Fit Of Logistic Regression Model Research Articles

Serum cytokine biomarker panels for discriminating pancreatic cancer from benign pancreatic disease.

BackgroundWe investigated whether combinations of serum cytokines, used with logistic disease predictor models, could facilitate the detection of pancreatic ductal adenocarcinoma (PDAC).MethodsThe serum levels of 27 cytokines were measured in 241 subjects, 127 with PDAC, 49 with chronic pancreatitis, 20 with benign biliary obstruction and 45 healthy controls. Samples were split randomly into independent training and test sets. Cytokine biomarker panels were selected by identifying the top performing cytokines in best fit logistic regression models during multiple rounds of resampling from the training dataset. Disease prediction by logistic models, built using the resulting cytokine panels, was evaluated with training and test sets and further examined using resampled performance evaluation.ResultsFor the discrimination of PDAC patients from patients with benign disease, a panel of IP-10, IL-6, PDGF plus CA19-9 offered improved diagnostic performance over CA19-9 alone in the training (AUC 0.838 vs. 0.678) and independent test set (AUC 0.884 vs. 0.798). For the discrimination of PDAC from CP, a panel of IL-8, CA19-9, IL-6 and IP-10 offered improved diagnostic performance over CA19-9 alone with the training (AUC 0.880 vs. 0.758) and test set (AUC 0.912 vs. 0.848). Finally, for the discrimination of PDAC in the presence of jaundice from benign controls with jaundice, a panel of IP-10, IL-8, IL-1b and PDGF demonstrated improvement over CA19-9 in the training (AUC 0.810 vs. 0.614) and test set (AUC 0.857 vs. 0.659).ConclusionsThese findings support the potential role for cytokine panels in the discrimination of PDAC from patients with benign pancreatic diseases and warrant additional study.

Willingness by people living with HIV/AIDS to utilize HIV services provided by village health teams in Kalungu district, central Uganda.

Introduction: Less than 25% of people in need of HIV services in Uganda have access. Village health teams (VHTs) provide a front-line service at the local level. Made up of volunteers with basic medical training the VHT members could integrate HIV service provision into their activities. This study assessed the willingness of people living with HIV (PLWAs) to utilize HIV services provided by VHT members in Kalungu district, central Uganda. Methods: A cross sectional study was conducted in health facilities providing anti-retroviral therapy. A random sample stratified by gender was used to recruit 312 people living with HIV/AIDS into the study. Participants were visited in their villages and a questionnaire was administered. Quantitative data were analyzed using Stata version 10 software. Univariate analysis was done for descriptive statistics. Bivariate analysis using odds ratios at 95% CI was done to identify important factors associated with willingness of utilization of VHT services. A forward fitting logistic regression model was used to assess the factors associated with willingness of PLWAs to utilize services provided by VHTs while controlling for confounders. Results: Overall, 49% (95% CI 43.43-56.42) of the study participants were willing to utilize HIV services provided by VHTs. This proportion increased to 75.6% (95% CI 70.76-80.36) if the VHT member was HIV positive. PLWAs who resided in urban areas were more likely to utilize HIV services provided by VHTs (Adjusted OR 0.24, 95% CI 0.06-0.87). Barriers to utilizing HIV services provided by VHTs included income level of greater than USD 40 (Adjusted OR 6.43 95% CI 1.19-34.68), being a business person (adjusted OR 8.71 95% CI 1.23-61.72), being a peasant (Adjusted OR 7.95 95% CI 1.37-46.19) lack of peer encouragement (Adjusted OR 6.33 95% CI 1.43-28.09), lack of encouragement from spouse (Adjusted OR 4.93 95% CI 1.23-19.82) and lack of encouragement from community leader (Adjusted OR 9.67 95% CI 3.35-27.92). Conclusion: PLWA are more willing to utilize HIV services provided by HIV positive VHT members compared to VHT members who are HIV negative. Having social support could improve willingness by PLWAs to utilize HIV services provided by VHTs. These factors can be used to increase access of HIV services to the general population.

