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.

Refolded dengue virus type 2 NS1 protein expressed in Escherichia coli preserves structural and immunological properties of the native protein

The dengue virus NS1 protein has been shown to be a protective antigen under different experimental conditions but the recombinant protein produced in bacterial expression systems is usually not soluble and loses structural and immunological features of the native viral protein In the present study, experimental conditions leading to purification and refolding of the recombinant dengue virus type 2 (DENV-2) NS1 protein expressed in Escherichia coil are described The refolded recombinant protein was recovered as heat-stable soluble dimers with preserved structural features, as demonstrated by spectroscopic methods In addition, antibodies against epitopes of the NS1 protein expressed in eukaryotic cells recognized the refolded protein expressed in E coli but not the denatured form or the same protein submitted to a different refolding condition Collectively, the results demonstrate that the recombinant NS1 protein preserved important conformation and antigenic determinants of the native virus protein and represents a valuable reagent either for the development of vaccines or for diagnostic methods. (C) 2010 Elsevier B V All rights reserved

Local Influence Under Parameter Constraints

Calculations of local influence curvatures and leverage have been well developed when the parameters are unrestricted. In this article, we discuss the assessment of local influence and leverage under linear equality parameter constraints with extensions to inequality constraints. Using a penalized quadratic function we express the normal curvature of local influence for arbitrary perturbation schemes and the generalized leverage matrix in interpretable forms, which depend on restricted and unrestricted components. The results are quite general and can be applied in various statistical models. In particular, we derive the normal curvature under three useful perturbation schemes for generalized linear models. Four illustrative examples are analyzed by the methodology developed in the article.

An R implementation for generalized Birnbaum-Saunders distributions

The Birnbaum-Saunders (BS) model is a positively skewed statistical distribution that has received great attention in recent decades. A generalized version of this model was derived based on symmetrical distributions in the real line named the generalized BS (GBS) distribution. The R package named gbs was developed to analyze data from GBS models. This package contains probabilistic and reliability indicators and random number generators from GBS distributions. Parameter estimates for censored and uncensored data can also be obtained by means of likelihood methods from the gbs package. Goodness-of-fit and diagnostic methods were also implemented in this package in order to check the suitability of the GBS models. in this article, the capabilities and features of the gbs package are illustrated by using simulated and real data sets. Shape and reliability analyses for GBS models are presented. A simulation study for evaluating the quality and sensitivity of the estimation method developed in the package is provided and discussed. (C) 2008 Elsevier B.V. All rights reserved.

Systematic risk estimation in symmetric models

The aim of this article is to discuss the estimation of the systematic risk in capital asset pricing models with heavy-tailed error distributions to explain the asset returns. Diagnostic methods for assessing departures from the model assumptions as well as the influence of observations on the parameter estimates are also presented. It may be shown that outlying observations are down weighted in the maximum likelihood equations of linear models with heavy-tailed error distributions, such as Student-t, power exponential, logistic II, so on. This robustness aspect may also be extended to influential observations. An application in which the systematic risk estimate of Microsoft is compared under normal and heavy-tailed errors is presented for illustration.

Diagnostic tool to analyse colour-magnitude diagrams of poorly populated stellar concentrations

The dynamical processes that lead to open cluster disruption cause its mass to decrease. To investigate such processes from the observational point of view, it is important to identify open cluster remnants (OCRs), which are intrinsically poorly populated. Due to their nature, distinguishing them from field star fluctuations is still an unresolved issue. In this work, we developed a statistical diagnostic tool to distinguish poorly populated star concentrations from background field fluctuations. We use 2MASS photometry to explore one of the conditions required for a stellar group to be a physical group: to produce distinct sequences in a colour-magnitude diagram (CMD). We use automated tools to (i) derive the limiting radius; (ii) decontaminate the field and assign membership probabilities; (iii) fit isochrones; and (iv) compare object and field CMDs, considering the isochrone solution, in order to verify the similarity. If the object cannot be statistically considered as a field fluctuation, we derive its probable age, distance modulus, reddening and uncertainties in a self-consistent way. As a test, we apply the tool to open clusters and comparison fields. Finally, we study the OCR candidates DoDz 6, NGC 272, ESO 435 SC48 and ESO 325 SC15. The tool is optimized to treat these low-statistic objects and to separate the best OCR candidates for studies on kinematics and chemical composition. The study of the possible OCRs will certainly provide a deep understanding of OCR properties and constraints for theoretical models, including insights into the evolution of open clusters and dissolution rates.

