Human toxocariasis in rural Brazilian Amazonia: Seroprevalence, risk factors, and spatial distribution

This population-based cross-sectional study of 403 rural settlers in Brazilian Amazonia revealed an overall rate of IgG seropositivity to Toxocara canis excretory-secretory larval antigen of 26.8% (95% confidence interval [CI], 22.5-31.4%). Multilevel logistic regression analysis identified current infection with hookworm (odds ratio [OR], 2.32; 95% CI, 1.11-4.86) and residence in the most recently occupied sectors of the settlement (OR, 1.81.; 95%CI, 1.3-2.52) as significant risk factors for Toxocara seropositivity; age > 14 years (OR, 0.46; 95% CI, 0.28-0.73) and the presence of cats in the household (OR, 0.57; 95% CI, 0.32-1.02) appeared to be protective. Two significant high-prevalence clusters were detected in the area, together comprising 38.9% of the seropositive subjects; households in the clusters had slightly lower socioeconomic status and were less likely to have cats as pets. The obstacles for controlling human toxocariasis in this and other tropical rural settings are discussed.

Malaria on the Amazonian frontier: Transmission dynamics, risk factors, spatial distribution, and prospects for control

Little follow-up data on malaria transmission in communities originating from frontier settlements in Amazonia are available. Here we describe a cohort study in a frontier settlement in Acre, Brazil, where 509 subjects contributed 489.7 person-years of follow-up. The association between malaria morbidity during the follow-up and individual, household, and spatial covariates was explored with mixed-effects logistic regression models and spatial analysis. Incidence rates for Plasmodium vivax and Plasmodium falciparum malaria were 30.0/100 and 16.3/100 person-years at risk, respectively. Malaria morbidity was strongly associated with land clearing and farming, and decreased after five years of residence in the area, suggesting that clinical immunity develops among subjects exposed to low malaria endemicity. Significant spatial clustering of malaria was observed in the areas of most recent occupation, indicating that the continuous influx of nonimmune settlers to forest-fringe areas perpetuates the cycle of environmental change and colonization that favors malaria transmission in rural Amazonia.

A Community-based Survey of Human Toxoplasmosis in Rural Amazonia: Seroprevalence, Seroconversion Rate, and Associated Risk Factors

IgG antibodies to Toxoplasma gondii were detected in, March-April 2004, in 65.8% (95% confidence interval, 60.8-70.8%) of 342 systematically sampled subjects 5-90 years of age (87.5% of the eligible) living in a rural settlement in Amazonia, with a seroconversion rate of 9% over I year of follow-up of 99 seronegative subjects. Multiple logistic regression analysis identified age as the only significant independent predictor of seropositivity at the baseline. Each additional year of age increases the odds of being seropositive by 6%, and 76.8% of the subjects are expected to be seropositive at 30 years of age. A single high-prevalence spatial cluster, comprising 11.9% of the seropositive subjects, was detected in the area; households in the cluster were less likely to have dogs as pets and their heads had a lower education level, when compared with households located outside the cluster. The challenges for preventing human toxoplasmosis in tropical rural settings are discussed.

Age-Dependent Acquisition of Protective Immunity to Malaria in Riverine Populations of the Amazon Basin of Brazil

Five community-based cross-sectional surveys of malaria morbidity and associated risk factors in remote riverine populations in northwestern Brazil showed average parasite rates of 4.2% (thick-smear microscopy) and 14.4% (polymerase chain reaction [PCR]) in the overall population, with a spleen rate of 13.9% among children 2-9 years of age. Plasmodium vivax was 2.8 times more prevalent than P. falciparum, with rare instances of P. malariae and mixed-species infections confirmed by PCR; 9.6% of asymptomatic subjects had parasitemias detected by PCR. Low-grade parasitemia detected by PCR only was a risk factor for anemia, after controlling for age and other covariates. Although clinical and subclinical infections occurred in all age groups, the risk of infection and disease decreased significantly with increasing age, after adjustment for several covariates in multilevel logistic regression models. These findings suggest that the continuous exposure to hypo- or mesoendemic malaria may induce significant anti-parasite and anti-disease immunity in native Amazonians.

