Why the study?
Heart failure is associated with high rates of unplanned readmissions, and existing models to predict these readmissions demonstrate suboptimal performance.
Do word embedding-based feature generation approaches improve the prediction of 30-day hospital readmission in heart failure patients compared to one-hot encoding?
Do word embedding-based feature generation approaches improve the prediction of 30-day hospital readmission in heart failure patients compared to one-hot encoding?
Word embedding approaches, particularly word2vec on terminology codes, significantly improve the performance of machine learning models predicting 30-day readmission in heart failure patients compared to traditional one-hot encoding.
May enhance HF readmission models; hypothesis-generating and requires prospective validation before clinical use.
Heart Failure (HF) is a public health concern with a wider impact on quality of life and cost of care. One of the major challenges in HF is the higher rate of unplanned readmissions and sub-optimal performance of models to predict the readmissions. Hence, in this study, we implemented embeddings-based approaches to generate features for improving model performance. Specifically, we compared three embedding approaches including word2vec on terminology codes and CUIs, and BERT on concept descriptions with baseline (one hot-encoding). We found that the embedding approaches significantly improved the performance of the prediction models, and word2vec on the study dataset outperformed pre-trained BERT model.
No takes yet. Share an insight, caveat, or question.
Khaneja et al. (2025) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: