which ITN efficacy is constant over time, which occurs for some range of parameters and is characterized by high malaria mortality in humans. This result implies that bringing R0 to less than one is not enough for malaria elimination but rather additional efforts will be necessary to control the disease. For the case in which ITN efficacy decays over time, we determine coverage levels required to control malaria for different ITN efficacies and demonstrate that ITNs with longer useful lifespans perform better in malaria control. We conclude that malaria control programs should focus on increasing bed-net coverage, which can be achieved by enhancing malaria education and increasing bed-net distribution in malaria endemic regions.

Optimizing human activity patterns using global sensitivity analysis.

Implementing realistic activity patterns for a population is crucial for modeling, for example, disease spread, supply and demand, and disaster response. Using the dynamic activity simulation engine, DASim, we generate schedules for a population that capture regular (e.g., working, eating, and sleeping) and irregular activities (e.g., shopping or going to the doctor). We use the sample entropy (SampEn) statistic to quantify a schedule's regularity for a population. We show how to tune an activity's regularity by adjusting SampEn, thereby making it possible to realistically design activities when creating a schedule. The tuning process sets up a computationally intractable high-dimensional optimization problem. To reduce the computational demand, we use Bayesian Gaussian process regression to compute global sensitivity indices and identify the parameters that have the greatest effect on the variance of SampEn. We use the harmony search (HS) global optimization algorithm to locate global optima. Our results show that HS combined with global sensitivity analysis can efficiently tune the SampEn statistic with few search iterations. We demonstrate how global sensitivity analysis can guide statistical emulation and global optimization algorithms to efficiently tune activities and generate realistic activity patterns. Though our tuning methods are applied to dynamic activity schedule generation, they are general and represent a significant step in the direction of automated tuning and optimization of high-dimensional computer simulations.

The impact of bed-net use on malaria prevalence.

Malaria infection continues to be a major problem in many parts of the world including the Americas, Asia, and Africa. Insecticide-treated bed-nets have shown to reduce malaria cases by 50%; however, improper handling and human behavior can diminish their effectiveness. We formulate and analyze a mathematical model that considers the transmission dynamics of malaria infection in mosquito and human populations and investigate the impact of bed-nets on its control. The effective reproduction number is derived and existence of backward bifurcation is presented. The backward bifurcation implies that the reduction of R below unity alone is not enough to eradicate malaria, except when the initial cases of infection in both populations are small. Our analysis demonstrate that bed-net usage has a positive impact in reducing the reproduction number R. The results show that if 75% of the population were to use bed-nets, malaria could be eliminated. We conclude that more data on the impact of human and mosquito behavior on malaria spread is needed to develop more realistic models and better predictions.

Transmission risk of Oropouche fever across the Americas.

BACKGROUND: Vector-borne diseases (VBDs) are important contributors to the global burden of infectious diseases due to their epidemic potential, which can result in significant population and economic impacts. Oropouche fever, caused by Oropouche virus (OROV), is an understudied zoonotic VBD febrile illness reported in Central and South America. The epidemic potential and areas of likely OROV spread remain unexplored, limiting capacities to improve epidemiological surveillance. METHODS: To better understand the capacity for spread of OROV, we developed spatial epidemiology models using human outbreaks as OROV transmission-locality data, coupled with high-resolution satellite-derived vegetation phenology. Data were integrated using hypervolume modeling to infer likely areas of OROV transmission and emergence across the Americas. RESULTS: Models based on one-support vector machine hypervolumes consistently predicted risk areas for OROV transmission across the tropics of Latin America despite the inclusion of different parameters such as different study areas and environmental predictors. Models estimate that up to 5 million people are at risk of exposure to OROV. Nevertheless, the limited epidemiological data available generates uncertainty in projections. For example, some outbreaks have occurred under climatic conditions outside those where most transmission events occur. The distribution models also revealed that landscape variation, expressed as vegetation loss, is linked to OROV outbreaks. CONCLUSIONS: Hotspots of OROV transmission risk were detected along the tropics of South America. Vegetation loss might be a driver of Oropouche fever emergence. Modeling based on hypervolumes in spatial epidemiology might be considered an exploratory tool for analyzing data-limited emerging infectious diseases for which little understanding exists on their sylvatic cycles. OROV transmission risk maps can be used to improve surveillance, investigate OROV ecology and epidemiology, and inform early detection.

Multi-dimensional resilience: A quantitative exploration of disease outcomes and economic, political, and social resilience to the COVID-19 pandemic in six countries.

The COVID-19 pandemic has highlighted a need for better understanding of countries' vulnerability and resilience to not only pandemics but also disasters, climate change, and other systemic shocks. A comprehensive characterization of vulnerability can inform efforts to improve infrastructure and guide disaster response in the future. In this paper, we propose a data-driven framework for studying countries' vulnerability and resilience to incident disasters across multiple dimensions of society. To illustrate this methodology, we leverage the rich data landscape surrounding the COVID-19 pandemic to characterize observed resilience for several countries (USA, Brazil, India, Sweden, New Zealand, and Israel) as measured by pandemic impacts across a variety of social, economic, and political domains. We also assess how observed responses and outcomes (i.e., resilience) of the COVID-19 pandemic are associated with pre-pandemic characteristics or vulnerabilities, including (1) prior risk for adverse pandemic outcomes due to population density and age and (2) the systems in place prior to the pandemic that may impact the ability to respond to the crisis, including health infrastructure and economic capacity. Our work demonstrates the importance of viewing vulnerability and resilience in a multi-dimensional way, where a country's resources and outcomes related to vulnerability and resilience can differ dramatically across economic, political, and social domains. This work also highlights key gaps in our current understanding about vulnerability and resilience and a need for data-driven, context-specific assessments of disaster vulnerability in the future.

SARS-CoV-2 variant transition dynamics are associated with vaccination rates, number of co-circulating variants, and convalescent immunity.

BACKGROUND: Throughout the COVID-19 pandemic, the SARS-CoV-2 virus has continued to evolve, with new variants outcompeting existing variants and often leading to different dynamics of disease spread. METHODS: In this paper, we performed a retrospective analysis using longitudinal sequencing data to characterize differences in the speed, calendar timing, and magnitude of 16 SARS-CoV-2 variant waves/transitions for 230 countries and sub-country regions, between October 2020 and January 2023. We then clustered geographic locations in terms of their variant behavior across several Omicron variants, allowing us to identify groups of locations exhibiting similar variant transitions. Finally, we explored relationships between heterogeneity in these variant waves and time-varying factors, including vaccination status of the population, governmental policy, and the number of variants in simultaneous competition. FINDINGS: This work demonstrates associations between the behavior of an emerging variant and the number of co-circulating variants as well as the demographic context of the population. We also observed an association between high vaccination rates and variant transition dynamics prior to the Mu and Delta variant transitions. INTERPRETATION: These results suggest the behavior of an emergent variant may be sensitive to the immunologic and demographic context of its location. Additionally, this work represents the most comprehensive characterization of variant transitions globally to date. FUNDING: Laboratory Directed Research and Development (LDRD), Los Alamos National Laboratory.

Economic analysis of the use of facemasks during pandemic (H1N1) 2009.

A large-scale pandemic could cause severe health, social, and economic impacts. The recent 2009 H1N1 pandemic confirmed the need for mitigation strategies that are cost-effective and easy to implement. Typically, in the early stages of a pandemic, as seen with pandemic (H1N1) 2009, vaccines and antivirals may be limited or non-existent, resulting in the need for non-pharmaceutical strategies to reduce the spread of disease and the economic impact. We construct and analyze a mathematical model for a population comprised of three different age groups and assume that some individuals wear facemasks. We then quantify the impact facemasks could have had on the spread of pandemic (H1N1) 2009 and examine their cost effectiveness. Our analyses show that an unmitigated pandemic could result in losses of nearly $832 billion in the United States during the length of the pandemic. Based on present value of future earnings, hospital costs, and lost income estimates due to illness, this study estimates that the use of facemasks by 10%, 25%, and 50% of the population could reduce economic losses by $478 billion, $570 billion, and $573 billion, respectively. The results show that facemasks can significantly reduce the number of influenza cases as well as the economic losses due to a pandemic.

