Search

SUMMARYBackgroundVenezuela has experienced an explosive resurgence in Plasmodium falciparum and Plasmodium vivax malaria incidence over the last decade due to various social, political, and economic factors. To ensure national and regional progress towards malaria elimination, there is an urgent need to better understand the epidemiological dynamics of this malaria outbreak at its epicenter in the southern state of Bolívar and to identify the sets of interventions that may be necessary to reduce transmission and incidence.MethodsWe fitted transmission models of P. falciparum and P. vivax to weekly incidence data in Bolívar, Venezuela during 2000-2018. We estimated the magnitude of local transmission for both Plasmodium spp. and inferred the contribution of relapses and reinfections to P. vivax incidence in the region. Compared to a business-as-usual scenario, we projected the impact of different interventions on Plasmodium spp. incidence during the period 2021-2023.FindingsWe estimated that 63·7 – 73·3% of all P. vivax infections in Bolívar are relapses, leading to as many as 51,800 observed relapses misclassified per year as reinfections in the routine surveillance data. Our estimates suggest that the reproduction number remains close to one for both Plasmodium spp., pointing towards the feasibility of control. Long-lasting insecticidal nets (LLINs) were projected to cause greater proportional reductions in P. falciparum incidence than P. vivax incidence, and mass drug administration (MDA) with an 8-aminoquinoline and a blood-stage partner drug was projected to cause the greatest reduction in P. vivax incidence, provided that adherence rates were high.InterpretationControl of the malaria outbreak in Southeastern Venezuela is feasible, should appropriate resources to support surveillance and control be brought to bear. Coupling the distribution of LLINs and a focal MDA with an 8-aminoquinoline and a blood-stage partner drug may lead to the greatest reduction in malaria incidence.FundingNational Science Foundation; University of Notre Dame; National Institute of General Medical Sciences (grant number 1R35GM143029-01 to TAP);RESEARCH IN CONTEXTEvidence before the studyWe searched PubMed, bioRxiv, and medRxiv for articles in English published on or before May 25th, 2021 using the following keywords: “Venezuela”, “malaria”, AND “model*”. Previous studies have applied statistical models to characterize the relationship between malaria incidence and climate in Venezuela, concluding that the reproduction number is low and suggesting the feasibility of control. A study fitting a mechanistic transmission model to epidemiological data to allow for projecting the impact of alternative approaches to control has not been performed.Added value of the studyWe fitted Plasmodium falciparum and Plasmodium vivax transmission models to 20 years of weekly incidence data to estimate the transmission of both Plasmodium spp. and characterize the contribution of relapses and reinfections to P. vivax incidence in Bolívar, Venezuela. We also projected the likely impact of interventions in the region under alternative scenarios about control.Implications of the available evidenceThe burden of Plasmodium vivax relapses in Bolívar is underestimated from routine surveillance data, so control interventions must target the hypnozoite reservoir in the region. Mass drug administration (MDA) is projected to be impactful for both Plasmodium spp., though tradeoffs between coverage and adherence suggest that a focal MDA with an 8-aminoquinoline and a blood-stage partner drug may yield the greatest impact.

Geographically stratified estimates of disease burden play an important role in setting priorities for the management of different diseases and for targeting interventions against a single disease. Such estimates involve numerous assumptions, which uncertainty about is not always well accounted for. We developed a framework for estimating the burden of yellow fever in Africa and evaluated its sensitivity to assumptions about the interpretation of serological data and choice of regression model. We addressed the latter with an ensemble approach, and we found that the former resulted in a nearly twentyfold difference in burden estimates (range of central estimates: 8.4×104-1.5×106 deaths in 2021-2030). Even so, statistical uncertainty made even greater contributions to variance in burden estimates (87%). Combined with estimates that most infections go unreported (range of 95% credible intervals: 99.65-99.99%), our results suggest that yellow fever’s burden will remain highly uncertain without major improvements in surveillance.

