Near real-time forecasting for cholera decision making in Haiti after Hurricane Matthew

Developing cholera outbreak forecasting through qualitative dynamics: Insights into Malawi case study

Cholera, an acute diarrheal disease, is a serious concern in developing and underdeveloped areas. A qualitative understanding of cholera epidemics aims to foresee transmission patterns based on reported data and mechanistic models. The mechanistic model is a crucial tool for capturing the dynamics of disease transmission and population spread. However, using real-time cholera cases is essential for forecasting the transmission trend. This prospective study seeks to furnish insights into transmission trends through qualitative dynamics followed by machine learning-based forecasting. The Monte Carlo Markov Chain approach is employed to calibrate the proposed mechanistic model. We identify critical parameters that illustrate the disease's dynamics using partial rank correlation coefficient-based sensitivity analysis. The basic reproduction number as a crucial threshold measures asymptotic dynamics. Furthermore, forward bifurcation directs the stability of the infection state, and Hopf bifurcation suggests that trends in transmission may become unpredictable as societal disinfection rates rise. Further, we develop epidemic-informed machine learning models by incorporating mechanistic cholera dynamics into autoregressive integrated moving averages and autoregressive neural networks. We forecast short-term future cholera cases in Malawi by implementing the proposed epidemic-informed machine learning models to support this. We assert that integrating temporal dynamics into the machine learning models can enhance the capabilities of cholera forecasting models. The execution of this mechanism can significantly influence future trends in cholera transmission. This evolving approach can also be beneficial for policymakers to interpret and respond to potential disease systems. Moreover, our methodology is replicable and adaptable, encouraging future research on disease dynamics.

Impact of a multi-pronged cholera intervention in an endemic setting

Cholera is a bacterial water-borne diarrheal disease transmitted via the fecal-oral route that causes high morbidity in sub-Saharan Africa and Asia. It is preventable with vaccination, and Water, Sanitation, and Hygiene (WASH) improvements. However, the impact of vaccination in endemic settings remains unclear. Cholera is endemic in the city of Kalemie, on the shore of Lake Tanganyika, in the Democratic Republic of Congo, where both seasonal mobility and the lake, a potential environmental reservoir, may promote transmission. Kalemie received a vaccination campaign and WASH improvements in 2013-2016. We assessed the impact of this intervention to inform future control strategies in endemic settings. We fit compartmental models considering seasonal mobility and environmentally-based transmission. We estimated the number of cases the intervention avoided, and the relative contributions of the elements promoting local cholera transmission. We estimated the intervention avoided 5,259 cases (95% credible interval: 1,576.6-11,337.8) over 118 weeks. Transmission did not rely on seasonal mobility and was primarily environmentally-driven. Removing environmental exposure or contamination could control local transmission. Repeated environmental exposure could maintain high population immunity and decrease the impact of vaccination in similar endemic areas. Addressing environmental exposure and contamination should be the primary target of interventions in such settings.

Impact of a multi-pronged cholera intervention in an endemic setting

Cholera is a bacterial water-borne diarrheal disease transmitted via the fecal-oral route that causes high morbidity in sub-Saharan Africa and Asia. It is preventable with vaccination, and Water, Sanitation, and Hygiene (WASH) improvements. However, the impact of vaccination in endemic settings remains unclear. Cholera is endemic in the city of Kalemie, on the shore of Lake Tanganyika, in the Democratic Republic of Congo, where both seasonal mobility and the lake, a potential environmental reservoir, may promote transmission. Kalemie received a vaccination campaign and WASH improvements in 2013-2016. We assessed the impact of this intervention to inform future control strategies in endemic settings. We fit compartmental models considering seasonal mobility and environmentally-based transmission. We estimated the number of cases the intervention avoided, and the relative contributions of the elements promoting local cholera transmission. We estimated the intervention avoided 5,259 cases (95% credible interval: 1,576.6-11,337.8) over 118 weeks. Transmission did not rely on seasonal mobility and was primarily environmentally-driven. Removing environmental exposure or contamination could control local transmission. Repeated environmental exposure could maintain high population immunity and decrease the impact of vaccination in similar endemic areas. Addressing environmental exposure and contamination should be the primary target of interventions in such settings.

