Prediction of mortality in Intensive Care Units: a multivariate feature selection.

CONTEXT: The critical nature of patients in Intensive Care Units (ICUs) demands intensive monitoring of their vital signs as well as highly qualified professional assistance. The combination of these needs makes ICUs very expensive, which requires investment to be prioritized. Administrative issues emerge, and health institutions face dilemmas such as: "How many beds should an ICU provide to serve the population, at the lowest costs" and "Which is the most critical body information to monitor in an ICU?". Due to financial and ethical implications, these judgments require technical and precise knowledge. Decisions have usually relied on clinical scores, like the APACHE (Acute Physiology And Chronic Health Evaluation) and SOFA (Sequential Organ Failure Assessment) scores, which are imprecise and outdated. The popularization of machine learning techniques has shed some light on the topic as a way to renew score purposes. In 2012, the PhysioNet/Computing in Cardiology launched the Challenge - ICU Patients. This Challenge aimed to stimulate the development of techniques to predict mortality in ICUs. Based on biometric and physiological features collected from patients, the participants predicted the patient's death risk by using their classifiers. Several participants achieved results that were better than the results produced by the SOFA and the APACHE scores; the prediction levels were ≈54%, which is weak. OBJECTIVES: Here, we investigate the reasons that led to these results as a means to ground our solution. Then, we propose alternative practices in an attempt to improve the results. Our main goal is to improve the prediction of mortality in ICUs by using the same data employed during the 2012 PhysioNet Challenge. Our specific objectives are (i) to simplify the problem by reducing the dimensionality; (ii) to reduce the uncontrolled variance, and (iii) to make classifiers less dependent on the training set. METHODS: Accordingly, we propose a methodology based on extensive steps, including sample filter and data normalization. To select features and to reduce the intra-group variance, we employ multivariate data analysis by using Principal Component Analysis, Factor Analysis, Spectral Clustering, and Tukey's HSD Test, recursively. After that, we use machine learning techniques to create classifiers according to different methods. We evaluate our results with the same metrics proposed by the 2012 PhysioNet Challenge. RESULTS: For classifiers constructed and tested by using independent datasets, our best classifier was a linear SVM, which provided results of ≈0.73. These results were significantly better than the ≈0.54 achieved in previous work at >99% confidence interval. Furthermore, our approach only demanded twelve features, which was consistently smaller than the number of features required by the previous approaches. CONCLUSION: Our results indicated that our approach presented: (a) higher performance to predict death risks (+20%); (b) smaller dependence on the training set; and (c) lower costs for ICU monitoring (few features). Besides the better prediction power, our approach also demanded lower costs for implementation and a more extensive range of potential ICUs. Future studies should employ our proposal to investigate the possibility of including some physiological features that were not available for the 2012 PhysioNet Challenge.

Vegetation Pattern Modulates Ground Arthropod Diversity in Semi-Arid Mediterranean Steppes.

The ecological functioning of dryland ecosystems is closely related to the spatial pattern of the vegetation, which is typically structured in patches. Ground arthropods mediate key soil functions and ecological processes, yet little is known about the influence of dryland vegetation pattern on their abundance and diversity. Here, we investigate how patch size and cover, and distance between patches relate to the abundance and diversity of meso-and microarthropods in semi-arid steppes. We found that species richness and abundance of ground arthropods exponentially increase with vegetation cover, patch size, and patch closeness. The communities under vegetation patches mainly respond to patch size, while the communities in the bare-soil interpatches are mostly controlled by the average distance between patches, independently of the concurrent changes in vegetation cover. Large patches seem to play a critical role as reserve and source of ground arthropod diversity. Our results suggest that decreasing vegetation cover and/or changes in vegetation pattern towards small and over-dispersed vegetation patches can fast lead to a significant loss of ground arthropods diversity in drylands.

Effect of fertilisation and harvest period on polar metabolites of Calendula oficcinalis

The present study evaluated the chemical profile of polar extracts of Calendula officinalis L., Asteraceae, that were grown under different cultivation conditions: chemical fertilisation, organic fertilisation and mulching. Furthermore, we investigated metabolite variations during plant development by comparing the metabolites from harvested plants at 60 and 120 days after planting. We used HPLC-DAD-MS/MS to tentatively identify metabolites. In total, we identified seven known compounds: five flavonoid glycosides and two caffeoylquinic acids derivatives. There were no statistically significant differences in the expression of metabolites from plants grown under the examined soil treatments. However, five substances varied according to harvest time, suggesting that the biosynthesis of polar metabolites of Calendula officinalis is not affected by changes in soil composition. Therefore, this plant could represent a source for phytomedicines with a constant content of polar metabolites.

A new HPLC-DAD-MS/MS method for the simultaneous determination of major compounds in the crude extract of Lychnophora salicifolia Mart. (Brazilian arnicão) leaves: application to chemical variability evaluation.

Lychnophora salicifolia Mart., which occurs in the Brazilian Cerrado in the states of Bahia and Minas Gerais as well as in the southeast of the state of Goiás, is the most widely distributed and also the most polymorphic species of the genus. This plant is popularly known to have anti-inflammatory and analgesic activities. In this work, we have studied the variation in terms of polar metabolites of ninety-three Lychnophora salicifolia Mart. specimens collected from different regions of the Brazilian Cerrado. Identification of the constituents of this mixture was carried out by analysis of the UV spectra and MS data after chromatographic separation. Twenty substances were identified, including chlorogenic acid derivatives, a flavonoid C-glucoside, and other sesquiterpenes. The analytical method was validated, and the reliability and credibility of the results was ensured for the purposes of this study. The concentration range required for analysis of content variability within the analyzed group of specimens was covered with appropriate values of limits of detection and quantitation, as well as satisfactory precision and recovery. A quantitative variability was observed among specimens collected from the same location, but on average they were similar from a chemical viewpoint. In relation to the study involving specimens from different locations, there were both qualitative and quantitative differences among plants collected from different regions of Brazil. Statistical analysis revealed that there is a correlation between geographical localization and polar metabolites profile for specimens collected from different locations. This is evidence that the pattern of metabolites concentration depends on the geographical distribution of the specimens.