Regularization Paths for Conditional Logistic Regression: TheclogitL1Package

We apply the cyclic coordinate descent algorithm of Friedman, Hastie, and Tibshirani (2010) to the fitting of a conditional logistic regression model with lasso [Formula: see text] and elastic net penalties. The sequential strong rules of Tibshirani, Bien, Hastie, Friedman, Taylor, Simon, and Tibshirani (2012) are also used in the algorithm and it is shown that these offer a considerable speed up over the standard coordinate descent algorithm with warm starts. Once implemented, the algorithm is used in simulation studies to compare the variable selection and prediction performance of the conditional logistic regression model against that of its unconditional (standard) counterpart. We find that the conditional model performs admirably on datasets drawn from a suitable conditional distribution, outperforming its unconditional counterpart at variable selection. The conditional model is also fit to a small real world dataset, demonstrating how we obtain regularization paths for the parameters of the model and how we apply cross validation for this method where natural unconditional prediction rules are hard to come by.

Achieving effective universal health coverage with equity: evidence from Chile

Chile's 'health guarantees' approach to providing universal and equitable coverage for quality healthcare in a dual public-private health system has generated global interest. The programme, called AUGE, defines legally enforceable rights to explicit healthcare benefits for priority health conditions, which incrementally covered 56 problems representing 75% of the disease burden between 2005 and 2009. It was accompanied by other health reform measures to increase public financing and public sector planning to secure the guarantees nationwide, as well as the state's stewardship role. We analysed data from household surveys conducted before and after the AUGE reform to estimate changes in levels of unmet health need, defined as the lack of a healthcare visit for a health problem occurring in the last 30 days, by age, sex, income, education, health insurance, residence and ethnicity; fitting logistic regression models and using predictive margins. The overall prevalence of unmet health need was much lower in 2009 (17.6%, 95% CI: 16.5%, 18.6%) than in 2000 (30.0%, 95% CI: 28.3%, 31.7%). Differences by income and education extremes and rural-urban residence disappeared. In 2009, people who had been in treatment for a condition covered by AUGE in the past year had a lower adjusted prevalence of unmet need for their recent problem (11.7%, 95% CI: 10.5%, 13.2%) than who had not (21.0%, 95% CI: 19.6%, 22.4%). Despite limitations including cross-sectional and self-reported data, our findings suggest that the Chilean health system has become more equitable and responsive to need. While these changes cannot be directly attributed to AUGE, they were coincident with the AUGE reforms. However, healthcare equity concerns are still present, relating to quality of care, health system barriers and differential access for health conditions that are not covered by AUGE.

Bayesian Estimation of Age and Length at 50% Maturity

AbstractFish age and length at 50% maturity are used extensively in the management of exploited fish populations. These parameters are historically estimated using logistic regression models (e.g., frequentist inference) for individual year‐classes and often fail to converge or result in insignificant results when a small sample size is used. The sample‐size problem motivated us to evaluate whether a hierarchical logistic regression model fit using frequentist inference or Bayesian inference, could improve our ability to fit these models. Our objective was to compare Bayesian and frequentist inference for estimating age and length at 50% maturity to determine whether the models produced similar values. To make this evaluation, we used a long‐term data set of Yellow Perch Perca flavescens from southern Lake Michigan. Frequentist inference of the year‐class‐specific models resulted in significant results when sample size was sufficiently large, a result that occurred in 76% of the models. The hierarchical model produces estimates of age (or length) at 50% maturity for all year‐classes using both frequentist and Bayesian inference. However, Bayesian inference of the hierarchical model resulted in more precise parameter estimates and provided the complete posterior distribution in one seamless and easy approach, and the computation time was 78% to 83% faster. We suggest that a hierarchical model fit using Bayesian inference of age (or length) at 50% maturity is an improvement over frequentist interference methods by providing more information about the population of interest, particularly when sample sizes are limited.

Identifying and quantifying secondhand smoke in source and receptor rooms: logistic regression and chemical mass balance approaches

Identifying and quantifying secondhand tobacco smoke (SHS) that drifts between multiunit homes is critical to assessing exposure. Twenty-three different gaseous and particulate measurements were taken during controlled emissions from smoked cigarettes and six other common indoor source types in 60 single-room and 13 two-room experiments. We used measurements from the 60 single-room experiments for (i) the fitting of logistic regression models to predict the likelihood of SHS and (ii) the creation of source profiles for chemical mass balance (CMB) analysis to estimate source apportionment. We then applied these regression models and source profiles to the independent data set of 13 two-room experiments. Several logistic regression models correctly predicted the presence of cigarette smoke more than 80% of the time in both source and receptor rooms, with one model correct in 100% of applicable cases. CMB analysis of the source room provided significant PM2.5 concentration estimates of all true sources in 9 of 13 experiments and was half-correct (i.e., included an erroneous source or missed a true source) in the remaining four. In the receptor room, CMB provided significant estimates of all true sources in 9 of 13 experiments and was half-correct in another two.

Interpreting Red Blood Cells in Lumbar Puncture: Distinguishing True Subarachnoid Hemorrhage From Traumatic Tap

The study purpose was to determine the optimal use of lumbar puncture (LP) red blood cell (RBC) counts to identify subarachnoid hemorrhage (SAH) when some blood remains in the final tube. A case series was performed at a tertiary emergency department (ED). Records of 4,496 consecutive adult patients billed for LPs between 2001 and 2009 were reviewed. Inclusion criteria were headache (HA), final tube RBCs ≥ 5, and neurovascular imaging within 2 weeks of the visit. Demographics, relevant history and physical examination components, LP results, and neuroimaging findings were recorded for 280 patients. True-positive (TP) and true-negative (TN) SAH were strictly defined. Primary outcomes were the areas under the receiver operating characteristic curves (AUC) for final tube RBC count, differential RBC count between the final and initial tubes, and absolute differential RBC count between the final and initial tubes divided by the mean RBC count of the two tubes (also called the percent change in RBC count). There were 26 TP and 196 TN results; 58 patients were neither. The TP group consisted of 19 patients with visible or possible SAH on imaging (17 on noncontrast head computed tomography [CT; 12 definite and five possible] and two on magnetic resonance imaging), six with xanthochromia and a vascular lesion (aneurysm or arteriovenous malformation [AVM] > 2 mm), and one with xanthochromia and polymerase chain reaction (PCR)-positive meningitis. As a test for SAH, final tube RBC AUC was 0.85 (95% confidence interval [CI] = 0.80 to 0.91). Interval likelihood ratios (LRs) for final tube RBC count were LR 0 (95% CI = 0 to 0.3) for RBCs < 100, LR 1.6 (95% CI = 1.1 to 2.3) for 100 < RBCs < 10,000, and LR 6.3 (95% CI = 4.8 to 23.4) for RBCs > 10,000. Differential RBC count was not associated with SAH, with AUC 0.45 (95% CI = 0.31 to 0.60). However, the percent change in RBC count between the final and initial tubes had an AUC 0.84 (95% CI = 0.78 to 0.90), and the optimal test threshold for SAH was 0.63, with positive LR 3.6 (95% CI = 2.7 to 4.7) and negative LR 0.10 (95% CI = 0.03 to 0.4) for percent change <63% and >63%, respectively. This test added additional independent information to the final tube RBC count based on improved logistic regression model fit and discriminatory ability as measured by the LR test and c statistic, respectively. Final LP tube RBC count and the percent change in RBC count, but not the simple differential count between the final and initial tubes, were associated with SAH. In this sample, there were no patients with SAH who had RBCs < 100 in the final tube, and RBCs > 10,000 increased the odds of SAH by a factor of 6.