Exploring Impacts to COVID-19 Herd Immunity Thresholds Under Demographic Heterogeneity that Lowers Vaccine Effectiveness.

The COVID-19 pandemic has caused severe health, economic, and societal impacts across the globe. Although highly efficacious vaccines were developed at an unprecedented rate, the heterogeneity in vaccinated populations has reduced the ability to achieve herd immunity. Specifically, as of Spring 2022, the 0-4 year-old population is still unable to be vaccinated and vaccination rates across 5-11 year olds are low. Additionally, vaccine hesitancy for older populations has further stalled efforts to reach herd immunity thresholds. This heterogeneous vaccine landscape increases the challenge of anticipating disease spread in a population. We developed an age-structured Susceptible-Infectious-Recovered-type mathematical model to investigate the impacts of unvaccinated subpopulations on herd immunity. The model considers two types of undervaccination - age-related and behavior-related - by incorporating four age groups based on available FDA-approved vaccines. The model accounts for two different types of vaccines, mRNA (e.g., Pfizer, Moderna) and vector (e.g., Johnson and Johnson), as well as their effectiveness. Our goal is to analyze different scenarios to quantify which subpopulations and vaccine characteristics (e.g., rate or efficacy) most impact infection levels in the United States, using the state of New Mexico as an example.

Assessing K-12 school reopenings under different COVID-19 Spread scenarios - United States, school year 2020/21: A retrospective modeling study.

INTRODUCTION: School-age children play a key role in the spread of airborne viruses like influenza due to the prolonged and close contacts they have in school settings. As a result, school closures and other non-pharmaceutical interventions were recommended as the first line of defense in response to the novel coronavirus pandemic (COVID-19). METHODS: We used an agent-based model that simulates communities across the United States including daycares, primary, and secondary schools to quantify the relative health outcomes of reopening schools for the period of August 15, 2020 to April 11, 2021. Our simulation was carried out in early September 2020 and was based on the latest (at the time) Centers for Disease Control and Prevention (CDC)'s Pandemic Planning Scenarios released in May 2020. We explored different reopening scenarios including virtual learning, in-person school, and several hybrid options that stratify the student population into cohorts in order to reduce exposure and pathogen spread. RESULTS: Scenarios where cohorts of students return to school in non-overlapping formats, which we refer to as hybrid scenarios, resulted in significant decreases in the percentage of symptomatic individuals with COVID-19, by as much as 75%. These hybrid scenarios have only slightly more negative health impacts of COVID-19 compared to implementing a 100% virtual learning scenario. Hybrid scenarios can significantly avert the number of COVID-19 cases at the national scale-approximately between 28 M and 60 M depending on the scenario-over the simulated eight-month period. We found the results of our simulations to be highly dependent on the number of workplaces assumed to be open for in-person business, as well as the initial level of COVID-19 incidence within the simulated community. CONCLUSION: In an evolving pandemic, while a large proportion of people remain susceptible, reducing the number of students attending school leads to better health outcomes; part-time in-classroom education substantially reduces health risks.

A time-varying vulnerability index for COVID-19 in New Mexico, USA using generalized propensity scores.

The coronavirus disease (COVID-19) pandemic has highlighted systemic inequities in the United States and resulted in a larger burden of negative social outcomes for marginalized communities. New Mexico, a state in the southwestern US, has a unique population with a large racial minority population and a high rate of poverty that may make communities more vulnerable to negative social outcomes from COVID-19. To identify which communities may be at the highest relative risk, we created a county-level vulnerability index. After the first COVID-19 case was reported in New Mexico on March 11, 2020, we fit a generalized propensity score model that incorporates sociodemographic factors to predict county-level viral exposure and thus, the generic risk to negative social outcomes such as unemployment or mental health impacts. We used four static sociodemographic covariates important for the state of New Mexico-population, poverty, household size, and minority population-and weekly cumulative case counts to iteratively run our model each week and normalize the exposure score to create a time-varying vulnerability index. We found the relative vulnerability between counties varied in the first eight weeks from the initial COVID-19 case before stabilizing. This framework for creating a location-specific vulnerability index in response to an ongoing disaster may be used as a quick, deployable metric to inform health policy decisions such as allocating state resources to the county level.