Analysis of environmental factors influencing endemic cholera risks in sub-Saharan Africa

The recurring cholera outbreaks in sub-Saharan Africa are of growing concern, especially considering the potential acceleration in the global trend of larger and more lethal cholera outbreaks due to the impacts of climate change. However, there is a scarcity of evidence-based research addressing the environmental and infrastructure factors that sustain cholera recurrence in Africa. This study adopts a statistical approach to investigate over two decades of endemic cholera outbreaks and their relationship with five environmental factors: water provision, sanitation provision, raising temperatures, increased rainfall and GDP. The analysis covers thirteen of the forty-two countries in the mainland sub-Saharan region, collectively representing one-third of the region's territory and half of its population. This breadth enables the findings to be generalised at a regional level. Results from all analyses consistently associate water provision with cholera reduction. The stratified model links increased water provision with a reduction in cholera risk that ranged from 4.2 % to 84.1 % among eight countries (out of 13 countries) as well as a reduction of such risk that ranged from 9.8 % to 68.9 % when there is increased sanitation provision, which was observed in nine countries (out of 13). These results indicate that the population's limited access to water and sanitation, as well as the rise in temperatures, are critical infrastructure and environmental factors contributing to endemic cholera and the heightened risk of outbreaks across the sub-Saharan region. Therefore, these are key areas for targeted interventions and cross-border collaboration to enhance resilience to outbreaks and lead to the end of endemic cholera in the region. However, it is important to interpret the results of this study with caution; hence, further investigation is recommended to conduct a more detailed analysis of the impact of infrastructure and environmental factors on reducing cholera risk.

Informing policy via dynamic models: Cholera in Haiti

Public health decisions must be made about when and how to implement interventions to control an infectious disease epidemic. These decisions should be informed by data on the epidemic as well as current understanding about the transmission dynamics. Such decisions can be posed as statistical questions about scientifically motivated dynamic models. Thus, we encounter the methodological task of building credible, data-informed decisions based on stochastic, partially observed, nonlinear dynamic models. This necessitates addressing the tradeoff between biological fidelity and model simplicity, and the reality of misspecification for models at all levels of complexity. We assess current methodological approaches to these issues via a case study of the 2010-2019 cholera epidemic in Haiti. We consider three dynamic models developed by expert teams to advise on vaccination policies. We evaluate previous methods used for fitting these models, and we demonstrate modified data analysis strategies leading to improved statistical fit. Specifically, we present approaches for diagnosing model misspecification and the consequent development of improved models. Additionally, we demonstrate the utility of recent advances in likelihood maximization for high-dimensional nonlinear dynamic models, enabling likelihood-based inference for spatiotemporal incidence data using this class of models. Our workflow is reproducible and extendable, facilitating future investigations of this disease system.

Optimal allocation strategies for prioritized geographical vaccination for Covid-19

While SARS-CoV-2 vaccine distribution campaigns are underway across the world communities, these efforts face the challenge of effective distribution of limited supplies. We wonder whether suitable spatial allocation strategies might significantly improve a campaignfls efficacy in averting damaging outcomes. In the context of a limited and intermittent COVID-19 supply, we investigate spatial prioritization strategies based on six metrics using the SLIR compartmental epidemic model. We found that the strategy based on the prevalence of susceptible individuals is optimal especially in early interventions and for intermediate values of vaccination rate. It minimizes the cumulative incidence and consequently averts most infections. Our results suggest also that a better performance is obtained if the single batch allocation is supplemented with one or more updating of the priority list. Moreover, the splitting of supply in two or more batches may significantly improve the optimality of the operation.

Optimal control of the spatial allocation of COVID-19 vaccines: Italy as a case study

While campaigns of vaccination against SARS-CoV-2 are underway across the world, communities face the challenge of a fair and effective distribution of a limited supply of doses. Current vaccine allocation strategies are based on criteria such as age or risk. In the light of strong spatial heterogeneities in disease history and transmission, we explore spatial allocation strategies as a complement to existing approaches. Given the practical constraints and complex epidemiological dynamics, designing effective vaccination strategies at a country scale is an intricate task. We propose a novel optimal control framework to derive the best possible vaccine allocation for given disease transmission projections and constraints on vaccine supply and distribution logistics. As a proof-of-concept, we couple our framework with an existing spatially explicit compartmental COVID-19 model tailored to the Italian geographic and epidemiological context. We optimize the vaccine allocation on scenarios of unfolding disease transmission across the 107 provinces of Italy, from January to April 2021. For each scenario, the optimal solution significantly outperforms alternative strategies that prioritize provinces based on incidence, population distribution, or prevalence of susceptibles. Our results suggest that the complex interplay between the mobility network and the spatial heterogeneities implies highly non-trivial prioritization strategies for effective vaccination campaigns. Our work demonstrates the potential of optimal control for complex and heterogeneous epidemiological landscapes at country, and possibly global, scales.

