FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2076864475> ?p ?o ?g. }

• W2076864475 endingPage "50" @default.
• W2076864475 startingPage "49" @default.
• W2076864475 abstract "An existing prediction model may suffer from errors when applied to new data, leading to wrong decisions. This can happen if the new data are collected in a center that was not involved in developing the model, or if a different data collection protocol was used. A useful solution is to adjust the model coefficients. Additionally, one may want to include an additional predictor because data is available or to align the model to one's clinical practices. In this study, these methods are illustrated on logistic regression model LR2 from phase 1 of the International Ovarian Tumor Analysis (IOTA) project. LR2 is modified based on 699 patients from centers that did not contribute to IOTA phase 1. In the first method all model coefficients are recalculated using the new data. In the second method the coefficients are recalculated and the color Doppler score of tumoral flow is added as a new predictor. The adjusted models were evaluated by the area under the ROC curve (AUC) and by odds ratios (OR). Given that IOTA phase 1 was a large multi-center study, and that IOTA protocol was followed when collecting the new data, we did not expect that model adjustment would greatly influence AUC in our study. The original LR2 model had an AUC of 0.947 on the new data. Recalculating coefficients (method 1) gave an AUC of 0.948, and changed some ORs. For example, the OR for papillary blood flow changed from 3.2 to 14.2. Method 2 resulted in an AUC of 0.952. Depending on the situation, medical centers could adjust an existing model to tailor it to their setting. This could be considered when suboptimal performance is suspected, if the value of an extra predictor is to be checked, or if there is specific interest in the ORs in the new setting. Due to the nature of the IOTA project, suboptimal performance of LR2 or another IOTA model is not very likely, conditional on protocol adherence. This is supported by our results." @default.
• W2076864475 created "2016-06-24" @default.
• W2076864475 creator A5002144411 @default.
• W2076864475 creator A5005038962 @default.
• W2076864475 creator A5010913750 @default.
• W2076864475 creator A5011341219 @default.
• W2076864475 creator A5012951954 @default.
• W2076864475 creator A5014259588 @default.
• W2076864475 creator A5026992091 @default.
• W2076864475 creator A5049055621 @default.
• W2076864475 creator A5052561794 @default.
• W2076864475 creator A5070972279 @default.
• W2076864475 creator A5074111460 @default.
• W2076864475 creator A5080961159 @default.
• W2076864475 creator A5086320066 @default.
• W2076864475 creator A5088038029 @default.
• W2076864475 date "2009-09-01" @default.
• W2076864475 modified "2023-10-02" @default.
• W2076864475 title "OC25.05: Adjusting prediction models for ovarian tumor classification to new clinical settings" @default.
• W2076864475 doi "https://doi.org/10.1002/uog.6609" @default.
• W2076864475 hasPublicationYear "2009" @default.
• W2076864475 type Work @default.
• W2076864475 sameAs 2076864475 @default.
• W2076864475 citedByCount "0" @default.
• W2076864475 crossrefType "journal-article" @default.
• W2076864475 hasAuthorship W2076864475A5002144411 @default.
• W2076864475 hasAuthorship W2076864475A5005038962 @default.
• W2076864475 hasAuthorship W2076864475A5010913750 @default.
• W2076864475 hasAuthorship W2076864475A5011341219 @default.
• W2076864475 hasAuthorship W2076864475A5012951954 @default.
• W2076864475 hasAuthorship W2076864475A5014259588 @default.
• W2076864475 hasAuthorship W2076864475A5026992091 @default.
• W2076864475 hasAuthorship W2076864475A5049055621 @default.
• W2076864475 hasAuthorship W2076864475A5052561794 @default.
• W2076864475 hasAuthorship W2076864475A5070972279 @default.
• W2076864475 hasAuthorship W2076864475A5074111460 @default.
• W2076864475 hasAuthorship W2076864475A5080961159 @default.
• W2076864475 hasAuthorship W2076864475A5086320066 @default.
• W2076864475 hasAuthorship W2076864475A5088038029 @default.
• W2076864475 hasConcept C119857082 @default.
• W2076864475 hasConcept C143998085 @default.
• W2076864475 hasConcept C154945302 @default.
• W2076864475 hasConcept C41008148 @default.
• W2076864475 hasConcept C71924100 @default.
• W2076864475 hasConceptScore W2076864475C119857082 @default.
• W2076864475 hasConceptScore W2076864475C143998085 @default.
• W2076864475 hasConceptScore W2076864475C154945302 @default.
• W2076864475 hasConceptScore W2076864475C41008148 @default.
• W2076864475 hasConceptScore W2076864475C71924100 @default.
• W2076864475 hasIssue "S1" @default.
• W2076864475 hasLocation W20768644751 @default.
• W2076864475 hasOpenAccess W2076864475 @default.
• W2076864475 hasPrimaryLocation W20768644751 @default.
• W2076864475 hasRelatedWork W1978808595 @default.
• W2076864475 hasRelatedWork W2417567749 @default.
• W2076864475 hasRelatedWork W2810549550 @default.
• W2076864475 hasRelatedWork W2938809977 @default.
• W2076864475 hasRelatedWork W3026063446 @default.
• W2076864475 hasRelatedWork W3046282410 @default.
• W2076864475 hasRelatedWork W3105352306 @default.
• W2076864475 hasVolume "34" @default.
• W2076864475 isParatext "false" @default.
• W2076864475 isRetracted "false" @default.
• W2076864475 magId "2076864475" @default.
• W2076864475 workType "article" @default.