An open challenge to advance probabilistic forecasting for dengue epidemics.

A wide range of research has promised new tools for forecasting infectious disease dynamics, but little of that research is currently being applied in practice, because tools do not address key public health needs, do not produce probabilistic forecasts, have not been evaluated on external data, or do not provide sufficient forecast skill to be useful. We developed an open collaborative forecasting challenge to assess probabilistic forecasts for seasonal epidemics of dengue, a major global public health problem. Sixteen teams used a variety of methods and data to generate forecasts for 3 epidemiological targets (peak incidence, the week of the peak, and total incidence) over 8 dengue seasons in Iquitos, Peru and San Juan, Puerto Rico. Forecast skill was highly variable across teams and targets. While numerous forecasts showed high skill for midseason situational awareness, early season skill was low, and skill was generally lowest for high incidence seasons, those for which forecasts would be most valuable. A comparison of modeling approaches revealed that average forecast skill was lower for models including biologically meaningful data and mechanisms and that both multimodel and multiteam ensemble forecasts consistently outperformed individual model forecasts. Leveraging these insights, data, and the forecasting framework will be critical to improve forecast skill and the application of forecasts in real time for epidemic preparedness and response. Moreover, key components of this project-integration with public health needs, a common forecasting framework, shared and standardized data, and open participation-can help advance infectious disease forecasting beyond dengue.

Seasonality and the effects of weather on Campylobacter infections.

BACKGROUND: Campylobacteriosis is a major public health concern. The weather factors that influence spatial and seasonal distributions are not fully understood. METHODS: To investigate the impacts of temperature and rainfall on Campylobacter infections in England and Wales, cases of Campylobacter were linked to local temperature and rainfall at laboratory postcodes in the 30 days before the specimen date. Methods for investigation included a comparative conditional incidence, wavelet, clustering, and time series analyses. RESULTS: The increase of Campylobacter infections in the late spring was significantly linked to temperature two weeks before, with an increase in conditional incidence of 0.175 cases per 100,000 per week for weeks 17 to 24; the relationship to temperature was not linear. Generalized structural time series model revealed that changes in temperature accounted for 33.3% of the expected cases of Campylobacteriosis, with an indication of the direction and relevant temperature range. Wavelet analysis showed a strong annual cycle with additional harmonics at four and six months. Cluster analysis showed three clusters of seasonality with geographic similarities representing metropolitan, rural, and other areas. CONCLUSIONS: The association of Campylobacteriosis with temperature is likely to be indirect. High-resolution spatial temporal linkage of weather parameters and cases is important in improving weather associations with infectious diseases. The primary driver of Campylobacter incidence remains to be determined; other avenues, such as insect contamination of chicken flocks through poor biosecurity should be explored.

Elevated risk of stillbirth in males: systematic review and meta-analysis of more than 30 million births.

BACKGROUND: Stillbirth rates have changed little over the last decade, and a high proportion of cases are unexplained. This meta-analysis examined whether there are inequalities in stillbirth risks according to sex. METHODS: A systematic review of the literature was conducted, and data were obtained on more than 30 million birth outcomes reported in observational studies. The pooled relative risk of stillbirth was estimated using random-effects models. RESULTS: The crude mean rate (stillbirths/1,000 total births) was 6.23 for males and 5.74 for females. The pooled relative risk was 1.10 (95% confidence interval (CI): 1.07-1.13). The attributable fraction in the whole population was 4.2% (95% CI: 3.70-4.63), and the attributable fraction among male fetuses was 7.8% (95% CI: 7.0-8.66). Study populations from countries with known sex-biased sex selection issues had anomalous stillbirth sex ratios and higher overall stillbirth risks than other countries, reflecting increased mortality among females. CONCLUSIONS: Risk of stillbirth in males is elevated by about 10%. The population-attributable risk is comparable to smoking and equates to approximately 100,000 stillbirths per year globally. The pattern is consistent across countries of varying incomes. Given current difficulties in reducing stillbirth rates, work to understand the causes of excess male risk is warranted. We recommend that stillbirths are routinely recorded by sex. This will also assist in exposing prenatal sex selection as elevated or equal risks of stillbirth in females would be readily apparent and could therefore be used to trigger investigation.

The development of an early warning system for climate-sensitive disease risk with a focus on dengue epidemics in Southeast Brazil.