Estimating underdiagnosis of COVID-19 with nowcasting and machine learning

OBJECTIVE

To analyze the underdiagnosis of COVID-19 through nowcasting with machine learning in a Southern Brazilian capital city.

METHODS

Observational ecological design and data from 3916 notified cases of COVID-19 from April 14th to June 2nd, 2020 in Florianópolis, Brazil. A machine-learning algorithm was used to classify cases that had no diagnosis, producing the nowcast. To analyze the underdiagnosis, the difference between data without nowcasting and the median of the nowcasted projections for the entire period and for the six days from the date of onset of symptoms were compared.

RESULTS

The number of new cases throughout the entire period without nowcasting was 389. With nowcasting, it was 694 (95%CI 496-897). During the six-day period, the number without nowcasting was 19 and 104 (95%CI 60-142) with nowcasting. The underdiagnosis was 37.29% in the entire period and 81.73% in the six-day period. The underdiagnosis was more critical in the six days from the date of onset of symptoms to diagnosis before the data collection than in the entire period.

CONCLUSION

The use of nowcasting with machine learning techniques can help to estimate the number of new disease cases.

The epidemicity index of recurrent SARS-CoV-2 infections

Several indices can predict the long-term fate of emerging infectious diseases and the effect of their containment measures, including a variety of reproduction numbers (e.g. [Formula: see text]). Other indices evaluate the potential for transient increases of epidemics eventually doomed to disappearance, based on generalized reactivity analysis. They identify conditions for perturbations to a stable disease-free equilibrium ([Formula: see text]) to grow, possibly causing significant damage. Here, we introduce the epidemicity index e0, a threshold-type indicator: if e0 > 0, initial foci may cause infection peaks even if [Formula: see text]. Therefore, effective containment measures should achieve a negative epidemicity index. We use spatially explicit models to rank containment measures for projected evolutions of the ongoing pandemic in Italy. There, we show that, while the effective reproduction number was below one for a sizable timespan, epidemicity remained positive, allowing recurrent infection flare-ups well before the major epidemic rebounding observed in the fall.

Infectious and non-infectious diseases burden among Haitian immigrants in Chile: a cross-sectional study

Chile has become a popular destination for migrants from South America and the Caribbean (low- and middle-income countries migration). Close to 200.000 Haitian migrants have arrived in Chile. Infectious and non-infectious disease burden among the Haitian adult population living in Chile is unknown. This study aimed to acquire the basic health information (selected transmissible and non-transmissible conditions) of the Haitian adult population living in Chile. A cross-sectional survey was performed, inviting Haitian-born residents in Chile older than 18 years old. Common conditions and risk factors for disease were assessed, as well as selected transmissible conditions (HIV, HBV, and HCV). 498 participants (60.4% female) from 10 communities in two regions of Chile were surveyed. Most subjects had never smoked (91.5%), and 80% drank less than one alcohol unit per month. The mean BMI was 25.6, with 45% of participants having a normal BMI (20-25). Hypertension was present in 31.5% (33% in the 25-44 age group). Prevalence of HIV was 2.4% (95 CI 1.3-4.2%), hepatitis B (HBsAg positive) was 3.4% (95 CI 2.1-5.5%), and hepatitis C was 0% (95 CI 0.0-0.9%). Quality of life showed a significant prevalence of depression and anxiety markers, particularly in those arriving in Chile less than 1 year ago. Low prevalence of obesity, diabetes, smoking, and drinking and estimated cardiovascular risk were found. Nonetheless, hypertension at a younger age, disproportionately higher prevalence of HIV and HBV infection and frequent markers of anxiety and depression were also found. Public policies for detecting and treating hypertension, HIV, and HBV screening, offering HBV vaccination, and organizing mental health programs for Haitian immigrants, are urgently needed.

Achieving coordinated national immunity and cholera elimination in Haiti through vaccination: a modelling study

BACKGROUND

Cholera was introduced into Haiti in 2010. Since then, more than 820 000 cases and nearly 10 000 deaths have been reported. Oral cholera vaccine (OCV) is safe and effective, but has not been seen as a primary tool for cholera elimination due to a limited period of protection and constrained supplies. Regionally, epidemic cholera is contained to the island of Hispaniola, and the lowest numbers of cases since the epidemic began were reported in 2019. Hence, Haiti may represent a unique opportunity to eliminate cholera with OCV.

