Contents © 2013 Flávio Codeço Coelho - Powered by Nikola
Preliminary list of Accepted Posters
This list is not final, if your work has been accepted but is not on the list below, it is probably going to be added soon.
Real world data: a tool for decision making in health
Maria Fernanda Mussolino Ribeiro, NewBD
Objective: to stimulate the use of secondary data in Health Economics analyses, based on a better understanding of available possibilities in Public Health database. Methods: analysis and observation of available data to develop Health Economics studies using Public Health database; based on that, exposes the data sources and this data capture process; qualifies Datasus database and explains the methodologies, finally, describes the steps to develop the analyses. Results: the most important public data in health care are Unified Health System (SUS); Outpatient Information (SIA/ SUS) - 3 million patients and expenditures of 8.7 billion reais in 2011; Inpatient Information (SIH/SUS) - more than 9 million patients and expenditures of 13.10 billion reais in 2011; the data national publication is updated monthly, in approximately 30 days from the production ending; the growing awareness about the information transparency and the use of databases as a tool for SUS planning and management, makes the quality of Public Health databases, significantly improve in recent years; these studies are possible, due the identification of the individual under treatment. Conclusions: even though SIA/SUS and SIH/SUS, are secondary data typically created by process that do not consider the scientific accuracy strict and proper of academic analyses, we should recognize that this possibility described in the article, introduces a new source of data to be considered.
Gene expression networks of advanced ovarian carcinoma associated to tumor grade, histology and stage.
Ulises Urzua (1 * §), James M Cherry (2 §), Ester Rozenblum (3), Sandra Ampuero (4), Luis Valenzuela (1), John I Powell (5) and David J Munroe (6)
* Corresponding author; § contributed equally
Background: Ovarian cancer has remained as the deadliest gynecological cancer in western countries for the last decades. Unspecific symptoms, lack of effective early screening markers, histological complexity and a controversial disease etiology have severely contributed to late-stage detection with poor survival. Objective: This study aimed to extract gene expression networks relating grade, stage and histology of advanced ovarian carcinomas (OC) and their possible biological and clinical significance. Methods: Gene expression profiles of 38 stage-III OC cases were obtained with DNA-microarrays and analyzed statistically with limma tests according primarily to conventional FIGO grades, i.e. poorly, moderately and well differentiated tumors. To make sense of profiling data, further clinicopathological information including histology (serous, papillary-serous, mucinous) and stage (IIIA, B, C) were added as statistical covariates. Network analysis was done with STRING 9.0 and GO/KEGG/pathway/disease analyses were performed with WebGestalt 2.0, DAVID 6.7 and HuGE Navigator 2.0. Results: Among differentially expressed genes across tumor grades, 819 were associated to histology and 628 to stage. The overlap between covariates reduced the list to 376 genes that comprised multiple functions including growth and proliferation, chromatin dynamics and microtubule-based processes. Enriched signaling pathways were Mapk, calcium, aurora kinases, Wnt and GnRH while aminoacids, glycerophospholipids and purines were the relevant metabolic routes altered. Nearly twenty genes corresponded to oncogenes and tumor suppressor genes and 10 genes (CCNB1, E2F5, ERBB4, HRAS, KCND3, MKI67, NFKB1, BIN1, FGFR2 and HBXIP) were associated to OC outcome in other studies. Conclusion: Grade-based gene expression profiling of OC considering histology and stage is able to identify inherent biological features of advanced tumors. While a great extent of signaling pathways are typical of tumor cells from various origins, distinctive metabolic and specific signaling pathways as well as protein families might be marker candidates of OC outcome and therapy.
Supported by NCI Contract No HSN261200800001E and Fondecyt 1130292.
Information management in public health: knowledge discovery in database and technologies
Carlos Roberto Valêncio[1], Paula Rahal[1], Pedro Luiz Pizzigatti Corrêa[2], Moacir Alves Borges[3], Gracio Tomaz Saturno[4], Ana Paula Tencarte[5], Marilda C. A. A. Rodrigues[6]
6. Prefeitura Municipal de São José do Rio Preto Contact: valencio@ibilce.unesp.br
A dataset which involves about 1.5 million of inhabitants in Brazil has been storaged and it is related to the following public health issues: hepatitis, epilepsy, mental illness, drug addiction, tumour tissues and work accidents. It is organized under a data archtecture supported by technologies in order to promote data integration, data warehousing, data cleaning, data mining, visual data mining, spatial data mining, data cloud and geographic data managment. In this work, the goal is the presentation of the experience in information management, considering the technologies to support the steps which compose the Knowledge Discovery in Database (KDD) process.