Multiracial Church Attendance and Support for Same‐Sex Romantic and Family Relationships

Research has found that attending a racially diverse congregation promotes more favorable attitudes toward interracial dating, marriage, and adoption, but does participation in an integrated faith community promote tolerance toward other non‐traditional romantic and family forms? This study examines the relationship between involvement in a racially diverse congregation and support for same‐sex romantic and family relationships. Data are taken from the 2005 Baylor Religion Survey. I fit logistic regression models to estimate the effect of attending a multiracial church on support for homosexuality, same‐sex marriage, and same‐sex adoption, net of sociodemographic and religious controls. Results indicate that persons who attend churches where between 25 and 75 percent of attendees are of another race are more likely to support gay sex, marriage, and adoption compared with those who attend more racially homogenous churches. This relationship generally holds when models are estimated for evangelicals and mainline Protestants separately, but not for Catholics. Findings suggest a link between religion‐based racial prejudice and heterosexism/homophobia and that increased exposure to racial diversity may promote a general tolerance toward non‐traditional romantic couples and families.

Predictors of Legionella occurrence in dental unit waterlines of a highly colonized dental hospital

Introduction. Legionella is frequently detected in Dental Unit Waterlines (DUWLs). Although such a high occurrence is not necessarily associated with high risk for Legionnaire's disease among patients and staff, it is prudent to monitor DUWLs for Legionella periodically. Since this procedure is long and expensive, surrogate markers are frequently used. Aim. To investigate whether surrogate markers are predictive of Legionella detection in DUWLs in a highly colonized dental hospital. Material and methods. DUWLs from a dental hospital where legionellae were detected intermittently throughout a period of ten years was considered. The investigated predictors were total viable flora (TVF) at 37°C and at 22°C, Pseudomonas (legionellae competitor) occurrence and season. Multivariate analysis was made and, using the best fitting logistic regression model, the probability to detect legionellae in water from DUWLs was estimated. Results. Legionellae were detected in 52% water samples collected in summertime and never detected in wintertime at levels ranging between 0 and 200 colony forming units(CFU)/L. The odds ratio of legionellae occurrence were 25.0 for Pseudomonas undetected vs. detected, 108.3 for summertime vs. wintertime, 2.2-2.3 for TVF levels at 37°C and 22°C >200 CFU/mL vs. ≤200 CFU/mL. A 29% probability to detect legionellae from DUWLs, where Pseudomonas was undetected, TVF levels were >200 CFU/mL and in summertime, was estimated. Conclusion. Despite legionellae were ubiquitous in the dental hospital during the study period, in the most favourable conditions for Legionella growth (lack of competitor, high biofilm and hot weather), legionellae were detected in almost one third of DUWLs.

Exploring the Association of Psychological Status with Self-Rated Diabetes Control: Results from the Montreal Evaluation of Diabetes Treatment Study

There is an increasing interest in single-item self-rated indicators of perceived health and control status in people with chronic illnesses such as diabetes. However, self-rated measures can be associated with indicators of psychological status. The aim of this paper is to explore the association of anxiety, depression, and diabetes distress with self-rated diabetes control. Telephone interviews were conducted with 1,787 people with type 2 diabetes taking oral hypoglycemic medication. Diabetes control, health behaviors, and outcomes, anxiety, depression, and diabetes distress were assessed by standardized questionnaires. Self-reported diabetes control was modeled using logistic regression. The best fit logistic regression model for self-rated poor diabetes control was a model that incorporated diabetes distress. When adjusted for age, sex, and all other health behaviors and outcomes, poor diabetes control was most associated with diabetes distress, physical inactivity, being overweight, and poor eating habits. Results from this study indicate that poor self-rated diabetes control shares the strongest associations with diabetes-specific distress along with perceptions of diabetes-specific healthcare behaviors and outcomes.

Prenatal, perinatal, early life, and sociodemographic factors underlying racial differences in the likelihood of high body mass index in early childhood.