METHODS

In this modelling study, we assessed the probability of elimination, time to elimination, and percentage of cases averted with OCV campaign scenarios in Haiti through simulations from four modelling teams. For a 10-year period from January 19, 2019, to Jan 13, 2029, we compared a no vaccination scenario with five OCV campaign scenarios that differed in geographical scope, coverage, and rollout duration. Teams used weekly department-level reports of suspected cholera cases from the Haiti Ministry of Public Health and Population to calibrate the models and used common vaccine-related assumptions, but other model features were determined independently.

FINDINGS

Among campaigns with the same vaccination coverage (70% fully vaccinated), the median probability of elimination after 5 years was 0-18% for no vaccination, 0-33% for 2-year campaigns focused in the two departments with the highest historical incidence, 0-72% for three-department campaigns, and 35-100% for nationwide campaigns. Two-department campaigns averted a median of 12-58% of infections, three-department campaigns averted 29-80% of infections, and national campaigns averted 58-95% of infections. Extending the national campaign to a 5-year rollout (compared to a 2-year rollout), reduced the probability of elimination to 0-95% and the proportion of cases averted to 37-86%.

INTERPRETATION

Models suggest that the probability of achieving zero transmission of Vibrio cholerae in Haiti with current methods of control is low, and that bolder action is needed to promote elimination of cholera from the region. Large-scale cholera vaccination campaigns in Haiti would offer the opportunity to synchronise nationwide immunity, providing near-term population protection while improvements to water and sanitation promote long-term cholera elimination.

FUNDING

Bill & Melinda Gates Foundation, Global Good Fund, Institute for Disease Modeling, Swiss National Science Foundation, and US National Institutes of Health.

Spatial dynamics and the basic reproduction number of the 1991-1997 Cholera epidemic in Peru

After being cholera free for over 100 years, Peru experienced an unprecedented epidemic of Vibrio cholerae O1 that began in 1991 and generated multiple waves of disease over several years. We developed a mechanistic transmission model that accounts for seasonal variation in temperature to estimate spatial variability in the basic reproduction number ([Formula: see text]), the initial concentration of vibrios in the environment, and cholera reporting rates. From 1991-1997, cholera spread following a multi-wave pattern, with weekly incidence concentrated during warm seasons. The epidemic first hit the coastal departments of Peru and subsequently spread through the highlands and jungle regions. The correlation between model predictions and observations was high (range in R2: 58% to 97%). Department-level population size and elevation explained significant variation in spatial-temporal transmission patterns. The overall R0 across departments was estimated at 2.1 (95% CI: 0.8,7.3), high enough for sustained transmission. Geographic-region level [Formula: see text] varied substantially from 2.4 (95% CI: 1.1, 7.3) for the coastal region, 1.9 (0.7, 6.4) for the jungle region, and 1.5 (0.9, 2.2) for the highlands region. At the department level, mean [Formula: see text] ranged from 0.8 to 6.9. Department-level [Formula: see text] were correlated with overall observed attack rates (Spearman ρ = 0.59, P = 0.002), elevation (ρ = -0.4, P = 0.04), and longitude (ρ = -0.6, P = 0.004). We find that both [Formula: see text] and the initial concentration of vibrios were higher in coastal departments than other departments. Reporting rates were low, consistent with a substantial fraction of asymptomatic or mild cases associated with the El Tor cholera biotype. Our results suggest that cholera vibrios, autochthonous to plankton in the natural aquatic environment, may have triggered outbreaks in multiple coastal locations along the Pacific coast of Peru. Our methodology could be useful to investigate multi-wave epidemics of cholera and could be extended to conduct near real-time forecasts and investigate the impact of vaccination strategies.

Spread and dynamics of the COVID-19 epidemic in Italy: Effects of emergency containment measures