Supported by Ministry of health and FAPESP
Intelligent System for Monitoring Patients applied to database Multi-parameter Intelligent Monitoring for Intensive Care
Cicília R. M. Leite - State University of Rio Grande do Norte - Brazil
Ana M. G. Guerreiro - Federal University of Rio Grande do Norte - Brazil
The area of the hospital automation has been the subject a lot of research, addressing relevant issues which can be automated, such as: management and control (electronic medical records, scheduling appointments, hospitalization, among others); communication (tracking patients, staff and materials), development of medical, hospital and laboratory equipment; monitoring (patients, staff and materials); and aid to medical diagnosis (according to each speciality). This work presents an architecture for a patient monitoring and alert systems. This architecture is based on intelligent systems techniques and is applied in hospital automation, specifically in the Intensive Care Unit (ICU) for the patient monitoring in hospital environment. The main goal of this architecture is to transform the multiparameter monitor data into useful information, through the knowledge of specialists and normal parameters of vital signs based on fuzzy logic that allows to extract information about the clinical condition of ICU patients and give a pre-diagnosis. Finally, alerts are dispatched to medical professionals in case any abnormality is found during monitoring. After the validation of the architecture, the fuzzy logic inferences were applied to the trainning and validation of an Artificial Neural Network for classification of the cases that were validated a priori with the fuzzy system.
Desenvolvimento de um Modelo para Estimativa de Internação Hospitalar Baseado na Rede Neural Artificial Perceptron de Múltiplas Camadas
Fabrício P. Härter, Anderson S. Nedel, Stephano R. Freitas, Gisele Vega
Universidade Federal de Pelotas
Nesta pesquisa tenta-se desenvolver um modelo estatístico, que permita ao usuário prever se determinadas variações no tempo implicam em internações hospitalares. Uma vez que este objetivo seja atingido, procura-se identificar picos de internações, bem como quantificar estas internações. Para tal, utiliza-se a Rede Neural Perceptron de Multiplas Camadas. Entre as principais propriedades e capacidades desta técnica, estão: não-linearidade, mapeamento de entrada-saída, adaptabilidade, aprender com exemplos, processamento paralelo e distribuído, capacidade de generalização e tolerância à falhas. Resultados parcias mostram que a rede é capaz de mapear a relação entre as variáveis meteorológicas e as internações hospiltalares, mas ainda não foi possível estimar novas internações com base em conjunto de dados do passado. Entretanto, diferentes parâmetros, arquiteturas e paradigmas de redes devem ser exploradas.
Titulo
Murilo Guedes, Faculdade de Tecnologia de Cruzeiro
Databases are a very essential tool, but current needs require a new approach to data modeling. This need was met by the creation of a new type of database; called NoSQL (Not Only SQL) that does not use language-based SQL language based Relational DBMSs. The work has as main focus, modeling NoSQL, specifically oriented model columns in this DBMS NoSQL Cassandra, from the perspective of creating a prototype for managing a commercial property, business, small and medium -sized developed in language of object-oriented programming and C # components. The aim is to contrast the relational model with the models of the NoSQL paradigm, analyzing the impacts of such differences during the implementation, especially the relational model with the paradigm CQL (Cassandra Query Language) database present in Cassandra. This is opposed approaches during the development of a prototype system for managing small and medium business, contributing also to the growth of NoSQL model. The NoSQL Database , in general , provide greater scalability , flexibility and performance , however , these qualities eventually become susceptible to data inconsistencies , fleeing the standard ACID ( Atomicity , Consistency , Isolation , Durability ) and adopting the principle BASE ( Basically Available , Soft state , Eventual consistency ) . The lack of consistency guarantee makes NoSQL bit suitable for some applications, such as financial institutions, but highly recommended to use applications in real-time, high volume data analysis and requiring high performance in reading and writing data. When considering the Apache Cassandra created by Facebook, is likely to say that it has a construction based on the orientation columns, being highly distributed, scalable, ensuring wide availability of information. Although it is a NoSQL Cassandra , or adopts a basic principle , however , making it in some way different from other Cassandra NoSQL DBMSs is the availability of controlling the levels of consistency but this mechanism , depending on the level consistency , compromises the overall performance of the database . Cassandra has architecture, at least in the category of tables in which case the Cassandra column families are called, not so far from the relational model, and may also be considered an evolution of the relational paradigm. The prototype created from the default NoSQL Cassandra, based the commands CQL, along with the programming language C #, presented during all tests performed, great performance, for systems developed with DBMS Relational in questions read, write, delete and update large amount of data made by algorithms developed for the test, and in relation to the issue of consistency, there has been no inconsistency problem pointed some information within the system prototype.
The SALVAR web-tool and big data applied to Public Health Monitoring
Luciana de Resende Londe,Leonardo Bacelar Lima Santos, Ana Elisa Silva, Aurelienne Souza, Jether Rodrigues, Mariane Assis Cemaden - Centro Nacional de Monitoramento e Alertas de Desastres Naturais
According to Hey et al. (2009), the current paradigm for the science is the data analysis and inference rules: relating experiments, modeling and simulation to support the knowledge generation. There is a demand for methods to manage large and increasing datasets, with several data formats provided by different sources. In this context, the Big Data approach appears as a powerful tool. Big data usually includes data sets which sizes exceed the ability of commonly used software tools to capture, manage, and process the data within a manageable elapsed time (Snijders et al., 2012). In countries with large surface areas, such as Brazil, there are many diseases breakouts, varying according to the incidence region. It is known that some diseases have a strong relationship with environment and climate characteristics and they are at some extension linked to the occurrence of natural disasters. The challenge of connecting these different issues performs a suitable occasion both to explore available big data and current tools and to address this information towards public health monitoring.