Previous studies demonstrate statistically significant associations between disease and climate variations, highlighting the potential for developing climate-based epidemic early warning systems. However, limitations include failure to allow for non-climatic confounding factors, limited geographical/temporal resolution, or lack of evaluation of predictive validity. Here, we consider such issues for dengue in Southeast Brazil using a spatio-temporal generalised linear mixed model with parameters estimated in a Bayesian framework, allowing posterior predictive distributions to be derived in time and space. This paper builds upon a preliminary study by Lowe et al. but uses extended, more recent data and a refined model formulation, which, amongst other adjustments, incorporates past dengue risk to improve model predictions. For the first time, a thorough evaluation and validation of model performance is conducted using out-of-sample predictions and demonstrates considerable improvement over a model that mirrors current surveillance practice. Using the model, we can issue probabilistic dengue early warnings for pre-defined 'alert' thresholds. With the use of the criterion 'greater than a 50% chance of exceeding 300 cases per 100,000 inhabitants', there would have been successful epidemic alerts issued for 81% of the 54 regions that experienced epidemic dengue incidence rates in February-April 2008, with a corresponding false alarm rate of 25%. We propose a novel visualisation technique to map ternary probabilistic forecasts of dengue risk. This technique allows decision makers to identify areas where the model predicts with certainty a particular dengue risk category, to effectively target limited resources to those districts most at risk for a given season.

A spatial approach for the epidemiology of antibiotic use and resistance in community-based studies: the emergence of urban clusters of Escherichia coli quinolone resistance in Sao Paulo, Brasil.

BACKGROUND: Population antimicrobial use may influence resistance emergence. Resistance is an ecological phenomenon due to potential transmissibility. We investigated spatial and temporal patterns of ciprofloxacin (CIP) population consumption related to E. coli resistance emergence and dissemination in a major Brazilian city. A total of 4,372 urinary tract infection E. coli cases, with 723 CIP resistant, were identified in 2002 from two outpatient centres. Cases were address geocoded in a digital map. Raw CIP consumption data was transformed into usage density in DDDs by CIP selling points influence zones determination. A stochastic model coupled with a Geographical Information System was applied for relating resistance and usage density and for detecting city areas of high/low resistance risk. RESULTS: E. coli CIP resistant cluster emergence was detected and significantly related to usage density at a level of 5 to 9 CIP DDDs. There were clustered hot-spots and a significant global spatial variation in the residual resistance risk after allowing for usage density. CONCLUSIONS: There were clustered hot-spots and a significant global spatial variation in the residual resistance risk after allowing for usage density. The usage density of 5-9 CIP DDDs per 1,000 inhabitants within the same influence zone was the resistance triggering level. This level led to E. coli resistance clustering, proving that individual resistance emergence and dissemination was affected by antimicrobial population consumption.

Confident interpretation of Bayesian decision tree ensembles for clinical applications.

Bayesian averaging (BA) over ensembles of decision models allows evaluation of the uncertainty of decisions that is of crucial importance for safety-critical applications such as medical diagnostics. The interpretability of the ensemble can also give useful information for experts responsible for making reliable decisions. For this reason, decision trees (DTs) are attractive decision models for experts. However, BA over such models makes an ensemble of DTs uninterpretable. In this paper, we present a new approach to probabilistic interpretation of Bayesian DT ensembles. This approach is based on the quantitative evaluation of uncertainty of the DTs, and allows experts to find a DT that provides a high predictive accuracy and confident outcomes. To make the BA over DTs feasible in our experiments, we use a Markov Chain Monte Carlo technique with a reversible jump extension. The results obtained from clinical data show that in terms of predictive accuracy, the proposed method outperforms the maximum a posteriori (MAP) method that has been suggested for interpretation of DT ensembles.

Random-effects models in investigating the effect of vitamin A in childhood diarrhea.

PURPOSE: By adopting more appropriate and powerful statistical methods that fully exploit longitudinal structure, we re-analyze and extend previously published results from a large community trial to investigate the effect of vitamin A supplementation on the prevalence and severity of diarrhea in young children. METHODS: Generalized linear mixed models were used to allow for repeated measures in a reanalysis of a double-blind, randomized, placebo-controlled community trial conducted in a cohort of children in northeastern Brazil during 1 year. The response variable was weekly number of days with diarrhea for each child, and Markov Chain Monte Carlo methods were used to estimate model parameters. RESULTS AND CONCLUSIONS: Random effects suitably accounted for the underlying heterogeneity between and within children, and our longitudinal analysis shows a significant beneficial effect of vitamin A supplementation that was inconclusive in previously reported simple summary analyses of these data. Risk for diarrhea infection was estimated to be 1.57 times greater for a child administered a placebo as opposed to vitamin A (95% credible interval, 1.17-2.12). Additionally, we identified previously unreported temporal effects in these data, showing a decreasing daily probability of diarrhea for both groups during the trial and treatment-time interaction.