The war in Tigray region of Ethiopia that started in November 2020 and is still ongoing has brought enormous damage to the health system. This analysis provides an assessment of the health system before and during the war. Evidence of damage was compiled from November 2020 to June 2021 from various reports by the interim government of Tigray, and also by international non-governmental organisations. Comparison was made with data from the prewar calendar year. Six months into the war, only 30% of hospitals, 17% of health centres, 11.5% of ambulances and none of the 712 health posts were functional. As of June 2021, the population in need of emergency food assistance in Tigray increased from less than one million to over 5.2 million. While the prewar performance of antenatal care, supervised delivery, postnatal care and children vaccination was 64%, 73%, 63% and 73%, respectively, but none of the services were likely to be delivered in the first 90 days of the war. A conservative estimate places the number of girls and women raped in the first 5 months of the war to be 10 000. These data indicate a widespread destruction of livelihoods and a collapse of the healthcare system. The use of hunger and rape as a weapon of war and the targeting of healthcare facilities are key components of the war. To avert worsening conditions, an immediate intervention is needed to deliver food and supplies and rehabilitate the healthcare delivery system and infrastructure.

When new pathogens emerge, numerous questions arise about their future spread, some of which can be addressed with probabilistic forecasts. The many uncertainties about the epidemiology of emerging pathogens can make it difficult to choose among model structures and assumptions, however. To assess the potential for uncertainties about emerging pathogens to affect forecasts of their spread, we evaluated the performance of a suite of 16 forecasting models in the context of the 2015-2016 Zika epidemic in Colombia. Each model featured a different combination of assumptions about the role of human mobility in driving transmission, spatiotemporal variation in transmission potential, and the number of times the virus was introduced. All models used the same core transmission model and the same iterative data assimilation algorithm to generate forecasts. By assessing forecast performance through time using logarithmic scoring with ensemble weighting, we found that which model assumptions had the most ensemble weight changed through time. In particular, spatially coupled models had higher ensemble weights in the early and late phases of the epidemic, whereas non-spatial models had higher ensemble weights at the peak of the epidemic. We compared forecast performance of the equally-weighted ensemble model to each individual model and identified a trade-off whereby certain individual models outperformed the ensemble model early in the epidemic but the ensemble model outperformed all individual models on average. On balance, our results suggest that suites of models that span uncertainty across alternative assumptions are necessary to obtain robust forecasts in the context of emerging infectious diseases.

Prediction of spatial heterogeneity in disease incidence based on measurable spatial factors is a major goal of spatial epidemiology. There are a number of applied goals of these predictions, including appropriately targeting resources for surveillance and intervention and accurately quantifying disease burden. Although spatial heterogeneity is evident in the epidemiology of many diseases, several aspects of the biology of vector-borne diseases amplify this form of heterogeneity. Here, we review several aspects of this biology, highlighting seven distinct ways in which the biology of vector-borne diseases impacts understanding spatial heterogeneity in disease incidence. Whereas traditional methods place emphasis on spatial regression and other forms of statistical analysis of empirical data, the goal here is to offer a perspective on potential pitfalls of analyses that take data at face value and do not acknowledge the complex, nonlinear, and dynamic relationships between spatial patterns of disease incidence and spatial heterogeneity in transmission.

Several hundred thousand Zika cases have been reported across the Americas since 2015. Incidence of infection was likely much higher, however, due to a high frequency of asymptomatic infection and other challenges that surveillance systems faced. Using a hierarchical Bayesian model with empirically-informed priors, we leveraged multiple types of Zika case data from 15 countries to estimate subnational reporting probabilities and infection attack rates (IARs). Zika IAR estimates ranged from 0.084 (95% CrI: 0.067–0.096) in Peru to 0.361 (95% CrI: 0.214–0.514) in Ecuador, with significant subnational variability in every country. Totaling infection estimates across these and 33 other countries and territories, our results suggest that 132.3 million (95% CrI: 111.3-170.2 million) people in the Americas had been infected by the end of 2018. These estimates represent the most extensive attempt to determine the size of the Zika epidemic in the Americas, offering a baseline for assessing the risk of future Zika epidemics in this region., Author summary During the recent Zika epidemic in the Americas millions of people were likely infected, but the true size of the epidemic is unknown because of gaps in the surveillance system. The infection attack rate (IAR)—defined as the proportion of the population that was infected over the course of the epidemic—has important implications for the longer-term epidemiology of Zika in the region, such as the timing, location, and likelihood of future outbreaks. To estimate the IAR and the total number of people infected, we leveraged multiple types of Zika case data from 15 countries and territories where subnational data were publicly available. Datasets included confirmed and suspected Zika cases in pregnant women and in the total population, Zika-associated Guillan-Barré syndrome cases, and cases of congenital Zika syndrome. We used a hierarchical Bayesian model with empirically-informed priors that leveraged the different case report types to simultaneously estimate national and subnational reporting probabilities, the fraction of symptomatic infections, and subnational IARs. In these 15 countries and territories, estimates of Zika IAR ranged from 0.084 (95% CrI: 0.067–0.096) in Peru to 0.361 (95% CrI: 0.214–0.514) in Ecuador. Totaling these infection estimates across these and 33 other countries and territories in the region, our results suggest that 132.3 million (95% CrI: 111.3-170.2 million) people in the Americas were infected with ZIKV by the end of 2018.

