Predicting smear negative pulmonary tuberculosis with classification trees and logistic regression: a cross-sectional study

Abstract Background Smear negative pulmonary tuberculosis (SNPT) accounts for 30% of pulmonary tuberculosis cases reported yearly in Brazil. This study aimed to develop a prediction model for SNPT for outpatients in areas with scarce resources. Methods The study enrolled 551 patients with clinical-radiological suspicion of SNPT, in Rio de Janeiro, Brazil. The original data was divided into two equivalent samples for generation and validation of the prediction models. Symptoms, physical signs and chest X-rays were used for constructing logistic regression and classification and regression tree models. From the logistic regression, we generated a clinical and radiological prediction score. The area under the receiver operator characteristic curve, sensitivity, and specificity were used to evaluate the model's performance in both generation and validation samples. Results It was possible to generate predictive models for SNPT with sensitivity ranging from 64% to 71% and specificity ranging from 58% to 76%. Conclusion The results suggest that those models might be useful as screening tools for estimating the risk of SNPT, optimizing the utilization of more expensive tests, and avoiding costs of unnecessary anti-tuberculosis treatment. Those models might be cost-effective tools in a health care network with hierarchical distribution of scarce resources.

Ethnicity and Cutaneous Melanoma in the City of Sao Paulo, Brazil: A Case-Control Study

Background: Over the last century the incidence of cutaneous melanoma has increased worldwide, a trend that has also been observed in Brazil. The identified risk factors for melanoma include the pattern of sun exposure, family history, and certain phenotypic features. In addition, the incidence of melanoma might be influenced by ethnicity. Like many countries, Brazil has high immigration rates and consequently a heterogenous population. However, Brazil is unique among such countries in that the ethnic heterogeneity of its population is primarily attributable to admixture. This study aimed to evaluate the contribution of European ethnicity to the risk of cutaneous melanoma in Brazil. Methodology/Principal Findings: We carried out a hospital-based case-control study in the metropolitan area of Sao Paulo, Brazil. We evaluated 424 hospitalized patients (202 melanoma patients and 222 control patients) regarding phenotypic features, sun exposure, and number of grandparents born in Europe. Through multivariate logistic regression analysis, we found the following variables to be independently associated with melanoma: grandparents born in Europe-Spain (OR = 3.01, 95% CI: 1.03-8.77), Italy (OR = 3.47, 95% CI: 1.41-8.57), a Germanic/Slavic country (OR = 3.06, 95% CI: 1.05-8.93), or >= 2 European countries (OR = 2.82, 95% CI: 1.06-7.47); eye color-light brown (OR = 1.99, 95% CI: 1.14-3.84) and green/blue (OR = 4.62; 95% CI 2.22-9.58); pigmented lesion removal (OR = 3.78; 95% CI: 2.21-6.49); no lifetime sunscreen use (OR = 3.08; 95% CI: 1.03-9.22); and lifetime severe sunburn (OR = 1.81; 95% CI: 1.03-3.19). Conclusions: Our results indicate that European ancestry is a risk factor for cutaneous melanoma. Such risk appears to be related not only to skin type, eye color, and tanning capacity but also to others specific characteristics of European populations introduced in the New World by European immigrants.

Response plateau models fitting via isotonic regression

Within the nutritional context, the supplementation of microminerals in bird food is often made in quantities exceeding those required in the attempt to ensure the proper performance of the animals. The experiments of type dosage x response are very common in the determination of levels of nutrients in optimal food balance and include the use of regression models to achieve this objective. Nevertheless, the regression analysis routine, generally, uses a priori information about a possible relationship between the response variable. The isotonic regression is a method of estimation by least squares that generates estimates which preserves data ordering. In the theory of isotonic regression this information is essential and it is expected to increase fitting efficiency. The objective of this work was to use an isotonic regression methodology, as an alternative way of analyzing data of Zn deposition in tibia of male birds of Hubbard lineage. We considered the models of plateau response of polynomial quadratic and linear exponential forms. In addition to these models, we also proposed the fitting of a logarithmic model to the data and the efficiency of the methodology was evaluated by Monte Carlo simulations, considering different scenarios for the parametric values. The isotonization of the data yielded an improvement in all the fitting quality parameters evaluated. Among the models used, the logarithmic presented estimates of the parameters more consistent with the values reported in literature.