The platform SALVAR (Sistema de Alerta e Visualização de Áreas de Risco – Alert and Visualization System of Areas Under Risk) is a web-based computational system developed to monitor natural disasters (Soler et al., 2013). The SALVAR provides to experts a Web Geographic Information System, using as map server the GeoServer – a reference implementation of OGC standards for Web Feature Service (WFS) and Web Map Service (WMS). Big data of meteorological, hydrological and geotechnical measurements represented as spatial variables at SALVAR may improve early warnings for disasters and moreover support the monitoring of some public health problems. To develop the first steps of this work, data about post-disaster diseases in Brazil were obtained from DATASUS (Departamento de Informática do Sistema Único de Saúde) and local data were obtained from the “Secretaria de Estado da Saúde de Santa Catarina”. We already tested data for municipalities from Santa Catarina state and in future researches we intend to study the relationship between disasters and diseases for other Brazilian municipalities. Using SALVAR and big data approaches, we sent hydrological warnings about flooding in the Amazon region during July/2013, especially for the municipalities Manaus (AM), Santarém (PA) and Óbidos (PA). In this region, river floods reach large areas during a long period of time. These characteristics, associated to precipitation dynamics, are related to varying disease spreads, with impacts to local communities. Considering these linkages, an extra issue has been added to current hydrological warnings, addressing risk of contamination by some diseases both during and after flooding, highlighting the actual application of this work.
References:
Hey, T.; Tansley, S.; Tolle, K. in. The fourth paradigm: data-intensive scientific discovery. Florianopolis: Microsoft Corporation, 2009. Snijders, C., Matzat, U., & Reips, U.-D. (2012). ‘Big Data’: Big gaps of knowledge in the field of Internet. International Journal of Internet Science, 7, 1-5. Soler et al. (2013). Challenges and perspectives of innovative digital ecosystems designed to monitor and warn natural disasters in Brazil. Accepted to ACM Conference on Management of Emergent Digital EcoSystems (ACM MEDES'13), Neumünster Abbey, Luxembourg.
Predicting obesity using machine learning techniques: Classification trees and random forests.
Hudson Golino, Universidade Federal de Minas Gerais / Faculdade Independente do Nordeste
The present study investigates the prediction of obesity (BMI> 29.9 kg / m²) by waist (WC) and hip circumference (HC), and waist hip ratio (WHR) using a machine learning technique named Classification Tree. Data were collected from 400 college students (56.3% women) from 16 to 63 years old (M = 23:14, SD = 6:03). The sample was divided into two sets of each sex (training and test) for cross-validation. Seven trees were calculated in training group for each sex, using different numbers and combinations of predictors. The result shows that for women WC and WHR is the combination that produces the best prediction, since it has the lowest deviance (16.46) and misclassification (.035), and the higher pseudo R2 (.69). This model had an accuracy of 96.5% in the classification of obese and non-obese people in the training set, and presented a sensitivity of 60% and a specificity of 100% in the test set. For men CQ showed the best prediction with the lowest deviance (9:53) and misclassification (.022), and the higher pseudo R2 (.84). This model had an accuracy of 97.8% in the classification of obese and non-obese people, and presented a sensitivity of 75% and a specificity of 96.20% in the test set. Since classification trees are subject to high variability when applied to new data sets (e.g. the difference between the accuracy in the training set and in the test set) due to overfit. So, sample and variable bootstrap can provide a better approximation of the true accuracy of the prediction, avoiding overfiting. Random forest is a set of computational and statistical procedures that bootstrap both the sample and the variables, growing a large number of prediction trees, in order to discover which variable are important in the prediction, and what is the “true” accuracy (i.e. the accuracy that is expected using the given predictors over all possible sample). Applying the random forest analysis in the men’s training set using WC, HC and WHR as predictors led to an accuracy of 95.45% in the classification, with a specificity of 99% and a sensibility of 66%. The importance analysis, provided by the Random Forest technique, shows how much each variable contributes to the prediction of obesity. Waist circumference and hip circumference are, respectively, 21.19% and 30.96% more helpful than chance to successfully classify men’s obesity, while WHR is only 0.8%. Inclusion of the latter reduces, in average, 4% of the classification error, while the former reduces 5%. The result of the random forest analysis in the women’s training sample showed an accuracy of 95.58%, with a specificity of 99.99% but a sensibility of 43%. In terms of importance, WC is 11% more helpful than chance to successfully classify women’s obesity, while HC is 21% and WHR is only 1% better than random. Finally, we’ll introduce an innovative way to visualize prediction accuracy by applying the Fruchterman Reingold algorithm in the proximity matrix generated by the random forest analysis. This algorithm computes a network layout in which the length of edges (cases) depends on their absolute weight, which in our case is the proximity value for each pair of cases.
Cloud computing technologies for genomic big data analysis
Fabrício Alves Barbosa da Silva, Programa de Computação Científica, Fiocruz.