IntroductionSince its emergence in late December 2019, COVID-19 has rapidly developed into a pandemic in mid of March with many countries suffering heavy human loss and declaring emergency conditions to contain its spread. The impact of the disease, while it has been relatively low in the sub-Saharan Africa (SSA) as of May 2020, is feared to be potentially devastating given the less developed and fragmented healthcare system in the continent. In addition, most emergency measures practised may not be effective due to their limited affordability as well as the communal way people in SSA live in relative isolation in clusters of large as well as smaller population centres.MethodsTo address the acute need for estimates of the potential impacts of the disease once it sweeps through the African region, we developed a process-based model with key parameters obtained from recent studies, taking local context into consideration. We further used the model to estimate the number of infections within a year of sustained local transmissions under scenarios that cover different population sizes, urban status, effectiveness and coverage of social distancing, contact tracing and usage of cloth face mask.ResultsWe showed that when implemented early, 50% coverage of contact tracing and face mask, with 33% effective social distancing policies can bringing the epidemic to a manageable level for all population sizes and settings we assessed. Relaxing of social distancing in urban settings from 33% to 25% could be matched by introduction and maintenance of face mask use at 43%.ConclusionsIn SSA countries with limited healthcare workforce, hospital resources and intensive care units, a robust system of social distancing, contact tracing and face mask use could yield in outcomes that prevent several millions of infections and thousands of deaths across the continent.

Predicting an infectious disease can help reduce its impact by advising public health interventions and personal preventive measures. While availability of heterogeneous data streams and sensors such as satellite imagery and the Internet have increased the opportunity to indirectly measure, understand, and predict global dynamics, the data may be prohibitively large and/or require intensive data management while also requiring subject matter experts to properly exploit the data sources (e.g., deriving features from fundamentally different data sets). Few efforts have quantitatively assessed the predictive benefit of novel data streams in comparison to more traditional data sources, especially at fine spatio-temporal resolutions. We have combined multiple traditional and non-traditional data streams (satellite imagery, Internet, weather, census, and clinical surveillance data) and assessed their combined ability to predict dengue in Brazil’s 27 states on a weekly and yearly basis over seven years. For each state, we nowcast dengue based on several time series models, which vary in complexity and inclusion of exogenous data. We also predict yearly cumulative risk by municipality and state. The top-performing model and utility of predictive data varies by state, implying that forecasting and nowcasting efforts in the future may be made more robust by and benefit from the use of multiple data streams and models. One size does not fit all, particularly when considering state-level predictions as opposed to the whole country. Our first-of-its-kind high resolution flexible system for predicting dengue incidence with heterogeneous (and still sometimes sparse) data can be extended to multiple applications and regions.

Human mobility, both short and long term, are important considerations in the study of numerous systems. Economic and technological advances have led to a more interconnected global community, further increasing the need for considerations of human mobility. While data on human mobility are better recorded in many developed countries, availability of such data remains limited in many low- and middle-income countries around the world, particularly at the fine temporal and spatial scales required by many applications. In this study, we used 5-year census-based internal migration microdata for 32 departments in Colombia (i.e., Admin-1 level) to develop a novel spatial interaction modeling approach for estimating migration, at a finer spatial scale, among the 1,122 municipalities in the country (i.e., Admin-2 level). Our modeling approach bridges a significant gap in the availability of migration data at administrative unit levels finer than those at which migration data are typically recorded. Due to the widespread availability of census-based migration microdata at the Admin-1 level, our modeling approach opens up for the possibilities of modeling migration patterns at Admin-2 and Admin-3 levels across many other countries where such data are currently lacking.

