Evaluating How the Social Restriction, the Government Response, the Health, and Economic Indices Affected the Prediction of the Number of Deaths Provoked by COVID-19 in Brazil Using Classical Statistical and Machine Learning Models

The COVID-19 death predictions are helpful for the formulation of public policies, allowing the use of more effective social isolation strategies with less economic and social impact. This article evaluates a wide range of forecasting methods to identify the best models for predicting cumulative and daily deaths caused by COVID-19 in Brazil, considering a forecast for a seven-day horizon. With the seven-day horizon, the predictions have more accuracy. The dataset is from Oxford Covid-19 Government Response Tracker. The jackknife resampling technique was implemented, thus providing an accurate estimate for evaluating the predictive capacity of the models. Each model was fitted with 266 jackknife samples considering 30-day training bases. The comparison between predictions was made using the average results, considering R-2, MAPE, RMSE, and MAE. Models from different classes were adopted: 1 ETS, 4 ARIMA, 18 regression models, and 7 machine learning algorithms. The cumulative death models produce better results than daily deaths, as the cumulative death models are less influenced by time series components: cycle and seasonality. The best results for predicting daily deaths were attained by the Ridge regression method. The best results for predicting cumulative deaths were obtained by the Cubist regression method.

Evaluating How the Social Restriction, the Government Response, the Health, and Economic Indices Affected the Prediction of the Number of Deaths Provoked by COVID-19 in Brazil Using Classical Statistical and Machine Learning Models

The COVID-19 death predictions are helpful for the formulation of public policies, allowing the use of more effective social isolation strategies with less economic and social impact. This article evaluates a wide range of forecasting methods to identify the best models for predicting cumulative and daily deaths caused by COVID-19 in Brazil, considering a forecast for a seven-day horizon. With the seven-day horizon, the predictions have more accuracy. The dataset is from Oxford Covid-19 Government Response Tracker. The jackknife resampling technique was implemented, thus providing an accurate estimate for evaluating the predictive capacity of the models. Each model was fitted with 266 jackknife samples considering 30-day training bases. The comparison between predictions was made using the average results, considering R2, MAPE, RMSE, and MAE. Models from different classes were adopted: 1 ETS, 4 ARIMA, 18 regression models, and 7 machine learning algorithms. The cumulative death models produce better results than daily deaths, as the cumulative death models are less influenced by time series components: cycle and seasonality. The best results for predicting daily deaths were attained by the Ridge regression method. The best results for predicting cumulative deaths were obtained by the Cubist regression method © 2022 by the authors. Submitted for possible open access publication under the terms and conditions of the Creative Commons Attribution (CC BY NC) license (https://creativecommons.org/licenses/by-nc/4.0/).

UTILIZAÇÃO DE HEMOCOMPONENTES EM PESSOAS ACOMETIDAS PELA COVID-19, EM UM HOSPITAL UNIVERSITÁRIO DO ESTADO DO RIO DE JANEIRO

Objetivos: Diante de uma doença viral desconhecida e de alto impacto sobre o sistema de saúde, este estudo visou analisar o perfil transfusional em portadores da COVID-19, no Hospital Universitário Pedro Ernesto- UERJ, referência para atendimento de casos graves. Material e métodos: Trata-se de estudo descritivo, transversal, observacional, realizado entre abril a julho/2020 (primeira onda) e janeiro a junho/2021 (segunda onda). Foram incluídos todos os casos de COVID-19, internados no HUPE, tanto em unidade fechada quanto em enfermarias. Adultos e crianças. As variáveis analisadas foram tipo de hemocomponentes, indicações, comorbidades, suporte ventilatório e hemodiálise, tipagem sanguínea e reação transfusional. A partir de revisão de prontuário eletrônico e consulta ao banco de dados do serviço de Hemoterapia do HUPE, foram obtidas as frequências simples das variáveis estudadas. O estudo foi aprovado pela CEP do HUPE, sob o CAAE 31421620.7.1001.5259. Resultados: Foram analisados 560 e 328 pacientes na primeira e segunda onda, respectivamente. Destes 19,7% (n = 110) e 22% (n = 72) receberam transfusão nos respectivos grupos. Na primeira e segunda ondas, concentrado de hemácias (CH) foi o hemocomponente mais utilizado (80% e 73%) e plasma fresco congelado (PFC) o segundo mais frequente (26,3% e 10,6%). Anemia e sangramento foram as principais causas para indicação de transfusão. Na primeira onda, 32 pacientes transfundidos tinham HAS e DM e os demais apresentavam estas isoladas, ou associadas a outras comorbidades. Ventilação mecânica e hemodiálise simultâneas, foram observadas em 36 dos pacientes transfundidos (32,7%) na primeira onda onde predominaram os tipos sanguíneos O (n = 45;40,9%) e A (n = 40;36,3%). Na segunda onda, a média de idade foi de 57,7 anos, sendo HAS e DM as comorbidades mais prevalentes. Não houve notificação de reação transfusional em nenhum dos grupos. Discussão: Tem sido observado uma “síndrome anêmica” em pacientes com COVID-19 sendo relatado em diversos estudos, a necessidade de reposição de CH. O uso de PFC, foi utilizado para eventos hemorrágicos ou na vigência de discrasia sanguínea, porém na primeira onda, havia um estudo sobre uso de plasma convalescente no HUPE, o que pode ter sido um viés na quantificação do uso de plasma, nesta fase. A não realização de busca ativa nas unidades COVID-19, e a não notificação ativa pelos prescritores, pode justificar a ausência de notificação de reação transfusional. Conclusão: A indicação de transfusão na COVID-19, tem sido baseada nas mesmas indicações para pacientes críticos. São necessários estudos analíticos para a construção de conhecimento transfusional na COVID-19, comparando-a com outros grupos de pacientes graves.

COVID-19 Lockdown and the Behavior Change on Physical Exercise, Pain and Psychological Well-Being: An International Multicentric Study

OBJECTIVE: To evaluate the impact of the COVID-19 pandemic and the following lockdown on physical exercise (PEx) practice, pain, and psychological well-being. METHODS: A cross-sectional multicentric study was performed using a nonrandom convenience sampling from the general population (>=18 years-old) of 6 countries (Brazil, Italy, France, Portugal, Germany, and Spain) adopting social isolation (SI). The validated self-administered online survey (PEF-COVID19) was used. The tests T-test and Chi-square with Bonferroni correction were used for statistical analysis and a multivariate logistic regression model (p < 0.05). RESULTS: We included 3194 replies and ~80% of the respondents were in SI. Brazilian sample was highly influenced by the pandemic considering PEx practice and habits, pain, anxiety, and stress (p < 0.05). Among the European countries, Italy presented the major changes. The model to predict the non-practice of PEx during SI showed that the variables countries, smoking, SI, and PEx level were significant predictors (p < 0.001). CONCLUSION: The pandemic changed the PEx practice and habits, and the psychological well-being of populations in different manners. Countries, smoking, SI, and PEx level were predictors for the non-practice of PEx. Public health strategies are suggested to avoid sedentary lifestyles and quality of life decrease.