In this talk we will present several cloud computing technologies that are being used for processing large volumes of genomic data. We also intend to present recent trends in cloud computing technologies and discuss how these new technologies can be used to increase throughput and decrease the processing time of genomic pipelines.
The quality of health information on-line should be assessed?
André Pereira Neto, Oswaldo Cruz Foundation
The Internet is the main vehicle of communication and dissemination of information in the contemporary world. With the Big Data the information available, the possibilities of interaction between individuals and the conditions for the production of information became even greater on various topics. Health emerged as one of the areas where there is increasing information available to a countless stakeholders. The people accessing the Internet to get some information about their health condition increasing everyday. This health information available on the Internet can be produced and maintained by businesses, individuals, professional associations, non-governmental organizations or public agencies without any filter. This aspect creates a problem that can have serious consequences for Public Health: Many virtual environments exhibit incomplete, inconsistent, incorrect or even fraudulent information. The information content can be accurate and up to date but it can be presented in a manner incomprehensible to a layman. On the other rand, information can be easily understood, although they are outdated. This information can hinder treatment, induce improper self-medication and even harm the patient. Given this reality the average person has difficulty distinguishing right from misleading or unheard from traditional information. Faced with the deluge of information, sponsored by the Internet, we ask: Is it necessary to assess the quality of health information on line? There are those who advocate that the collaborative mechanisms will become filters. They will make one information more reliable than another. Others believe that the information on line can not be measured with the same parameters used in tangible objects. They claim that there are no evaluation criteria for intangible goods, as information. For them, the certification of information online becomes a utopia. There are those who consider this assessment an instrument of freedom restriction in the digital world. In our point of view, the market of ideas should not be considered a qualifier for information. The different environments Nazis and xenophobes on-line suggest that the virtual world suffers the same problems experienced in the real world. The lack of evaluation indicators of intangible assets (such as information), did not prevent different public agencies and non-governmental organizations to built criteria, tools and other mechanisms to assess the quality of health information on-line. This was the case, for example, of the Health Information Technology Institute (Hiti), in the United States; the Discern Questionnaire (DQ), in the National Health Services of England and the Health on the Net Foundation (HON) – Swiss non-governmental organization recognized by the Council social and Economic UN. The purpose of this communication is to debate the importance of quality assessment of health information on-line.
BREAST CANCER DATA QUALITY AT THE POPULATION-BASED CANCER REGISTRY OF SÃO PAULO: IMPLICATIONS FOR HEALTH PLANNING
DANIELE PINTO DA SILVEIRA, AGÊNCIA NACIONAL DE SAÚDE SUPLEMENTAR
Background:Information based on population data, such as Cancer Registries, are essential to endorse health planning and provision of cancer care services. This paper aims to present a data quality evaluation and coverage of breast cancer incidence data from the Population Based Cancer Registry of the Municipality of São Paulo, Brazil. The evaluation was based on international quality standards adopted by international agencies worldwide. Methods: The study included 46,305 new breast cancer cases recorded during a two five-year periods (1997-2001 and 2002-2006). Population-based cohort study was led focusing on variables such as: date and age at diagnosis, method of diagnosis, clinical stage and topography. Data collected by other Registries, in other countries, were used as a data proxy for coverage comparisons with São Paulo Cancer Registry. Results: The study has shown that some variables considered in the literature as essential to the validity and comparability of Registries, such as the diagnosis by histopathology exam and age of the patients, has a good information, while than for other critical variables, such as clinical staging, there are problems of data completeness. Diagnostic criteria and classification based on International Classification of Disease (ICD -10) presented 100% of completeness. However, the completeness of death registration was around 20%, less than the expected for a cancer registry. Conclusion: Record linkage methodologies can be applied to improve data quality in population-based registries and comparisons with others population-based information systems can be a good method of accounting for missing information and amplifying the usability of population-based cancer registries.
The influence of chemical control techniques in the emergence of resistant populations of Aedes aegypti
CRYSTTIAN ARANTES PAIXÃO e Flavio Codeço Coelho, FUNDAÇÃO GETULIO VARGAS
Among the numerous public health problems to which the world population is subjected, Dengue fever stands out. Dengue fever is a disease transmitted by mosquitoes in tropical and subtropical regions of the world, i.e., a considerable proportion of the world’s population lives in risk areas. The main vector of dengue is the mosquito hematophagous of the genus Aedes. It belongs to the species aegypti, albopictus and polynesienses. In Brazil, the majority of infestations is caused by Aedes aegypti, which is spread over a large part of the Brazilian territory. Because it is a viral disease that still has no definitive treatment, vector control is one way of trying to combat the disease. Different forms of combat are used in an attempt to eliminate the vector: mechanical chemical and biological control. Through mechanical control, the sites where the vector breeds are eliminated to prevent growth of vector population. Through biological control, pathogens or predators are used to eliminate the vector. Through chemical control, one of the most frequently used methods, the vectors are subjected to insecticides of organic and inorganic origin. It has been used most often since they are easily manipulated to be applied in large scale. Its intensive use has, however, caused some problems. One of the problems which deserves attention and which is the focus of this study relates to the emergence of insecticide-resistant mosquito populations. This resistance decreases the efficiency of insecticides and makes the method impracticable. The indiscriminate use of these products generates a selection pressure on the vector population which is genetically resistant to the active ingredient used. This, in turn, allows for the reproduction of resistant vectors. The objective of this study is to describe the dynamics of resistance in vector populations using individual-based models, which emphasizes the bitstring technique. Within this model, the most relevant characteristics of the individuals are added to bit strips and then manipulated to simulate their behavior. In particular, the resistance is characterized by two alleles in which the resistance is encoded by the dominant one. Individuals are subjected to a form of control using insecticide and populations grow by sexual reproduction. Thus, with such a model, it becomes possible for parents to pass on the resistant characteristics to offspring, and it is possible to observe the dynamics of resistance in the population over time for different control methodology settings. Moreover, it is possible to test different methods of control, highlighting the optimum technique. With this technique, we seek to find the most desirable plan for the application of chemical control in order to minimize the number of mosquitoes and not favor the emergence of a resistant population.
