Adelmo Inácio Bertolde

Epidemiology of extrapulmonary tuberculosis in Brazil: A hierarchical model

BackgroundAlthough extrapulmonary tuberculosis (EPTB) is less frequent than Pulmonary Tuberculosis (PTB) and is a secondary target for national TB control programs, its significance has increased worldwide during the HIV epidemic. The objective of this study was to examine the epidemiology of EPTB in Brazil between 2007 and 2011.MethodsCross-sectional study involving all cases of TB reported to the Brazilian Notifiable Diseases Surveillance System (Sistema de Informações de Agravo de Notificação - SINAN) in Brazil between 2007 and 2011. Sociodemographic and clinical characteristics of patients with exclusively PTB and exclusively EPTB were compared. Following analysis with Pearson’s chi-square test, variables with p < 0.05 were included in a hierarchical regression model. Variables with p < 0.05 in the corresponding level were kept in the model.ResultsA total of 427,548 cases of TB were included. Of these, 356,342 cases (83.35%; 95% confidence interval (CI) 83.23% - 83.45%) were PTB, 57,217 (13.37%; 95% CI 13.28% - 13.48%) were EPTB, 13,989 (3.27%; 95% CI 3.21% - 3.32%) were concurrent pulmonary and extrapulmonary TB. Patients with EPTB were mainly white (16.7%), and most (29.1%) patients had five to eight years of education. Among comorbidities, HIV infection was prominent (OR 2.15; 95% CI 2.09 – 2.21), although the proportion of cases awaiting test results or untested was high (39%). Ethanol use (OR 0.45; 95% CI 0.43 – 0.46), diabetes mellitus (OR 0.54; 95% CI 0.51 – 0.57) and mental illness (OR 0.88; 95% CI 0.82 – 0.95) were associated with PTB.ConclusionsThirteen percent of patients diagnosed with TB in Brazil have only EPTB. More effective diagnostic strategies and control measures are needed to reduce the number of cases of extrapulmonary TB in Brazil.

Genotypic and Spatial Analysis of Mycobacterium tuberculosis Transmission in a High-Incidence Urban Setting

BACKGROUND Genotyping Mycobacterium tuberculosis isolates allows study of dynamics of tuberculosis transmission, while geoprocessing allows spatial analysis of clinical and epidemiological data. Here, genotyping data and spatial analysis were combined to characterize tuberculosis transmission in Vitória, Brazil, to identify distinct neighborhoods and risk factors associated with recent tuberculosis transmission. METHODS From 2003 to 2007, 503 isolates were genotyped by IS6110 restriction fragment length polymorphism (RFLP) and spoligotyping. The analysis included kernel density estimation, K-function analysis, and a t test distance analysis. Mycobacterium tuberculosis isolates belonging to identical RFLP patterns (clusters) were considered to represent recent tuberculosis infection (cases). RESULTS Of 503 genotyped isolates, 242 (48%) were categorized into 70 distinct clusters belonging to 12 RFLP families. The proportion of recent transmission was 34.2%. Kernel density maps indicated 3 areas of intense concentration of cases. K-function analysis of the largest RFLP clusters and families showed they co-localized in space. The distance analysis confirmed these results and demonstrated that unique strain patterns (controls) randomly distributed in space. A logit model identified young age, positive smear test, and lower Index of Quality of Urban Municipality as risk factors for recent transmission. The predicted probabilities for each neighborhood were mapped and identified neighborhoods with high risk for recent transmission. CONCLUSIONS Spatial and genotypic clustering of M. tuberculosis isolates revealed ongoing active transmission of tuberculosis caused by a small subset of strains in specific neighborhoods of the city. Such information provides an opportunity to target tuberculosis transmission control, such as through rigorous and more focused contact investigation programs.

Spatial analysis of distribution of dengue cases in Espírito Santo, Brazil, in 2010: use of Bayesian model

OBJECTIVE To study the relationship between the risk of dengue and sociodemographic variables through the use of spatial regression models fully Bayesian in the municipalities of Espírito Santo in 2010. METHOD This is an ecological study and exploration that used spatial analysis tools in preparing thematic maps with data obtained from SinanNet. An analysis by area, taking as unit the municipalities of the state, was performed. Thematic maps were constructed by the computer program R 2.15.00 and Deviance Information Criterion (DIC), calculated in WinBugs, Absolut and Normalized Mean Error (NMAE) were the criteria used to compare the models. RESULTS We were able to geocode 21,933 dengue cases (rate of 623.99 cases per 100 thousand habitants) with a higher incidence in the municipalities of Vitória, Serra and Colatina; model with spatial effect with the covariates trash and income showed the best performance at DIC and Nmae criteria. CONCLUSION It was possible to identify the relationship of dengue with factors outside the health sector and to identify areas with higher risk of disease.