Evaluating probabilistic dengue risk forecasts from a prototype early warning system for Brazil.

Recently, a prototype dengue early warning system was developed to produce probabilistic forecasts of dengue risk three months ahead of the 2014 World Cup in Brazil. Here, we evaluate the categorical dengue forecasts across all microregions in Brazil, using dengue cases reported in June 2014 to validate the model. We also compare the forecast model framework to a null model, based on seasonal averages of previously observed dengue incidence. When considering the ability of the two models to predict high dengue risk across Brazil, the forecast model produced more hits and fewer missed events than the null model, with a hit rate of 57% for the forecast model compared to 33% for the null model. This early warning model framework may be useful to public health services, not only ahead of mass gatherings, but also before the peak dengue season each year, to control potentially explosive dengue epidemics.

Biclustering models for structured microarray data.

Microarrays have become a standard tool for investigating gene function and more complex microarray experiments are increasingly being conducted. For example, an experiment may involve samples from several groups or may investigate changes in gene expression over time for several subjects, leading to large three-way data sets. In response to this increase in data complexity, we propose some extensions to the plaid model, a biclustering method developed for the analysis of gene expression data. This model-based method lends itself to the incorporation of any additional structure such as external grouping or repeated measures. We describe how the extended models may be fitted and illustrate their use on real data.

Dengue outlook for the World Cup in Brazil: an early warning model framework driven by real-time seasonal climate forecasts.

BACKGROUND: With more than a million spectators expected to travel among 12 different cities in Brazil during the football World Cup, June 12-July 13, 2014, the risk of the mosquito-transmitted disease dengue fever is a concern. We addressed the potential for a dengue epidemic during the tournament, using a probabilistic forecast of dengue risk for the 553 microregions of Brazil, with risk level warnings for the 12 cities where matches will be played. METHODS: We obtained real-time seasonal climate forecasts from several international sources (European Centre for Medium-Range Weather Forecasts [ECMWF], Met Office, Meteo-France and Centro de Previsão de Tempo e Estudos Climáticos [CPTEC]) and the observed dengue epidemiological situation in Brazil at the forecast issue date as provided by the Ministry of Health. Using this information we devised a spatiotemporal hierarchical Bayesian modelling framework that enabled dengue warnings to be made 3 months ahead. By assessing the past performance of the forecasting system using observed dengue incidence rates for June, 2000-2013, we identified optimum trigger alert thresholds for scenarios of medium-risk and high-risk of dengue. FINDINGS: Our forecasts for June, 2014, showed that dengue risk was likely to be low in the host cities Brasília, Cuiabá, Curitiba, Porto Alegre, and São Paulo. The risk was medium in Rio de Janeiro, Belo Horizonte, Salvador, and Manaus. High-risk alerts were triggered for the northeastern cities of Recife (p(high)=19%), Fortaleza (p(high)=46%), and Natal (p(high)=48%). For these high-risk areas, particularly Natal, the forecasting system did well for previous years (in June, 2000-13). INTERPRETATION: This timely dengue early warning permits the Ministry of Health and local authorities to implement appropriate, city-specific mitigation and control actions ahead of the World Cup. FUNDING: European Commission's Seventh Framework Research Programme projects DENFREE, EUPORIAS, and SPECS; Conselho Nacional de Desenvolvimento Científico e Tecnológico and Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro.

Semiparametric modeling of the spatial distribution of occupational accident risk in the casual labor market, Piracicaba, southeast Brazil.

The objective of this study was to estimate the spatial distribution of work accident risk in the informal work market in the urban zone of an industrialized city in southeast Brazil and to examine concomitant effects of age, gender, and type of occupation after controlling for spatial risk variation. The basic methodology adopted was that of a population-based case-control study with particular interest focused on the spatial location of work. Cases were all casual workers in the city suffering work accidents during a one-year period; controls were selected from the source population of casual laborers by systematic random sampling of urban homes. The spatial distribution of work accidents was estimated via a semiparametric generalized additive model with a nonparametric bidimensional spline of the geographical coordinates of cases and controls as the nonlinear spatial component, and including age, gender, and occupation as linear predictive variables in the parametric component. We analyzed 1,918 cases and 2,245 controls between 1/11/2003 and 31/10/2004 in Piracicaba, Brazil. Areas of significantly high and low accident risk were identified in relation to mean risk in the study region (p < 0.01). Work accident risk for informal workers varied significantly in the study area. Significant age, gender, and occupational group effects on accident risk were identified after correcting for this spatial variation. A good understanding of high-risk groups and high-risk regions underpins the formulation of hypotheses concerning accident causality and the development of effective public accident prevention policies.