Spatial clusters of suicide in the municipality of Sao Paulo 1996-2005: an ecological study

Background: In a classical study, Durkheim mapped suicide rates, wealth, and low family density and realized that they clustered in northern France. Assessing others variables, such as religious society, he constructed a framework for the analysis of the suicide, which still allows international comparisons using the same basic methodology. The present study aims to identify possible significantly clusters of suicide in the city of Sao Paulo, and then, verify their statistical associations with socio-economic and cultural characteristics. Methods: A spatial scan statistical test was performed to analyze the geographical pattern of suicide deaths of residents in the city of Sao Paulo by Administrative District, from 1996 to 2005. Relative risks and high and/or low clusters were calculated accounting for gender and age as co-variates, were analyzed using spatial scan statistics to identify geographical patterns. Logistic regression was used to estimate associations with socioeconomic variables, considering, the spatial cluster of high suicide rates as the response variable. Drawing from Durkheim's original work, current World Health Organization (WHO) reports and recent reviews, the following independent variables were considered: marital status, income, education, religion, and migration. Results: The mean suicide rate was 4.1/100,000 inhabitant-years. Against this baseline, two clusters were identified: the first, of increased risk (RR = 1.66), comprising 18 districts in the central region; the second, of decreased risk (RR = 0.78), including 14 districts in the southern region. The downtown area toward the southwestern region of the city displayed the highest risk for suicide, and though the overall risk may be considered low, the rate climbs up to an intermediate level in this region. One logistic regression analysis contrasted the risk cluster (18 districts) against the other remaining 78 districts, testing the effects of socioeconomic-cultural variables. The following categories of proportion of persons within the clusters were identified as risk factors: singles (OR = 2.36), migrants (OR = 1.50), Catholics (OR = 1.37) and higher income (OR = 1.06). In a second logistic model, likewise conceived, the following categories of proportion of persons were identified as protective factors: married (OR = 0.49) and Evangelical (OR = 0.60). Conclusions: This risk/ protection profile is in accordance with the interpretation that, as a social phenomenon, suicide is related to social isolation. Thus, the classical framework put forward by Durkheim seems to still hold, even though its categorical expression requires re-interpretation.

Relationship Between Occlusion and Lisping in Children With Cleft Lip and Palate

Introduction: The literature suggests that individuals with history of cleft lip and palate who present with midfacial growth deficiency are at higher risk of presenting lisping. The relationship between distortions during production of linguoalveolar fricative sounds and the severity of malocclusion, however, has not been established for the population with cleft. Objective: To study the association between lisping and dental arch relationship. Methodology: Speech samples and dental arch casts were obtained from 106 children with operated unilateral cleft lip and palate (UCLP) during the stage of mixed dentition and before orthodontic treatment. Videotaped productions of the phrase/u saci saiw sedu/were rated by speech-language pathologists for the identification of lisping during [s]. Dental arch casts were rated by orthodontists using the Goslon Yardstick and the Five-Year Index to establish dental arch relationship. Results: Multiple logistic regression showed no significant association between lisping and dento-occlusal index (p = .802) and age (p = .662). Substantial interjudge agreement during auditory-perceptual ratings was found (kappa = .63). Almost perfect agreement was found between orthodontists while establishing the dental arch relationship (kappa = .81). Discussion: This study failed to reveal an association between lisping and dental arch relationship in children with operated UCLP. Multiple variables may play a role in determining occurrence of lisping, warranting further investigation.

On the impact of disproportional samples in credit scoring models: An application to a Brazilian bank data

Statistical methods have been widely employed to assess the capabilities of credit scoring classification models in order to reduce the risk of wrong decisions when granting credit facilities to clients. The predictive quality of a classification model can be evaluated based on measures such as sensitivity, specificity, predictive values, accuracy, correlation coefficients and information theoretical measures, such as relative entropy and mutual information. In this paper we analyze the performance of a naive logistic regression model (Hosmer & Lemeshow, 1989) and a logistic regression with state-dependent sample selection model (Cramer, 2004) applied to simulated data. Also, as a case study, the methodology is illustrated on a data set extracted from a Brazilian bank portfolio. Our simulation results so far revealed that there is no statistically significant difference in terms of predictive capacity between the naive logistic regression models and the logistic regression with state-dependent sample selection models. However, there is strong difference between the distributions of the estimated default probabilities from these two statistical modeling techniques, with the naive logistic regression models always underestimating such probabilities, particularly in the presence of balanced samples. (C) 2012 Elsevier Ltd. All rights reserved.