Treatment outcomes of MDR-tuberculosis patients in Brazil: a retrospective cohort analysis

BackgroundMultidrug-resistant tuberculosis (MDR-TB) is a threat for the global TB epidemic control. Despite existing evidence that individualized treatment of MDR-TB is superior to standardized regimens, the latter are recommended in Brazil, mainly because drug-susceptibility tests (DST) are often restricted to first-line drugs in public laboratories. We compared treatment outcomes of MDR-TB patients using standardized versus individualized regimens in Brazil, a high TB-burden, low resistance setting.MethodsThe 2007–2013 cohort of the national electronic database (SITE-TB), which records all special treatments including drug-resistance, was analysed. Patients classified as MDR-TB in SITE-TB were eligible. Treatment outcomes were classified as successful (cure/treatment completed) or unsuccessful (failure/relapse/death/loss to follow-up). The odds for successful treatment according to type of regimen were controlled for demographic and clinical variables.ResultsOut of 4029 registered patients, we included 1972 recorded from 2010 to 2012, who had more complete outcome data. The overall success proportion was 60%. Success was more likely in non-HIV patients, sputum-negative at baseline, with unilateral disease and without prior DR-TB. Adjusted for these variables, those receiving standardized regimens had 2.7-fold odds of success compared to those receiving individualized treatments when failure/relapse were considered, and 1.4-fold odds of success when death was included as an unsuccessful outcome. When loss to follow-up was added, no difference between types of treatment was observed. Patients who used levofloxacin instead of ofloxacin had 1.5-fold odds of success.ConclusionIn this large cohort of MDR-TB patients with a low proportion of successful outcomes, standardized regimens had superior efficacy than individualized regimens, when adjusted for relevant variables. In addition to the limitations of any retrospective observational study, database quality hampered the analyses. Also, decision on the use of standard or individualized regimens was possibly not random, and may have introduced bias. Efforts were made to reduce classification bias and confounding. Until higher-quality evidence is produced, and DST becomes widely available in the country, our findings support the Brazilian recommendation for the use of standardized instead of individualized regimens for MDR-TB, preferably containing levofloxacin. Better quality surveillance data and DST availability across the country are necessary to improve MDR-TB control in Brazil.

Single-function approach to calibrating whatman no. 42 filter paper based on suction versus water content relationships

Several suction–water-content (s-w) calibrations for the filter paper method (FPM) used for soil-suction measurement have been published. Most of the calibrations involve a bilinear function (i.e., two different equations) with an inflection point occurring at 60 kPa

Introducing risk inequality metrics in tuberculosis policy development

Global stakeholders including the World Health Organization rely on predictive models for developing strategies and setting targets for tuberculosis care and control programs. Failure to account for variation in individual risk leads to substantial biases that impair data interpretation and policy decisions1,2. Anticipated impediments to estimating heterogeneity for each parameter are discouraging despite considerable technical progress in recent years. Here we identify acquisition of infection as the single process where heterogeneity most fundamentally impacts model outputs, due to cohort selection imposed by dynamic forces of infection. Individuals with higher risk of acquiring infection are predominantly affected by the pathogen, leaving the unaffected pool with those whose intrinsic risk is lower. This causes susceptibility pools to attain average risks which are lower under higher forces of infection. Interventions that modify the force of infection change the strength of selection, and therefore alter average risks in the pools which feed further incidence. Inability to account for these dynamics is what makes homogenous models unsuitable. We introduce concrete metrics to approximate risk inequality in tuberculosis, demonstrate their utility in mathematical models, and pack the information into a risk inequality coefficient which can be calculated and reported by national tuberculosis programs for use in policy development and modeling.

Social determinants of tuberculosis via a zero-inflated model in small areas of a city in Southeastern Brazil

INTRODUCTION This study aimed to analyze social factors involved in the spatial distribution and under-reporting of tuberculosis (TB) in the city of Vitória, Espírito Santo State, Brazil. METHODS This was an ecological study of the reported cases of TB between 2009 and 2011, according to census tracts. The outcome was TB incidence for the study period and the variables of exposure were proportions of literacy, inhabitants with an income of up to half the minimum monthly wage (MMW), and inhabitants associated with sewer mains or with access to safe drinking water. We used a zero-inflated process, zero-inflated negative binomial regression (ZINB), and selected an explanatory model based on the Akaike Information Criterion (AIC). RESULTS A total of 588 cases of tuberculosis were reported in Vitória during the study period, distributed among 223 census tracts (38.6%), with 354 (61.4%) tracts presenting zero cases. In the ZINB model, the mean value of p i was 0.93, indicating that there is a 93% chance that an observed false zero could be due to sub-notification. CONCLUSIONS It is important to prioritize areas exhibiting determinants that influence the occurrence of TB in the municipality of Vitória. The zero-inflated model can be useful to the public health sector since it identifies the percentage of false zeros, generating an estimate of the real epidemiological condition of TB in Vitória.

UMA COMPARAÇÃO DE MÉTODOS DE PREVISÃO DE DEMANDA DE PEÇAS DE REPOSIÇÃO: UMA APLICAÇÃO AO TRANSPORTE FERROVIÁRIO

O objetivo principal deste trabalho e comparar o desempenho de modelos de previsao de demanda para pecas de reposicao utilizadas em manutencoes, apresentando suas aplicacoes em dados de uma industria real. A importância crescente do papel da manutencao no novo ambiente empresarial tem provocado o aumento da quantidade de pesquisas na area de modelagem e otimizacao da manutencao. Nesse contexto, este trabalho compara o desempenho de modelos de previsao de demanda para pecas de reposicao utilizadas em manutencoes de motores a diesel, a partir de uma aplicacao a um conjunto de dados de uma industria real, que atua na area de transporte ferroviario. Os modelos experimentados foram Media Movel, Suavizacao Exponencial Simples, Holt, Croston e ARIMA, alem do metodo utilizado pela empresa estudada nesse trabalho, que tem o seu nome omitido. Apos analisar uma grande amostra de pecas de reposicao, os resultados mostraram uma forte evidencia de superioridade do modelo de Suavizacao Exponencial Simples em relacao ao metodo de previsao utilizado pela empresa em questao. Este trabalho tem como principal contribuicao apresentar a necessidade da utilizacao de metodos mais fundamentados na previsao de demanda de consumo de materiais na manutencao para que os impactos no nivel de atendimento e os custos na cadeia de suprimento possam ser controlados.