We investigated early childhood disparities in high body mass index (BMI) between Black and White US children. We compared differences in Black and White children's prevalence of sociodemographic, prenatal, perinatal, and early life risk and protective factors; fit logistic regression models predicting high BMI (≥ 95th percentile) at age 4 to 5 years to 2 nationally representative samples followed from birth; and performed separate and pooled-survey estimations of these models. After adjustment for sample design-related variables, models predicting high BMI in the 2 samples were statistically indistinguishable. In the pooled-survey models, Black children's odds of high BMI were 59% higher than White children's (odds ratio [OR] = 1.59; 95% confidence interval [CI]= 1.32, 1.92). Sociodemographic predictors reduced the racial disparity to 46% (OR = 1.46; 95% CI = 1.17, 1.81). Prenatal, perinatal, and early life predictors reduced the disparity to nonsignificance (OR = 1.18; 95% CI = 0.93, 1.49). Maternal prepregnancy obesity and short-duration or no breastfeeding were among predictors for which racial differences in children's exposures most disadvantaged Black children. Racial disparities in early childhood high BMI were largely explained by potentially modifiable risk and protective factors.

Determinants of Small for Gestational Age Birth at Term

Being born small for gestational age (SGA) is an indicator of intrauterine growth restriction (IUGR) and later health risks. This study investigated determinants of severe and moderate SGA (respectively, birthweight <3rd percentile and 3rd to <10th percentile for gestational age and sex). A total of 2195 term pregnancies from a prospective cohort were studied. Prenatal data arose from maternal interview at 10-22 weeks of gestation and perinatal data were collected from hospital charts. Severe and moderate SGA were classified by Canadian population standards. Risk factors for SGA were identified from fitting multivariable logistic regression models. Multivariable associations with severe SGA were: maternal age ≥ 35 [odds ratio (OR) 3.2 [95% confidence interval (CI) 1.4, 6.9]], maternal smoking during pregnancy (OR 5.3 [95% CI 2.4, 11.7]), preeclampsia (OR 4.6 [95% CI 1.6, 13.2]) and threatened preterm labour (OR 3.9 [95% CI 1.3, 11.4]). Primiparity was associated with both severe and moderate SGA with OR 2.4 [95% CI 1.1, 5.1] and OR 1.9 [95% CI 1.3, 2.9] respectively. Underweight pre-pregnancy body mass index was associated with moderate SGA (OR 2.4 [95% CI 1.2, 5.0]). Inclusion of placental weight, in the final model attenuated the associations. This study demonstrated different determinants for severe and moderate SGA. We speculate that the majority of severe SGA infants are IUGR while moderate SGA infants may be a mixture of IUGR and constitutionally small newborns. This study has also contributed evidence linking preterm labour and SGA as two, potentially related, outcomes of overlapping causal mechanisms reflective of ischaemic placental disease.

Mark‐resight and sightability modeling of a western Washington elk population

AbstractThe North Cascades (Nooksack) elk (Cervus elaphus) population declined during the 1980s, prompting a closure to state and tribal hunting in 1997 and an effort to restore the herd to former abundance. In 2005, we began a study to assess the size of the elk population, judge the effectiveness of restoration efforts, and develop a practical monitoring strategy. We concurrently evaluated 2 monitoring approaches: sightability correction modeling and mark‐resight modeling. We collected data during February–April helicopter surveys and fit logistic regression models to predict the sightability of elk groups based on group and environmental variables. We used an information‐theoretic criterion to compare 9 models of varying complexity; the best model predicted sightability of elk groups based on 1) transformed (log2) group size, 2) forest canopy cover (%), and 3) a categorical activity variable (active vs. bedded). The sightability model indicated relatively steady and modest herd growth during 2006–2011, but estimates were less than minimum‐known‐alive counts. We also used the logit‐normal mixed effects (LNME) mark‐resight model to generate estimates of total elk population size and the sizes of the adult female and branch‐antlered male subpopulations. We explored 15 LNME models to predict total population size and 12 models to predict subpopulations. Our results indicated individual heterogeneity in resighting probabilities and variation in resighting probabilities across sexes and some years. Model‐averaged estimates of total population size increased from 639 (95% CI = 570–706) in spring 2006 to 1,248 (95% CI = 1,094–1,401) in 2011. We estimated the adult female subpopulation increased from 381 (95% CI = 338–424) in spring 2006 to 573 (95% CI = 507–639) by 2011. The branch‐antlered male subpopulation estimates increased from 87 (95% CI = 54–119) to 180 (95% CI = 118–241) from spring 2006 to spring 2011. The LNME model estimates were greater than sightability model estimates and minimum‐known‐alive counts. We concluded that mark‐resight performed better and was a viable approach for monitoring this small elk population and possibly others with similar characteristics (i.e., small population and landscape scales), but this approach requires periodic marking of elk; we estimated mark‐resight costs would be about 40% greater than sightability model application costs. The utility of sightability‐correction modeling was limited by a high proportion of groups with low detectability on our densely forested landscape. © 2012 The Wildlife Society.