Use of spatial factors in the analysis of heavy metals in sediments in a Brazilian coastal region

Environmental phenomena often generate complex data which are not only multivariate , but where spatial structure and dependency may also be important. One potential area of interest is that of identifying combinations of variables which exhibit distinctive spatial structure in a set of multivariate and geographically indexed observations, e.g. long, as opposed to short, range spatial contiguity. Such components may serve not only to characterize the main determinants in the geographical distribution of a set of environmental measurements , but may also provide useful insight into the physical or environmental processes underlying and goverming that distribution. This article suggests one form of multivariate technique which may be useful in such a context. We demonstrate the method in relation to data comprising measures of various heavy metals in sediments in a Brazilian coastal region.

Decreasing in-hospital mortality after hip fracture: invest on expertise or primary attention?

The number of frailty hip fractures is increasing leading to necessary adjustments of healthcare systems. The majority of these fractures require specialized care (surgery), recovery is usually long and some patients die during hospitalization or a short period after. It is important to understand whether outcomes are related to provider profile´s in order to plan future health investments. The aim was to assess determinants of death during hospitalization due to hip fracture in order to understand which characteristics: individual, provider or geographical area, should be targeted by future health investments. National secondary data from Portugal 2000- 2008 was used. All patients over 49 years diagnosed with hip fracture (coded ICD9-CM 820.x) caused by low/moderate trauma were evaluated. Patients´variables (National Hospital Discharge Register) portrayed individual characteristics (age, sex and comorbidities inputted in the Charlson Comorbidity Index). Hospital variables (National Hospital Inquiry) were associated with structure/expertise (volume, ratio of resources to total). Area variables (Statistics Portugal) portrayed differences in health equipment and/or performance. A multi-level Bayesian modelling framework (patient/hospital/area) was adopted with in-hospital fatality as the outcome. A total of 53,684 hospitalizations due to hip fractures were included. There were initial differences between providers regarding in hospital fatality. However, after the model adjustment all providers had the same effect. Variables that were significantly associated with the probability of death, during an hospitalization to treat a hip fracture, were manly individual variables: age [log-odds scale with 95% credible intervals (CrI), 0.07 per year (0.06, 0.07)], sex [smaller effect for females -0.72 (-0.80, -0.63)], length of stay [-0.01 per day (-0.02, -0.01)], and severity (Charlson's Index) [highest versus lowest category, 2.50 (2.30, 2.71)]. Providers had different time-effects. Results indicate that if hospitals were treating equal patients they would have equal performance. Therefore, given the current expertise, by improving the population's health the in-hospital probability of death related to the treatment of hip fracture, can be reduced. Health attention/investments should concentrate on maintaining the population's health rather than invest on more sophisticated methods/equipment for this particular situation.

Spatial statisticl methods in health

O estudo da distribuiçäo geográfica da incidência de doenças e da sua relaçäo com fatores de risco potenciais (chamada aqui de "epidemiologia geográfica") vem constituindo um terreno fértil para a aplicaçäo e desenvolvimento de métodos e modelos estatísticos. Nos últimos anos, foram desenvolvidas ferramentas cada vez mais poderosas e versáteis nesta área de aplicaçäo. Discute as classes gerais de problemas na epidemiologia geográfica e faz uma revisäo dos principais métodos estatísticos utilizados atualmente em cada uma das áreas de aplicaçäo identificadas. Näo procura cobrir exaustivamente todos os medos e modelos possíveis, mas fornece referências bibliográficas para outros detalhes e abordagens. O objetivo geral é dar um panorama do "estado da arte" no uso de métodos estatísticos espaciais na pesquisa em epidemiologia e saúde pública. Após a revisäo metodológica, discute-se os principais ambientes de software disponíveis para implementar tais métodos. Conclui com algumas reflexöes gerais sobre as implicaçöes, para a epidemiologia e a saúde pública, do uso de métodos estatísticos espaciais em saúde, além dos benefícios e problemas associados.