Poly-bagging predictors for classification modelling for credit scoring

Credit scoring modelling comprises one of the leading formal tools for supporting the granting of credit. Its core objective consists of the generation of a score by means of which potential clients can be listed in the order of the probability of default. A critical factor is whether a credit scoring model is accurate enough in order to provide correct classification of the client as a good or bad payer. In this context the concept of bootstraping aggregating (bagging) arises. The basic idea is to generate multiple classifiers by obtaining the predicted values from the fitted models to several replicated datasets and then combining them into a single predictive classification in order to improve the classification accuracy. In this paper we propose a new bagging-type variant procedure, which we call poly-bagging, consisting of combining predictors over a succession of resamplings. The study is derived by credit scoring modelling. The proposed poly-bagging procedure was applied to some different artificial datasets and to a real granting of credit dataset up to three successions of resamplings. We observed better classification accuracy for the two-bagged and the three-bagged models for all considered setups. These results lead to a strong indication that the poly-bagging approach may promote improvement on the modelling performance measures, while keeping a flexible and straightforward bagging-type structure easy to implement. (C) 2011 Elsevier Ltd. All rights reserved.

Comparing diagnostic tests with missing data

When missing data occur in studies designed to compare the accuracy of diagnostic tests, a common, though naive, practice is to base the comparison of sensitivity, specificity, as well as of positive and negative predictive values on some subset of the data that fits into methods implemented in standard statistical packages. Such methods are usually valid only under the strong missing completely at random (MCAR) assumption and may generate biased and less precise estimates. We review some models that use the dependence structure of the completely observed cases to incorporate the information of the partially categorized observations into the analysis and show how they may be fitted via a two-stage hybrid process involving maximum likelihood in the first stage and weighted least squares in the second. We indicate how computational subroutines written in R may be used to fit the proposed models and illustrate the different analysis strategies with observational data collected to compare the accuracy of three distinct non-invasive diagnostic methods for endometriosis. The results indicate that even when the MCAR assumption is plausible, the naive partial analyses should be avoided.

Early Onset Obsessive-Compulsive Disorder With and Without Tics

Introduction: Research suggests that obsessive-compulsive disorder (OCD) is not a unitary entity, but rather a highly heterogeneous condition, with complex and variable clinical manifestations. Objective: The aims of this study were to compare clinical and demographic characteristics of OCD patients with early and late age of onset of obsessive-compulsive symptoms (OCS); and to compare the same features in early onset OCD with and without tics. The independent impact of age at onset and presence of tics on comorbidity patterns was investigated. Methods: Three hundred and thirty consecutive outpatients meeting Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria for OCD were evaluated: 160 patients belonged to the ""early onset"" group (EOG): before 11 years of age, 75 patients had an ""intermediate onset"" (IOG), and 95 patients were from the ""late onset"" group (LOG): after 18 years of age. From the 160 EOG, 60 had comorbidity with tic disorders. The diagnostic instruments used were: the Yale-Brown Obsessive Compulsive Scale and the Dimensional Yale-Brown Obsessive Compulsive Scale (DY-BOCS), Yale Global Tics Severity Scale; and Structured Clinical Interview for DSM-IV Axis I Disorders-patient edition. Statistical tests used were: Mann-Whitney, full Bayesian significance test, and logistic regression. Results: The EOG had a predominance of males, higher frequency of family history of OCS, higher mean scores on the ""aggression/violence"" and ""miscellaneous"" dimensions, and higher mean global DY-BOCS scores. Patients with EOG without tic disorders presented higher mean global DY-BOCS scores and higher mean scores in the ""contamination/cleaning"" dimension. Conclusion: The current results disentangle some of the clinical overlap between early onset OCD with and without tics. CNS Spectr. 2009; 14(7):362-370

Comparison for three Classification Techniques

In this paper we show the results of a comparison simulation study for three classification techniques: Multinomial Logistic Regression (MLR), No Metric Discriminant Analysis (NDA) and Linear Discriminant Analysis (LDA). The measure used to compare the performance of the three techniques was the Error Classification Rate (ECR). We found that MLR and LDA techniques have similar performance and that they are better than DNA when the population multivariate distribution is Normal or Logit-Normal. For the case of log-normal and Sinh(-1)-normal multivariate distributions we found that MLR had the better performance.