Regression calibration with more surrogates than mismeasured variables

In a recent paper (Weller EA, Milton DK, Eisen EA, Spiegelman D. Regression calibration for logistic regression with multiple surrogates for one exposure. Journal of Statistical Planning and Inference 2007; 137: 449-461), the authors discussed fitting logistic regression models when a scalar main explanatory variable is measured with error by several surrogates, that is, a situation with more surrogates than variables measured with error. They compared two methods of adjusting for measurement error using a regression calibration approximate model as if it were exact. One is the standard regression calibration approach consisting of substituting an estimated conditional expectation of the true covariate given observed data in the logistic regression. The other is a novel two-stage approach when the logistic regression is fitted to multiple surrogates, and then a linear combination of estimated slopes is formed as the estimate of interest. Applying estimated asymptotic variances for both methods in a single data set with some sensitivity analysis, the authors asserted superiority of their two-stage approach. We investigate this claim in some detail. A troubling aspect of the proposed two-stage method is that, unlike standard regression calibration and a natural form of maximum likelihood, the resulting estimates are not invariant to reparameterization of nuisance parameters in the model. We show, however, that, under the regression calibration approximation, the two-stage method is asymptotically equivalent to a maximum likelihood formulation, and is therefore in theory superior to standard regression calibration. However, our extensive finite-sample simulations in the practically important parameter space where the regression calibration model provides a good approximation failed to uncover such superiority of the two-stage method. We also discuss extensions to different data structures.

Association of Inflammatory and Noninflammatory Breast Cancer with Socioeconomic Characteristics in the Surveillance, Epidemiology, and End Results Database, 2000–2007

Inflammatory breast cancer (IBC) is a rare and highly aggressive form of primary breast cancer. Little is known about the risk factors for IBC, specifically the association with socioeconomic position (SEP). The association between breast cancer type (IBC vs. non-IBC) with county-level SEP in the Surveillance, Epidemiology, and End Results database for cases diagnosed from 2000 to 2007 was examined. County-level SEP characteristics included metropolitan versus non-metropolitan residence, percentage below the poverty level, percentage less than high-school graduate, and an index combining the poverty and high-school variables. IBC and non-IBC age-adjusted incidence rates were calculated, stratified on SEP and race/ethnicity. The odds of IBC versus non-IBC given a particular SEP characteristic, adjusting for age and race/ethnicity, was examined through fitting of hierarchical logistic regression models (HLM). Incidence rates for IBC generally increased as SEP decreased, whereas the opposite was found for non-IBC. HLM results showed that low SEP is associated with higher odds of IBC: highest (≥ 20%) versus lowest (<10%) persons below the poverty level [OR (95% confidence interval, CI) = 1.25 (1.09-1.43)]; highest (>28.76%) versus lowest (≤ 15.99%) persons less than high-school graduate [OR (95% CI) = 1.25 (1.10-1.42)]; and low SEP as measured by poverty-high school index versus high SEP [OR (95% CI)= 1.26 (1.11-1.44)]. Overall breast cancer has been found to be positively associated with SEP, whereas in this analysis, IBC was associated with decreasing SEP. Studies focused on understanding the disparity in IBC incidence, as well as interventions to eliminate these differences are needed.

