RESUMO
ObjectiveTo provide a thorough comparative study among state-of-the-art machine learning methods and statistical methods for determining in-hospital mortality in COVID-19 patients using data upon hospital admission; to study the reliability of the predictions of the most effective methods by correlating the probability of the outcome and the accuracy of the methods; to investigate how explainable are the predictions produced by the most effective methods. Materials and MethodsDe-identified data were obtained from COVID-19 positive patients in 36 participating hospitals, from March 1 to September 30, 2020. Demographic, comorbidity, clinical presentation and laboratory data were used as training data to develop COVID-19 mortality prediction models. Multiple machine learning and traditional statistics models were trained on this prediction task using a folded cross-validation procedure, from which we assessed performance and interpretability metrics. ResultsThe Stacking of machine learning models improved over the previous state-of-the-art results by more than 26% in predicting the class of interest (death), achieving 87.1% of AUROC and macro F1 of 73.9%. We also show that some machine learning models can be very interpretable and reliable, yielding more accurate predictions while providing a good explanation for the why. ConclusionThe best results were obtained using the meta-learning ensemble model - Stacking. State-of the art explainability techniques such as SHAP-values can be used to draw useful insights into the patterns learned by machine-learning algorithms. Machine-learning models can be more explainable than traditional statistics models while also yielding highly reliable predictions.
RESUMO
The SARS-CoV-2 pandemic have been affecting millions of people worldwide, since the beginning of 2020. COVID-19 can cause a wide range of clinical symptoms, which varies from asymptomatic presentation to severe respiratory insufficiency, exacerbation of immune response, disseminated microthrombosis and multiple organ failure, which may lead to dead. Due to the rapid spread of SARS-CoV-2, the development of vaccines to minimize COVID-19 severity in the world population is imperious. One of the employed techniques to produce vaccines against emerging viruses is the synthesis of recombinant proteins, which can be used as immunizing agents. Based on the exposed, the aim of the present study was to verify the systemic and immunological effects of IM administration of recombinant Nucleocapsid protein (NP), derived from SARS-CoV-2 and produced by this research group, in 2 different strains of rats (Rattus norvegicus); Wistar and Lewis. For this purpose, experimental animals received 4 injections of NP, once a week, and were submitted to biochemical and histological analysis. Our results showed that NP inoculations were safe for the animals, which presented no clinical symptoms of worrying side effects, nor laboratorial alterations in the main biochemical and histological parameters, suggesting the absence of toxicity induced by NP. Moreover, NP injections successfully triggered the production of specific anti-SARS-CoV-2 IgG antibodies by both Wistar and Lewis rats, showing the sensitization to have been well sufficient for the immunization of these strains of rats. Additionally, we observed the local lung activation of the Bronchus-Associated Lymphoid Tissue (BALT) of rats in the NP groups, suggesting that NP elicits specific lung immune response. Although pre-clinical and clinical studies are still required, our data support the recombinant NP produced by this research group as a potential immunizing agent for massive vaccination, and may represent advantages upon other recombinant proteins, since it seems to induce specific pulmonary protection.
RESUMO
ObjectiveTo evaluate the perception of midwifery interns regarding hospital practices during COVID-19. Material and methodsStudy of qualitative approach, of phenomenological design, where 80 midwifery interns from the different regions of Peru participated, who are also representatives of their hospital headquarters. An in-depth interview was applied where the perception of hospital practices was addressed according to: i) current problems and ii) solution proposals. ResultsMidwifery interns have been removed from hospital practices, mainly due to the absence of personal protective equipment and health insurance, financially affecting those who must continue to make rent and food payments; Likewise, a large part of the universities have not offered proposals for solutions to the delay in internships, raising concerns about delays in administrative procedures, even more so for students from non-licensed universities. Among the proposals, those who are close to graduating suggest being exempted from the months when there were no activities, so as not to delay future processes such as tuition and rural service; likewise, suspend payment for these months and strengthen knowledge through the discussion of clinical cases, which could be virtual. ConclusionsThe cessation of hospital practice responds to a lack of guarantees in the health care of the student, generating economic repercussions and a negative perception regarding university management. Finally, solutions that could be considered for the next decisions made by the institutions are reported.
RESUMO
In this work, we proposed a variant of the SIR model, taking as based on models used to describe the epidemic outbreak in South Korea and Portugal, to study the SARS-CoV-2 epidemic curve in Brazil. The model presented here describes with reasonable agreement the number of COVID-19 cases registered in Brazil between February 26 and April 25, 2020 based on the hypothesis that there a large number no notified cases (11 to 1) and variation in contagion rate according to social isolation measures and greater or lesser exposure to the virus (highest rate in beginning from epidemic). To this end, we introduced an exposure factor, called {beta}1/{beta}2, which allows us to describe the influence of factors such as social isolation on dispersal from disease. The results also corroborate a phenomenon observed in countries that registered a high growth in cases in short period of time, to example of Italy, Spain and USA: if isolation measures are imposed late, the total number of cases explodes when the epidemic is approaching from peak, which implies a higher exposure rate in the first days of case registration. The model also predicts that the peak epidemic outbreak in Brazil, based on the number of cases, will occur around May 20, 2020.
RESUMO
Bandejas com diferente número de células (128, 200 e 288) e dois substratos (fibra de casca de coco e casca de pinus) foram avaliados para a produção de mudas de tomilho e sobre o desempenho das plantas em sistema hidropônico. Para ambos os experimentos, o delineamento foi em blocos ao acaso. As mudas provenientes de estaquia foram avaliadas aos 30 dias quanto ao comprimento e massa seca de raízes. Aos 40 dias após o transplante das mudas, as plantas cultivadas em sistema hidropônico foram avaliadas quanto à massa fresca da parte aérea, massa seca da parte aérea e massa seca das raízes. A maior massa seca de raízes foi obtida para as mudas cultivadas em substrato à base de casca de fibra de coco. As mudas com maior comprimento de raízes foram obtidas nas bandejas com 128 e 200 células. As plantas de tomilho provenientes de mudas produzidas em bandejas de 128 células apresentaram maior massa fresca da parte aérea, massa seca da parte aérea e massa seca das raízes.
Trays with different number of cells (128, 200 and 288) and two substrates (coir and pine bark-based) were studied for thyme seedling production and plant development in hydroponic system. In both experiments, experimental design was in randomized blocks. Seedlings from cuttings were evaluated at 30 days as to root length and dry matter. At 40 days after transplanting, plants grown in hydroponic system were evaluated for shoot fresh matter, shoot dry matter and root dry matter. The highest root dry matter was detected in seedlings grown in coir-based substrate. Seedlings that presented the greatest root length were obtained by using trays with 128 and 200 cells. Thyme plants from seedlings produced in trays with 128 cells had higher shoot fresh matter, shoot dry matter and root dry matter.