The area of influence of the BR 163 in the state of Mato Grosso - Brazil. Deforestation and its relation to environmental health in the municipality of Peixoto de Azevedo
Lilian Rose Lemos Rock ( 1 ) ( 1 ) Master in Public Policy and Sustainable Development / CSD - Unb . Coordinator of the Postgraduate Diploma in Law Uniceub , Doctoral Program in Health Sciences and Technologies - Unb- lilian.rocha@uniceub.br
This summary is part of the doctoral thesis which aims to assess the increase of deforestation ( burning) and their effects on the health of the population and the number of outpatient visits for respiratory problems among children aged 0-05 years and seniors over sixty and five years of age, living in the area of influence of the BR163 , which unites the states of Mato Grosso and Pará the area chosen was the subarea of the far north of the state of Mato Grosso . The area was subjected to colonization of INCRA settlements by private companies and government , which led to an economic model based on the extraction of natural resources, taking as a result of deforestation via which caused extensive devastation and decay areas , and water pollution and air contamination by overusing pesticides , bringing substantial risk to health and the environment. The socio-ecological degradation is strongly linked to the onset of respiratory disease ( RD) , epidemic , and others arising from excessive use of pesticides . The opening of roads , mining , cattle ranching expansion , the timber, the increase of fires , causes the area of influence of the BR 163 make the region a large field of research. The reason for the choice of Peixoto de Azevedo is due to the fact the city is included in the arc of deforestation due to high deforestation rates and the number of fire outbreaks , and present indicators of morbidity and mortality for DR among children aged 0 to 05 years the period 2009-2010 . During yet rated , considering the total deforestation and increase in the same period . According to data from Prodes the total deforestation in 2009 km ² was 9762.4 ( 67.79 % ) , the increase in the period 2008/2009 km ² was 72.1 km ² (0.50 % ) , compared with the same period the total deforested by 2008 which was 9762.4 ( 67.79 % ) and the increase in 2007/ 2008 was 25.1 km ² 0:17 % increase in fires caused by deforestation . The outpatients were collected on the platform of the Ministry of Health , contract - DATASUS health information through the records annual distribution rate of hospitalizations for different causes and age - CID during 2009 , representing 4.3 % ace DR . of total admissions in the municipality . Analysis and construction of indicators will be developed between the proportion of the number of outpatients in the periods described above and its relation with the increase of deforestation ( burned ) in the city now under development in doctoral research .
Large-Scale Data-driven Modeling for Detection and Prediction of Threats to Human and Wildlife Health.
Marcia Chame[1], Livia Abdalla[1], Eduardo Krempser[2], Douglas Augusto[2]
[1] Fundação Oswaldo Cruz - FIOCRUZ
[2] Laboratório Nacional de Computação Científica - LNCC
The abundance and increasing data availability related to biodiversity and public health represent nowadays a great research opportunity that connects environmental imbalance with the transmission of diseases between animal and human species. At the same time, treatment, management and integration of these large amounts of data, especially in highly diverse countries as Brazil, represent a complex challenge. In this sense, one can find a proposal for the organization and development of the Information System on Wildlife Health ("Sistema de Informação em Saúde Silvestre") - SISS, which is the central component of the Information Center on Wildlife Health ("Centro de Informação em Saúde Silvestre") - CISS, which in turn is the objective of Fiocruz in National Project of Integrated Public-Private Actions for Biodiversity ("Projeto Nacional de Ações Integradas Público-Privadas para a Biodiversidade") - PROBIO II. SISS will integrate from basic data of wildlife occurrences, with the help of several society segments, up to environment and social data, through direct integration of georeferenced data bases of many partner governmental institutions. The storage of such massive amount of data will take place on georeferenced data banks, where it will systematize and standardize cartography features of the data, allowing the construction of spatial models of alert detection and prediction. The modeling to be developed will be initially based on parameters previously defined in the system such as territorial distance of the occurrences, frequency of the records, taxonomy similarity of the involved groups, and physical conditions of the observed animals. These parameters are associated with environment and social informations such as use and coverage of the land, topography, types of vegetation, climate, human impacts, and notifications of threats to health systems. The alert detection model is divided into three phases: the clustering of occurrences into groups (through unsupervised learning), the extraction of certain features of the resulting groups (such as frequency of the events), and finally the classification of them according to whether they represent an alert or not (by means of supervised learning). The alert model will evolve as new data arrive as a result of the confirmation of the alerts emitted by the system, which will be done with the special aid of the Wildlife Health Diagnostic Network. Through those confirmations the model will be able to tune its efficiency and it is expected that it will play an important role with respect to the prediction of threats to human and wildlife health. All in all, it is hoped that the discussed modeling, which will handle massive amount of data, be able to identify alert and emergency situations of threats to fauna and also help to advance the understanding of the complex relations that connect human and wildlife health together.