베이지안 로지스틱 회귀모형에서의 추론에 대한 연구

기존의 frequentist 추론에 비해 Bayesian 추론에서의 가설 검정 및 모형 선택 문제는 학자들 간에 일치된 견해를 보이지 못하고 있으며 아직도 논란이 되는 것들이 많다. Bayesian 추론에서 가설 검정 및 모형 선택의 기준으로 널리 쓰이는 Bayes factor는 이해하기 쉬우나 여러 경우에 구하기 어려운 단점이 존재한다. 그 외에 다른 기준으로 Spiegelhalter 등 (2002)가 제시한 DIC(Deviance Information Criterion)과 frequentist 추론에서의 P-value에 대비되는 Bayesian P-value가 있다. 본 논문에서는 Swiss banknote 자료를 Bayesian 로지스틱 회귀모형으로 분석하고 관련 기준들을 구하여 각 기준들이 일관성 있는 결론을 보이는지 확인하고자 한다. Model selection and hypothesis testing problems in Bayesian inference are still debated between scholars. Bayesian factors traditionally used as a criterion in Bayesian hypothesis testing and model selection, are easy to understand but sometimes hard to compute. In addition, there are other model selection criterions such as DIC(Deviance Information Criterion) by Spiegelhalter et al. (2002) and Bayesian P-values for testing. In this paper, we briefly introduce the Bayesian hypothesis testing and model selection procedure. In addition we have applied a Bayesian inference to Swiss banknote data by a fitting logistic regression model and computing several test statistics to see if they provide consistent results.

Olfactory response of Typhlodromus pyri (Acari: Phytoseiidae) to synthetic methyl salicylate in laboratory bioassays

The herbivore-induced plant volatile (HIPV) methyl salicylate (MeSA) is widely present in the chemical profile of several plant species and is known to attract natural enemies, including predatory mites. In this study, the response of Typhlodromus pyri, a key predator of pest mites in west coast vineyards, to synthetically produced MeSA was tested using a Y-tube olfactometer in laboratory bioassays. Six doses ranging from 0.002 to 200 μg of MeSA diluted in 0.1 ml hexane were tested. Significantly higher proportions of T. pyri preferred MeSA at doses 0.02, 0.2 and 20 μg. No differences in response to MeSA were detected at the highest (200 μg), intermediate (2 μg) and lowest (0.002 μg) doses. Mite response to MeSA was a function of dose when fitting polynomial and logistic regression models using dose and square of the log dose prediction factors. Results indicate that synthetic MeSA may be applied to attract predatory arthropod populations in vineyards to enhance biological control of pest mites.

Determinantes de la prescripción de antibióticos en atención primaria

To determine patient and physician-related factors associated with variability in antibiotic prescription. Observational study of the prevalence of antibacterial medication prescription >14 years old. official prescriptions, clinical histories and individual health cards. Patient-related variables were: age, sex, number of medical visits-year, comorbidity, antibacterials dispensed with prescription. Physician-related variables were age, sex, number of patients assigned, place of work and rurality. Variables associated with prescription were studied by estimating the odds ratio (OR) from the fit of the multilevel logistic regression models. The rate of antibiotic prescription-year in the population was 31.4%. Factors associated with prescription were high rate of visits (users with more than 5 annual visits multiply the probability of receiving antibiotics, compared to those who madeno visits: OR=10.8), age (non-linearly, with a greater likelihood in the young and the elderly) and sex, with a higher rate in women (OR=1.5). No association was found between prescription and age and sex of the physician, but an association was found with workload: the higher the physician's workload, the higher the likelihood of antibiotic prescription. The most important factor associated with the increase in prescription rate was the frequency of visits. In addition, women, the young and the elderly receive more antibiotics. A multi-factor intervention focusing on demand, patients, and physicians should be carried out to reduce prescription rates.