Research on Candida dubliniensis in a Brazilian yeast collection obtained from cardiac transplant, tuberculosis, and HIV-positive patients, and evaluation of phenotypic tests using agar screening methods

The aim of this study was to research Candida dubliniensis among isolates present in a Brazilian yeast collection and to evaluate the main phenotypic methods for discrimination between C. albicans and C. dubliniensis from oral cavity. A total of 200 isolates, presumptively identified as C. albicans or C. dubliniensis obtained from heart transplant patients under immunosuppressive therapy, tuberculosis patients under antibiotic therapy, HIV-positive patients under antiretroviral therapy, and healthy subjects, were analyzed using the following phenotypic tests: formation and structural arrangement of chlamydospores on corn meal agar, casein agar, tobacco agar, and sunflower seed agar; growth at 45 degrees C; and germ tube formation. All strains were analyzed by polymerase chain reaction (PCR). In a preliminary screen for C. dubliniensis, 48 of the 200 isolates on corn meal agar, 30 of the 200 on casein agar, 16 of the 200 on tobacco agar, and 15 of the 200 on sunflower seed agar produced chlamydoconidia; 27 of the 200 isolates showed no or poor growth at 45 degrees C. All isolates were positive for germ tube formation. These isolates were considered suggestive of C. dubliniensis. All of them were subjected to PCR analysis using C. dubliniensis-specific primers. C. dubliniensis isolates were not found. C. dubliniensis isolates were not recovered in this study done with immunocompromised patients. Sunflower seed agar was the medium with the smallest number of isolates of C. albicans suggestive of C. dubliniensis. None of the phenotypic methods was 100% effective for discrimination between C. albicans and C. dubliniensis. (C) 2011 Elsevier Inc. All rights reserved.

Cathepsin L-like genes of Trypanosoma vivax from Africa and South America - characterization, relationships and diagnostic implications

We characterized sequences from genes encoding cathepsin L-like (CatL-like) cysteine proteases from African and South American isolates of Trypanosoma vivax and T. vivax-like organisms, and evaluated their suitability as genetic markers for population structure analysis and diagnosis. Phylogenetic analysis of sequences corresponding to CatL-like catalytic domains revealed substantial polymorphism, and clades of sequences (TviCatL1-9) were separated by large genetic distances. TviCatL1-4 sequences were from cattle isolates from West Africa (Nigeria and Burkina Faso) and South America (Brazil and Venezuela), which belonged to the same T. vivax genotype. T. vivax-like genotypes from East Africa showed divergent sequences, including TviCatL5-7 for isolates from Mozambique and TviCatL8-9 for an isolate from Kenya. Phylogenetic analysis of CatL-like gene data supported the relationships among trypanosome species reflected in the phylogenies based on the analysis of small subunit (SSU) of ribosomal RNA gene sequence data. The discovery of different CatL-like sequences for each genotype, defined previously by ribosomal DNA data, indicate that these sequences provide useful targets for epidemiological and population genetic studies. Regions in CatL-like sequences shared by all T. vivax genotypes but not by other trypanosomes allowed the establishment of a specific and sensitive diagnostic PCR for epidemiological studies in South America and Africa. (C) 2008 Elsevier Ltd. All rights reserved.

Bayesian Estimation for Performance Measures of Two Diagnostic Tests in the Presence of Verification Bias

Sensitivity and specificity are measures that allow us to evaluate the performance of a diagnostic test. In practice, it is common to have situations where a proportion of selected individuals cannot have the real state of the disease verified, since the verification could be an invasive procedure, as occurs with biopsy. This happens, as a special case, in the diagnosis of prostate cancer, or in any other situation related to risks, that is, not practicable, nor ethical, or in situations with high cost. For this case, it is common to use diagnostic tests based only on the information of verified individuals. This procedure can lead to biased results or workup bias. In this paper, we introduce a Bayesian approach to estimate the sensitivity and the specificity for two diagnostic tests considering verified and unverified individuals, a result that generalizes the usual situation based on only one diagnostic test.

Diagnostic reference levels for the most frequent radiological examinations carried out in Brazil

Objectives. A large-scale survey of doses to patients undergoing the most frequent radiological examinations was carried out in health services in Sao Paulo (347 radiological examinations per 1 000 inhabitants), the most populous Brazilian state. Methods. A postal dosimetric kit with thermoluminescence dosimeters was used to evaluate the entrance surface dose (ESD) to patients. A stratified sampling technique applied to the national health database furnished important data on the distribution of equipment and the annual number of examinations. Chest, head (skull and sinus), and spine (cervical, thoracic, and lumbar) examinations were included in the trial. A total of 83 rooms and 868 patients were included, and 1 415 values of ESD were measured. Results. The data show large coefficients of variation in tube charge, giving rise to large variations in ESD values. Also, a series of high ESD values associated with unnecessary localizing fluoroscopy were detected. Diagnostic reference levels were determined, based on the 75th percentile (third quartile) of the ESD distributions. For adult patients, the diagnostic reference levels achieved are very similar to those obtained in international surveys. However, the situation is different for pediatric patients: the ESD values found in this survey are twice as large as the international recommendations for chest radiographs of children. Conclusions. Despite the reduced number of ESD values and rooms for the pediatric patient group, it is recommended that practices in chest examinations be revised and that specific national reference doses and image quality be established after a broader survey is carried out.