Since the first Zika virus (ZIKV) infection was confirmed in Brazil in May 2015, several hundred thousand cases have been reported across the Americas. This figure gives an incomplete picture of the epidemic, however, given that asymptomatic infections, imperfect surveillance, and variability in reporting rates imply that the incidence of infection was likely much higher. The infection attack rate (IAR)—defined as the proportion of the population that was infected over the course of the epidemic—has important implications for the longer-term epidemiology of Zika in the region, such as the timing, location, and likelihood of future outbreaks. To estimate the IAR and the total number of people infected, we leveraged multiple types of Zika case data from 15 countries and territories where subnational data were publicly available. Datasets included confirmed and suspected Zika cases in pregnant women and in the total population, Zika-associated Guillan-Barré syndrome cases, and cases of congenital Zika syndrome. We used a hierarchical Bayesian model with empirically-informed priors that leveraged the different case report types to simultaneously estimate national and subnational reporting rates, the fraction of symptomatic infections, and subnational IARs. In these 15 countries and territories, estimates of Zika IAR ranged from 0.084 (95% CrI: 0.067 − 0.096) in Peru to 0.361 (95% CrI: 0.214 − 0.514) in Ecuador, with significant subnational variability in IAR for every country. Totaling these infection estimates across these and 33 other countries and territories in the region, our results suggest that 132.3 million (95% CrI: 111.3-170.2 million) people in the Americas have been infected by ZIKV since 2015. These estimates represent the most extensive attempt to date to determine the size of the Zika epidemic in the Americas, and they offer an important baseline for assessing the risk of future Zika epidemics in this region.

Using data for the states of Brazil, we construct a polynomial distributed lag model under different truncation lag criteria to predict reported dengue cases. Accurately predicting dengue cases provides the framework to develop forecasting models, which would provide public health professionals time to create targeted interventions for areas at high risk of dengue outbreaks. Others have shown that variables of interest such as temperature and vegetation can be used to predict dengue cases. These models did not detail how truncation lag criteria was chosen for their respective models when polynomial distributed lag was used. We explore current truncation lag selection methods used widely in the literature (marginal and minimized AIC) and determine which of these methods works best for our given data set. While minimized AIC truncation lag selection produced the best fit to our data, this method used substantially more data to inform its prediction compared to the marginal truncation lag selection method. Finally, the following variables were found to be significant predictors of dengue in this region: normalized difference vegetation index (NDVI), green-based normalized difference water index (NDWI), normalized burn ratio (NBR), and temperature. These best predictors were derived from multispectral remote sensing imagery as well as temperature data.

Despite a long history of mosquito-borne virus epidemics in the Americas, the impact of the Zika virus (ZIKV) epidemic of 2015–2016 was unexpected. The need for scientifically informed decision-making is driving research to understand the emergence and spread of ZIKV. To support that research, we assembled a data set of key covariates for modeling ZIKV transmission dynamics in Colombia, where ZIKV transmission was widespread and the government made incidence data publically available. On a weekly basis between January 1, 2014 and October 1, 2016 at three administrative levels, we collated spatiotemporal Zika incidence data, nine environmental variables, and demographic data into a single downloadable database. These new datasets and those we identified, processed, and assembled at comparable spatial and temporal resolutions will save future researchers considerable time and effort in performing these data processing steps, enabling them to focus instead on extracting epidemiological insights from this important data set. Similar approaches could prove useful for filling data gaps to enable epidemiological analyses of future disease emergence events.

Background: Temporal incidence patterns provide a crucial window into the dynamics of emerging infectious diseases, yet their utility is limited by the spatially aggregated form in which they are often presented. Weekly incidence data from the 2015-2016 Zika epidemic were available only at the national level for most countries in the Americas. One exception was Colombia, where data at departmental and municipal scales were made publicly available in real time, providing an opportunity to assess the degree to which national-level data are reflective of temporal patterns at local levels. Methods: To characterize differences in epidemic trajectories, our analysis centered on classifying proportional cumulative incidence curves according to six features at three levels of spatial aggregation. This analysis used the partitioning around medoids algorithm to assign departments and municipalities to groups based on these six characteristics. Examination of the features that differentiated these groups and exploration of their temporal and spatial patterns were performed. Simulations from a stochastic transmission model provided data that were used to assess the extent to which groups identified by the classification algorithm could be associated with differences in underlying drivers of transmission. Results: The timing of departmental-level epidemic peaks varied by three months, and departmental-level estimates of the time-varying reproduction number, R(t), showed patterns that were distinct from a national-level estimate. The classification algorithm identified moderate support for two to three clusters at the departmental level and somewhat stronger support for this at the municipal level. Variability in epidemic duration, the length of the tail of the epidemic, and the consistency of cumulative incidence data with a cumulative normal distribution function made the greatest contributions to distinctions across these groups. Applying the classification algorithm to simulated data showed that municipalities with basic reproduction number, R0, greater than 1 were consistently associated with a particular group. Municipalities with R0 < 1 displayed more diverse patterns, although in this case that may be due to simplifications of how the model represented spatial interaction among municipalities. Conclusions: The diversity of temporal incidence patterns at local scales uncovered by this analysis underscores the value of spatially disaggregated data and the importance of locally tailored strategies for responding to emerging infectious diseases.