The spread of coronavirus disease 2019 (COVID-19) in Italy prompted drastic measures for transmission containment. We examine the effects of these interventions, based on modeling of the unfolding epidemic. We test modeling options of the spatially explicit type, suggested by the wave of infections spreading from the initial foci to the rest of Italy. We estimate parameters of a metacommunity Susceptible-Exposed-Infected-Recovered (SEIR)-like transmission model that includes a network of 107 provinces connected by mobility at high resolution, and the critical contribution of presymptomatic and asymptomatic transmission. We estimate a generalized reproduction number ([Formula: see text] = 3.60 [3.49 to 3.84]), the spectral radius of a suitable next-generation matrix that measures the potential spread in the absence of containment interventions. The model includes the implementation of progressive restrictions after the first case confirmed in Italy (February 21, 2020) and runs until March 25, 2020. We account for uncertainty in epidemiological reporting, and time dependence of human mobility matrices and awareness-dependent exposure probabilities. We draw scenarios of different containment measures and their impact. Results suggest that the sequence of restrictions posed to mobility and human-to-human interactions have reduced transmission by 45% (42 to 49%). Averted hospitalizations are measured by running scenarios obtained by selectively relaxing the imposed restrictions and total about 200,000 individuals (as of March 25, 2020). Although a number of assumptions need to be reexamined, like age structure in social mixing patterns and in the distribution of mobility, hospitalization, and fatality, we conclude that verifiable evidence exists to support the planning of emergency measures.

Assessing the potential impact of vector-borne disease transmission following heavy rainfall events: a mathematical framework

Predicting the impact of natural disasters such as hurricanes on the transmission dynamics of infectious diseases poses significant challenges. In this paper, we put forward a simple modelling framework to investigate the impact of heavy rainfall events (HREs) on mosquito-borne disease transmission in temperate areas of the world such as the southern coastal areas of the USA. In particular, we explore the impact of the timing of HREs relative to the transmission season via analyses that test the sensitivity of HRE-induced epidemics to variation in the effects of rainfall on the dynamics of mosquito breeding capacity, and the intensity and temporal profile of human population displacement patterns. The recent Hurricane Harvey in Texas motivates the simulations reported. Overall, we find that the impact of vector-borne disease transmission is likely to be greater the earlier the HREs occur in the transmission season. Simulations based on data for Hurricane Harvey suggest that the limited impact it had on vector-borne disease transmission was in part because of when it occurred (late August) relative to the local transmission season, and in part because of the mitigating effect of the displacement of people. We also highlight key data gaps related to models of vector-borne disease transmission in the context of natural disasters. This article is part of the theme issue 'Modelling infectious disease outbreaks in humans, animals and plants: approaches and important themes'. This issue is linked with the subsequent theme issue 'Modelling infectious disease outbreaks in humans, animals and plants: epidemic forecasting and control'.

Conditions for transient epidemics of waterborne disease in spatially explicit systems

Waterborne diseases are a diverse family of infections transmitted through ingestion of-or contact with-water infested with pathogens. Outbreaks of waterborne infections often show well-defined spatial signatures that are typically linked to local eco-epidemiological conditions, water-mediated pathogen transport and human mobility. In this work, we apply a spatially explicit network model describing the transmission cycle of waterborne pathogens to determine invasion conditions in metacommunities endowed with a realistic spatial structure. Specifically, we aim to define conditions under which pathogens can temporarily colonize a set of human communities, thus triggering a transient epidemic outbreak. To that end, we apply generalized reactivity analysis, a recently developed methodological framework for the study of transient dynamics in ecological systems subject to external perturbations. The study of pathogen invasion is complemented by the detection of the spatial signatures associated with the perturbations to a disease-free system that are expected to be amplified the most over different time scales. Understanding the drivers of waterborne disease dynamics over time scales that are relevant to epidemic and/or endemic transmission is a crucial, cross-disciplinary challenge, as large portions of the developing world still struggle to cope with the burden of these infections.

Unexplored Opportunities: Use of Climate- and Weather-Driven Early Warning Systems to Reduce the Burden of Infectious Diseases

PURPOSE OF REVIEW

Weather and climate influence multiple aspects of infectious disease ecology. Creating and applying early warning systems based on temperature, precipitation, and other environmental data can identify where and when outbreaks of climate-sensitive infectious diseases could occur and can be used by decision makers to allocate resources. Whether an outbreak actually occurs depends heavily on other social, political, and institutional factors.

RECENT FINDINGS

Improving the timing and confidence of seasonal climate forecasting, coupled with knowledge of exposure-response relationships, can identify prior conditions conducive to disease outbreaks weeks to months in advance of outbreaks. This information could then be used by public health professionals to improve surveillance in the most likely areas for threats. Early warning systems are well established for drought and famine. And while weather- and climate-driven early warning systems for certain diseases, such as dengue fever and cholera, are employed in some regions, this area of research is underdeveloped. Early warning systems based on temperature, precipitation, and other environmental data provide an opportunity for early detection leading to early action and response to potential pathogen threats, thereby reducing the burden of disease when compared with passive health indicator-based surveillance systems.