Infraestrutura computacional para suporte e avaliação dos efeitos dos Programas Sociais com base em coorte populacional referenciada no Cadastro Único.
Maurício L. Barreto, Raimundo José A. Macêdo, Marcos E. Barreto, Alírio S. Sá, Margarete Sá.
Universidade Federal da Bahia
Os programas sociais têm contribuído na redução da pobreza e das desigualdades socioeconômicas. Existem evidências de que tais programas geram efeitos positivos nas áreas de Saúde e de Educação, embora alguns efeitos negativos também sejam relatados.
As agências governamentais responsáveis pelos programas, juntamente com centros de pesquisa e agências internacionais de cooperação têm envidado esforços para conhecer os variados efeitos destes programas sobre as populações beneficiadas e o seu entorno. Tais estudos fornecem informações tanto para o processo de tomada das decisões necessárias ao aprimoramento das políticas e dos programas públicos quanto para os beneficiários e interessados nestas intervenções, qualificando a transparência dos investimentos.
Os estudos observacionais quase experimentais são frequentemente empregados para tal finalidade, dada sua eficiência e maior abrangência quando comparados aos estudos randomizados. Dentre os métodos disponíveis, aqueles baseados em coortes populacionais apresentam resultados mais robustos, apesar de questões metodológicas, operacionais e de custos.
O presente trabalho visa desenvolver uma prática inovadora de avaliação e monitoramento do Programa Bolsa Família e de outros programas sociais que utilizam o Cadastro Único como referência. A metodologia proposta permitirá o amplo estudo das determinantes sociais dos desfechos estudados. No campo da Saúde, por exemplo, permitirá a exploração de hipóteses relacionadas com os determinantes sociais da ocorrência e da evolução de diferentes doenças e eventos (por exemplo, tuberculose, câncer, violências etc).
Para tanto, uma infraestrutura computacional deverá ser projetada e implementada. Tal infraestrutura deverá auxiliar na constituição de uma coorte populacional, a partir do Cadastro Único, a qual poderá acumular progressivamente informações provenientes de outras bases de dados através de processos de linkage.
As funcionalidades previstas compreendem: i) gerência de dados: suporte para diferentes visões de dados, de acordo com as áreas de interesse (Saúde, Educação, Trabalho); suporte para a integração de diferentes bases de dados (CadÚnico, SUS); rotinas de limpeza, mineração e armazenamento de dados; ii) processamento: baseado em algoritmos parametrizáveis e no modelo MapReduce para o suporte de grandes massas de dados (big data); iii) visualização de dados: através de modelos de visualização disponíveis em portais Web; iv) segurança: rotinas para autenticação de usuários, acesso aos dados, conexões seguras e criptografia de dados sensíveis e v) dependabilidade: baseada no monitoramento da execução das aplicações, no ajuste autonômico dos recursos da infraestrutura e no aprimoramento dos algoritmos de processamento de dados.
Como resultado, espera-se que tal infraestrutura contribua no desenvolvimento e aplicação de desenhos e métodos analíticos apropriados para a estimação dos impactos dos programas sociais sobre diferentes desfechos (Saúde, Trabalho, Educação etc). Espera-se que a plataforma proposta possibilite a execução de métodos analíticos especiais, tais como Desenho de Regressão Descontínua, Pareamento com Escore de Propensão e Estimação de Variável Instrumental, além de métodos de análise multivariados clássicos, para o levantamento robusto e eficaz dos efeitos das diferentes intervenções.
GWAS in childhood asthma in a cohort of Salvador-Brazil.
Gustavo Nunes de Oliveira Costa[1], Rosemeire Fiaccone[1], Jackson Santos Conceição[1], Thiago Magalhães da Silva[1], Frank Dudbridge[2], Laura Rodrigues[2], Mauricio Lima Barreto[1].
[1] Instituto de Saúde Coletiva – UFBA
[2] London School Hygiene and Tropical Medicine (LSHTM)
Asthma is a complex disease of the respiratory tract with distinct phenotypes and has a huge impact on mortality, morbidity and quality of life. In Latin America is associated with urban and social inequality. The understanding about the genetic contribution to the causation of asthma is important and is growing in importance. Many associations have not been replicated because of false identifications, low replication power or gene-environment interactions. Studies between genetic architecture and complex have predominantly been developed in populations of European origin, whose profile differs greatly from the asthma established in Latin American populations. The aim of this work is estimating small genetic effects on the development of asthma by Genome Wide Association Study (GWAs).