Evaluation of graphical diagnostics for assessing goodness of fit of logistic regression models

The aim of the current work was to evaluate graphical diagnostics for assessment of the fit of logistic regression models. Assessment of goodness of fit of a model to the data set is essential to ensure the model provides an acceptable description of the binary variables seen. For logistic regression the most common diagnostic used for this purpose is binning the data and comparing the empirical probability of the occurrence of a dependent variable with the model predicted probability against the mean covariate value in the bin. Although intuitively appealing this method, which we term simple binning, may not have consistent properties for diagnosing model problems. In this report we describe and evaluate two different diagnostic procedures, random binning and simplified Bayes marginal model plots. These procedures were assessed via simulation under three different designs. Design 1: studies which were balanced on binary variables and a continuous covariate. Design 2: studies that were balanced on binary variables but unbalanced on the continuous covariate. Design 3: studies that were unbalanced on both the binary variables and the covariate. Each simulated study consisted of 500 individuals. Thirty studies were simulated. The covariate of interest was dose which could range from 0 to 20 units. The data were simulated with the dose being related to the outcome according to an E (max) model on the logit scale. A logit E (max) model (correct model) and a logit linear model (wrong model) were fitted to all data sets. The performance of the above diagnostics, in addition to simple binning, was compared. For all designs the proposed diagnostics performed at least as well and in many instances better than simple binning. In case of design 1 random binning and simple binning are identical. In the case of designs 2 and 3 random binning and simplified Bayes marginal model plots were superior in assessing the model fit when compared to simple binning. For the examples tested, both random binning and simplified Bayesian marginal model plots performed acceptably.

Determinantes da experimentação do cigarro e do início precoce do tabagismo entre adolescentes escolares em Salvador (BA)

OBJETIVO: Identificar os fatores associados à experimentação do cigarro e ao início precoce do tabagismo por adolescentes escolares (11-14 anos). MÉTODOS: Estudo transversal de base populacional com adolescentes escolares (ensino fundamental e médio) de escolas públicas e particulares de Salvador, Bahia. Foi utilizado um protocolo de amostragem em múltiplos estágios. Os participantes da pesquisa responderam a um questionário anônimo, de preenchimento voluntário e autoaplicável. As variáveis dependentes foram o uso experimental do cigarro e o início precoce do tabagismo. Variáveis sociodemográficas, educacionais, relacionadas à influência da mídia e ao consumo de álcool foram consideradas variáveis independentes. Foi realizada a análise de regressão logística multivariada, seguida da análise da adequação do modelo de regressão logística final. RESULTADOS: A prevalência de experimentação do cigarro foi de 16,1% (IC95%: 15,1-17,2). Dos 5.347 alunos avaliados, 757 experimentaram cigarros, e 462 (61,2%) desses 757 o fizeram antes dos 15 anos de idade (média de idade de experimentação de cigarros, 13,9 ± 1,8 anos). O início precoce do tabagismo associou-se com o tabagismo paterno (OR = 1,53; IC83%: 1,07-2,19) e com a influência da mídia (OR = 2,43; IC83%: 1,46-4,04). A experimentação de cigarros associou-se com o consumo do álcool (OR = 6,04; IC83%: 4,62-7,88), o tabagismo de amigos (OR = 3,18; IC83%: 2,54-3,99) e de namorados (OR = 3,42; IC83%: 1,78-6,56), assim como a intervenção educacional agressiva/coercitiva dos pais (OR = 1,60; IC83%: 1,26-2,04). CONCLUSÕES: O consumo do álcool e a influência de pessoas da convivência do jovem apresentaram uma forte associação com a experimentação de cigarros. O tabagismo paterno e a influência da mídia para os produtos do tabaco aumentaram as chances de consumo desta substância.