Machine learning to deal with missing disability status: Ascertainment and imputation of outcomes should be distinguished

[摘要] We appreciate the comments by Kent and Steyerberg.1While the authors acknowledge the value of thisapproach to enhance real-world data, they bring upimportant considerations related to our novel approachusing machine learning to create an estimate of theexpanded disability status scale (eEDSS) scores fromunstructured, neurologist-written clinical notes to substantially increase the number of encounters with anendpoint.