1247 children were part of study population. This population was part of an observational cohort study conducted to study the impact of sanitation on diarrhoea, and the data analysed here on asthma and genetics was collected in a single moment, as part of the from the project “Social Changes. We used the ISAAC questionnaire translated into Portuguese of Brazil. Children were classified as asthmatic when parents reported wheezing in the 12 months preceding the questionnaire or diagnosis of asthma at some point in their lives. For genotyping we used the commercial panel HumanOmni2.5-8 BeadChip Kit (Illumina, Sand Diego, CA, USA), which has 2.5 million of variants genotyped distributed over the 23 human chromosomes. We used logistic regression to examine the association, assuming an additive model for the disease. PCA was performed to control for confounding. Genotypic effect were estimated using logistic regression analysis.
In this study the most important variant is on chromosome 2, rs6714508 (OR: 2.40; 95%CI: 3.38-5.0; p-value: 5.24x10-7) in GALNT14. GALNT14 was mapped to the human chromosome 2 at p23.2, is a member of the human UDP-N-Acetyl-D-Galactosamine: Polypeptide N-acetylgalactosaminyltransferase gene family, with at least twenty known subtypes. This gene is very expressed in mucosal cells (Bennett, 2012) and linked in apoptosis regulation; invasion, metastasis and proliferation of many carcinomatous cells well as glycosylation of proteins, especially mucin. Mucin is found in the airways and is responsible for protecting and lubricating the epithelium, however, they are over-expressed in chronic diseases of the airways, contributing to its obstruction in patients with asthma. In relation to chromosomal region, two studies had similar results to ours. The region 2p21 was associated with asthma for the first time in a linkage study in French population and 2p23 was associated with asthma in a meta-analysis of genome wide linkage study. The authors don’t think if this region is the causative for asthma. Despite being associated SNPs in the intronic region. We conclude that GALNT14 may be indirectly associated with asthma by regulating mucin or possibly be in linkage disequilibrium with SNPs truly causal.
Management of traffic accidents through data extracted from social media: a case study
Sibele Fausto [1], Adílson Luiz Pinto [2], Manuela Soares da Fonseca [2], Sérgio Salustiano da Silva [3]
[1] Escola de Comunicações e Artes, Universidade de São Paulo, São Paulo, Brazil
[2] Universidade Federal de Santa Catarina, Santa Catarina, Brazil
[3] Independent researcher, Rio de Janeiro, Brazil
Technologies and web-based tools open new possibilities to improve and facilitate the process of research in Public Health, contributing to the diagnosis, understanding and solving various current issues related to health in society, as traffic accidents, a serious public health problem perceived in big urban areas. The high number of accidents in cities as São Paulo requires extensive studies for their correct diagnosis and coping. In this sense, social media tools such as Twitter show great potential as data sources. The Fire Brigade of the São Paulo State adopted a new way to record and to report incidents attended, including traffic accidents: via the social media tool Twitter, disseminating all occurrences met with updates within 24 hours, and on March 10, 2013 had more than 36,000 tweets. This study sought to verify the validity of data extract from traffic accidents through the events registered by @BombeirosPMESP profile, as well as the use of new online tools for processing and analyzing data.
Methodology Occurrences tweeted by @BombeirosPMESP follow a common sequence: Time, Type of incident, Location (address), Number of victims and Referral to health services. We test a hypothesis that this sequence could facilitate data extraction from incidents recorded through these tweets, using the tools Google Refine and Google Fusion Tables. Data collection was performed considering tweets from @BombeirosPMESP recorded in February 2013, reporting only traffic accidents. The data collected from Twitter and organized into Google Refine spreadsheets been exported to Fusion Tables, which features integration with the Google Maps allowing geographic mapping. Also, in order to analyze the followers network and information dissemination based on posts from @BombeirosPMESP, we used social network analysis (ARS), through crawling with the plugin NodeXL, and data visualization was performed using the free software Gephi to obtain characteristics and properties of @BombeirosPMESP network.
Results and conclusions In February 2013 @BombeirosPMESP recorded 1,358 tweets reporting several incidents occurred at São Paulo city, from which 358 were traffic accidents. Geographic mapping showed its widespread distribution throughout the municipal area. @BombeirosPMESP’s network had 17,986 followers, and a sample of 179 generated a graph with 9,371 nodes and 9,933 edges, showing a high potential for information dissemination, because this profile can to reach 941.602 people. Regarding the data source, our initial hypothesis that the information sequence of the tweets from @BombeirosPMESP could facilitate data extraction was not validated because tweets do not have enough proper structure. It was possible to extract the addresses of occurrences to make a geographical mapping, but there is a great potential to extract more relevant information such as the type of accidents, number of victims and referrals to health services, thus enabling a comprehensive analysis of events and contributing to their correct management by public health policies. @BombeirosPMESP could to adopt a more rigorous standardization for their records on Twitter, e.g. separating the information categories by markers such as semicolons, thus facilitating data mining.
