Matches in SemOpenAlex for { <https://semopenalex.org/work/W4295771990> ?p ?o ?g. }
Showing items 1 to 68 of
68
with 100 items per page.
- W4295771990 endingPage "50" @default.
- W4295771990 startingPage "50" @default.
- W4295771990 abstract "To develop and validate an artificial neural network (ANN) model using clinical and biochemical data that allows the early prediction of the evolution of mild to severe COVID-19. The architecture of the ANN model consisted of an input layer of maternal features, a hidden layer with activation functions, and an output layer, the prediction of severe COVID-19 in pregnant women. Layers are connected through weights and biases, which allow the model to be adjusted. Input variables included: anthropometric, clinical, and blood panel profile. The dataset was randomly divided into learning (75%) and validation/test (25%). Five-fold cross-validation was used to minimise overfitting. Back-propagation neural networks were utilised for training and validating the models using the Levenberg-Marquardt algorithm with Matlab software and the Deep Learning Toolbox The performance for cross-validation and validation/test were evaluated with three metrics: the Root Mean Square Error. The program was run 30,000 times with 100 iterations by each neuron. 98 pregnant women were studied, 30% developed severe COVID-19, including 4 maternal deaths. We conducted a first input selection process, considering variables associated with COVID-19 severity. 12 input variables were included: age, diabetes, chronic hypertension, BMI, MAP, neutrophils, lymphocytes, platelets, troponin, PlGF, and sFlt-1. In the internal validation, the model showed good performance in training and testing with an R2 of 0.926 between the real data and the values estimated by the model. In predicting whether pregnant women at admission will be diagnosed with severe COVID-19, the model showed an AUC of 0.714 with a sensitivity of 57.1%, a specificity of 100%, a PPV of 100% and NPV of 75%. The implementation of ANN models using clinical and biochemical data in COVID-19 pregnant women allows prediction of the evolution of C0VID-19 severe form. A risk calculator could be helpful in the initial selection of women susceptible to closer surveillance and would allow to implement strategies that could modify the outcome of severe COVID-19." @default.
- W4295771990 created "2022-09-15" @default.
- W4295771990 creator A5034206051 @default.
- W4295771990 creator A5043181763 @default.
- W4295771990 creator A5047146532 @default.
- W4295771990 creator A5050741144 @default.
- W4295771990 creator A5060627517 @default.
- W4295771990 creator A5063851157 @default.
- W4295771990 creator A5070951144 @default.
- W4295771990 creator A5090380504 @default.
- W4295771990 date "2022-09-01" @default.
- W4295771990 modified "2023-09-26" @default.
- W4295771990 title "OP02.01: An artificial neural network model for severe COVID‐19 prediction in pregnant women" @default.
- W4295771990 doi "https://doi.org/10.1002/uog.25113" @default.
- W4295771990 hasPublicationYear "2022" @default.
- W4295771990 type Work @default.
- W4295771990 citedByCount "0" @default.
- W4295771990 crossrefType "journal-article" @default.
- W4295771990 hasAuthorship W4295771990A5034206051 @default.
- W4295771990 hasAuthorship W4295771990A5043181763 @default.
- W4295771990 hasAuthorship W4295771990A5047146532 @default.
- W4295771990 hasAuthorship W4295771990A5050741144 @default.
- W4295771990 hasAuthorship W4295771990A5060627517 @default.
- W4295771990 hasAuthorship W4295771990A5063851157 @default.
- W4295771990 hasAuthorship W4295771990A5070951144 @default.
- W4295771990 hasAuthorship W4295771990A5090380504 @default.
- W4295771990 hasBestOaLocation W42957719902 @default.
- W4295771990 hasConcept C105795698 @default.
- W4295771990 hasConcept C119857082 @default.
- W4295771990 hasConcept C139945424 @default.
- W4295771990 hasConcept C154945302 @default.
- W4295771990 hasConcept C22019652 @default.
- W4295771990 hasConcept C33923547 @default.
- W4295771990 hasConcept C41008148 @default.
- W4295771990 hasConcept C45804977 @default.
- W4295771990 hasConcept C50644808 @default.
- W4295771990 hasConcept C71924100 @default.
- W4295771990 hasConceptScore W4295771990C105795698 @default.
- W4295771990 hasConceptScore W4295771990C119857082 @default.
- W4295771990 hasConceptScore W4295771990C139945424 @default.
- W4295771990 hasConceptScore W4295771990C154945302 @default.
- W4295771990 hasConceptScore W4295771990C22019652 @default.
- W4295771990 hasConceptScore W4295771990C33923547 @default.
- W4295771990 hasConceptScore W4295771990C41008148 @default.
- W4295771990 hasConceptScore W4295771990C45804977 @default.
- W4295771990 hasConceptScore W4295771990C50644808 @default.
- W4295771990 hasConceptScore W4295771990C71924100 @default.
- W4295771990 hasIssue "S1" @default.
- W4295771990 hasLocation W42957719901 @default.
- W4295771990 hasLocation W42957719902 @default.
- W4295771990 hasOpenAccess W4295771990 @default.
- W4295771990 hasPrimaryLocation W42957719901 @default.
- W4295771990 hasRelatedWork W1996541855 @default.
- W4295771990 hasRelatedWork W2985459377 @default.
- W4295771990 hasRelatedWork W2989932438 @default.
- W4295771990 hasRelatedWork W3011996705 @default.
- W4295771990 hasRelatedWork W3099765033 @default.
- W4295771990 hasRelatedWork W3128220493 @default.
- W4295771990 hasRelatedWork W3175189414 @default.
- W4295771990 hasRelatedWork W4210794429 @default.
- W4295771990 hasRelatedWork W4224929651 @default.
- W4295771990 hasRelatedWork W4313448588 @default.
- W4295771990 hasVolume "60" @default.
- W4295771990 isParatext "false" @default.
- W4295771990 isRetracted "false" @default.
- W4295771990 workType "article" @default.