Background Mathematical models of transmission dynamics are routinely fitted to epidemiological time series, which must inevitably be aggregated at some spatial scale. Weekly case reports of chikungunya have been made available nationally for numerous countries in the Western Hemisphere since late 2013, and numerous models have made use of this data set for forecasting and inferential purposes. Motivated by an abundance of literature suggesting that the transmission of this mosquito-borne pathogen is localized at scales much finer than nationally, we fitted models at three different spatial scales to weekly case reports from Colombia to explore limitations of analyses of nationally aggregated time series data. Methods We adapted the recently developed Disease Transmission Kernel (DTK)-Dengue model for modeling chikungunya virus (CHIKV) transmission, given the numerous similarities of these viruses vectored by a common mosquito vector. We fitted versions of this model specified at different spatial scales to weekly case reports aggregated at different spatial scales: (1) single-patch national model fitted to national data; (2) single-patch departmental models fitted to departmental data; and (3) multi-patch departmental models fitted to departmental data, where the multiple patches refer to municipalities within a department. We compared the consistency of simulations from fitted models with empirical data. Results We found that model consistency with epidemic dynamics improved with increasing spatial granularity of the model. Specifically, the sum of single-patch departmental model fits better captured national-level temporal patterns than did a single-patch national model. Likewise, multi-patch departmental model fits better captured department-level temporal patterns than did single-patch departmental model fits. Furthermore, inferences about municipal-level incidence based on multi-patch departmental models fitted to department-level data were positively correlated with municipal-level data that were withheld from model fitting. Conclusions Our model performed better when posed at finer spatial scales, due to better matching between human populations with locally relevant risk. Confronting spatially aggregated models with spatially aggregated data imposes a serious structural constraint on model behavior by averaging over epidemiologically meaningful spatial variation in drivers of transmission, impairing the ability of models to reproduce empirical patterns. Electronic supplementary material The online version of this article (10.1186/s12916-018-1127-2) contains supplementary material, which is available to authorized users.

Numerous Zika virus vaccines are being developed. However, identifying sites to evaluate the efficacy of a Zika virus vaccine is challenging due to the general decrease in Zika virus activity. We compare results from three different modeling approaches to estimate areas that may have increased relative risk of Zika virus transmission during 2017. The analysis focused on eight priority countries (i.e., Brazil, Colombia, Costa Rica, Dominican Republic, Ecuador, Mexico, Panama, and Peru). The models projected low incidence rates during 2017 for all locations in the priority countries but identified several subnational areas that may have increased relative risk of Zika virus transmission in 2017. Given the projected low incidence of disease, the total number of participants, number of study sites, or duration of study follow-up may need to be increased to meet the efficacy study endpoints.

Epidemic growth rate, r, provides a more complete description of the potential for epidemics than the more commonly studied basic reproduction number, R0, yet the former has never been described as a function of temperature for dengue virus or other pathogens with temperature-sensitive transmission. The need to understand the drivers of epidemics of these pathogens is acute, with arthropod-borne virus epidemics becoming increasingly problematic. We addressed this need by developing temperature-dependent descriptions of the two components of r—R0 and the generation interval—to obtain a temperature-dependent description of r. Our results show that the generation interval is highly sensitive to temperature, decreasing twofold between 25 and 35°C and suggesting that dengue virus epidemics may accelerate as temperatures increase, not only because of more infections per generation but also because of faster generations. Under the empirical temperature relationships that we considered, we found that r peaked at a temperature threshold that was robust to uncertainty in model parameters that do not depend on temperature. Although the precise value of this temperature threshold could be refined following future studies of empirical temperature relationships, the framework we present for identifying such temperature thresholds offers a new way to classify regions in which dengue virus epidemic intensity could either increase or decrease under future climate change., Author summary Recurrent, rapidly intensifying epidemics of dengue–the world’s most prevalent mosquito-borne viral disease–pose a challenge to healthcare systems throughout the tropical and subtropical world. An acute disease that tends to respond well to proper treatment, the sometimes intense nature of dengue epidemics has been known to overwhelm healthcare systems and elevate the morbidity and mortality of patients left without adequate medical treatment under peak epidemic conditions. Here, we quantify the temperature dependence of dengue epidemic intensity by quantifying two distinct determinants of epidemic growth rate: the average number of secondary infections arising from each primary infection and the average time between successive infections in humans. Our results show that the time between successive infections in humans decreases steadily with increasing temperatures, whereas the average number of secondary infections peaks at intermediate temperatures. Altogether, this suggests a peak temperature for dengue epidemic intensity. Applying this result to global temperature projections under future climate change scenarios suggests that dengue epidemics in many regions of the world could become more intense under future temperature increases.