HIV and HTLV Codon Usage Support a Combined Therapy Based on Transfer RNA Inactivation
Diego Frias[1], Joana Paixao Monteiro-Cunha[2,3], Aline Cristina Mota-Miranda[2,3], Bernardo Galvao-Castro[3,4], Luiz Carlos Junior Alcantara[4].
[1] Bahia State University, Salvador, Bahia, Brazil
[2] Federal University of Bahia, Salvador, Bahia, Brazil
[3] Bahia School of Medicine and Public Health/Bahia Foundation for Science Development, Salvador, Bahia, Brazil
[4] Gonçalo Moniz Research Center, Oswaldo Cruz Foundation, Salvador, Bahia, Brazil.
The purpose of this study was to investigate the balance between tRNA supply and demand in retrovirus-infected cells, seeking the best targets for antiretroviral therapy based on the hypothetical Transfer RNA Inhibition Therapy (TRIT). Codon usage and tRNA gene data were retrieved from public databases. Based on logistic principles, a therapeutic score (T-score) was calculated for all sense codons, in each retrovirus-host system. Codons that are critical for viral protein translation, but not as critical for the host have the highest T-score values. Theoretically, inactivating the cognate tRNA species should imply a severe reduction of the elongation rate during viral mRNA translation. We developed a method to predict tRNA species critical for retroviral protein synthesis. Four of the best TRIT targets in HIV-1 and HIV-2 encode Large Hydrophobic Residues (LHR), which have a central role in protein folding. One of them, codon CUA, is also a TRIT target in both HTLV-1 and HTLV-2. Therefore, a drug designed for inactivating or reducing the cytoplasmatic concentration of tRNA species with anticodon TAG could attenuate significantly both HIV and HTLV protein synthesis rates. Inversely, replacing codons ending in UA by synonymous codons should increase the expression, which is relevant for DNA vaccine design.
Brazilian Clinical Trials Registry - Rebec: achievements and challenges of the Brazilian open access solution for clinical research registration
Luiza Rosângela da Silva (Rebec/Icict/Fiocruz,UERJ), Josué Laguardia (Rebec/Icict/Fiocruz), Jorge Otávio Maia Barreto (CGGC/Decit/MS, UFPI), Diego Tostes (Rebec/Icict/Fiocruz, UFF), Alexandre Moretto Ribeiro (UFRGS, Instituto Communitas), Daniel Pereira Eiras (Rebec/Icict/Fiocruz, Unirio), Marcelo Alves (Icict/Fiocruz, PUC-MG), Carlos Eduardo Ribeiro (PUC-RJ)
This work focuses on the trajectory of the Brazilian Clinical Trials Registry - Rebec, which among the 15 members of the International Clinical Trials Research Platform (ICTRP / WHO) is the only fully designed and conducted according to the philosophy of openness, using Open Trials, open source and providing unrestricted access to data supplied by the registrant. This prevents rework, contributes to decision-making and to the social accountability of public investment in health. It works 24/7 as the interface with a clinical research transparency network, both nationally (Plataforma Brasil, a system that integrates all ethics committees decisions about studies registered in its platform) and internationally (ICTRP / WHO).
The initiative is the only primary ICTRP registry in Portuguese and it is also bilingual (English), a feature that increases global visibility and reuse of data. These specificities, and their convergence with new trends of the vast field understood as information /knowledge management – open data , threaded publications etc. – give the platform an avant-garde position worldwide. Rebec, which was created in 2010, has over 200 thousand views that come from Brazil, USA, Australia and dozens of other countries - including India and the Lusophone Africa and Asia, showing its potential for South-South and/or intra-BRICs cooperation. There are more than 1600 registries in Rebec’s database - of which 88 are recruiting - and room for initial cooperation contacts in Latin America and the Caribbean to become formal.
On the horizon brought about by the open access movement for scientific communication, the role of information technology to provide security, ease of access, accountability and transparency to health data is undeniable. Monitoring and disseminating the results of experiments on human beings show a global trend focusing open primary data as an strategy to prevent journals’ publishing bias and to ensure data reuse in different activities by means of systems interoperability. This trend converges with the increased awareness of scientists and social movements regarding the social responsibility/accountability of science, especially in the Third World – and, in the case of Brazil, it converges with advances in legislation that make the clinical trials registration in electronic databases mandatory.
Among the technological challenges faced by Rebec´s team, is the preparation of a more friendly interface design and channeling efforts to create an efficient protocol for exchanging information with the 15 ICTRP registries and with Plataforma Brazil; today, it is done through an API and APIs statistical analysis written using Python and NumPy library. The main sustainability issue is, however, financial: although both Ministry of Health and Fiocruz show acceptance and certainty Rebec is a priority, the project is still new, which demands a new co-operational equation providing and allocating budget resources and skilled personnel for its development, for the management and exchange of experiences with national and international related initiatives, as well as for the permanent interface with varied audiences, stakeholders and partners worldwide.