Altitude Sickness Whether the incidence of malaria will be (or has been) affected by the warming climate is poorly resolved. Global-scale analyses are fraught with technical difficulties, including problems with separating out changes resulting from destruction of mosquito habitat, insecticide use, and antimalarial drug use from multidecadal trends in climate change. Siraj et al. (p. 1154 ) performed parallel analyses of highland malaria in Ethiopia and Colombia to ask whether interannual changes in temperature could explain variation in malaria incidence with altitude. Modeling the data confirmed that malaria moves up in elevation in warmer years and allowed estimates of how many more millions of cases could be expected in tropical highland areas if the mean temperature increased by 1° to 3°C.

Objective Tropical highland malaria intensifies and shifts to higher altitudes during exceptionally warm years. Above-normal temperatures associated with El Nino during boreal winter months (December- March) may intensify malaria in East African highlands. We assessed the malaria risk for Oromia, the largest region of Ethiopia with around 30 million inhabitants. Methods Simple linear regression and spatial analyses were used to associate sea surface temperatures (SST) in the Pacific and surface temperatures in Ethiopia with annual malaria risk in Oromia, based on confirmed cases of malaria between 1982 and 2005. Results A strong association (R2 = 0.6, p

Zika virus is a mosquito-borne pathogen that is rapidly spreading across the Americas. Due to associations between Zika virus infection and a range of fetal maladies1,2, the epidemic trajectory of this viral infection poses a significant concern for the nearly 15 million children born in the Americas each year. Ascertaining the portion of this population that is truly at risk is an important priority. One recent estimate3suggested that 5.42 million childbearing women live in areas of the Americas that are suitable for Zika occurrence. To improve on that estimate, which did not take into account the protective effects of herd immunity, we developed a new approach that combines classic results from epidemiological theory with seroprevalence data and highly spatially resolved data about drivers of transmission to make location-specific projections of epidemic attack rates. Our results suggest that 1.65 (1.45–2.06) million childbearing women and 93.4 (81.6–117.1) million people in total could become infected before the first wave of the epidemic concludes. Based on current estimates of rates of adverse fetal outcomes among infected women2,4,5, these results suggest that tens of thousands of pregnancies could be negatively impacted by the first wave of the epidemic. These projections constitute a revised upper limit of populations at risk in the current Zika epidemic, and our approach offers a new way to make rapid assessments of the threat posed by emerging infectious diseases more generally.

A better understanding of malaria persistence in highly seasonal environments such as highlands and desert fringes requires identifying the factors behind the spatial reservoir of the pathogen in the low season. In these ‘unstable’ malaria regions, such reservoirs play a critical role by allowing persistence during the low transmission season and therefore, between seasonal outbreaks. In the highlands of East Africa, the most populated epidemic regions in Africa, temperature is expected to be intimately connected to where in space the disease is able to persist because of pronounced altitudinal gradients. Here, we explore other environmental and demographic factors that may contribute to malaria's highland reservoir. We use an extensive spatio-temporal dataset of confirmed monthly Plasmodium falciparum cases from 1995 to 2005 that finely resolves space in an Ethiopian highland. With a Bayesian approach for parameter estimation and a generalized linear mixed model that includes a spatially structured random effect, we demonstrate that population density is important to disease persistence during the low transmission season. This population effect is not accounted for in typical models for the transmission dynamics of the disease, but is consistent in part with a more complex functional form of the force of infection proposed by theory for vector-borne infections, only during the low season as we discuss. As malaria risk usually decreases in more urban environments with increased human densities, the opposite counterintuitive finding identifies novel control targets during the low transmission season in African highlands.