Translation between Structured EHRs and Clinical Notes
Keywords: Text Generation, Clinical Notes, Electronic Health Records
Abstract: Electronic health records (EHRs) consist of structured, formulaic data and unstructured, free-text descriptions that contain valuable patient information are significant for clinical diagnosis. However, it is time-consuming for clinical staff to summarize text descriptions from structured EHRs or to populate the structured EHRs from the text descriptions. This project proposes generation models are proposed to support translation between structured EHRs and text descriptions. A new dataset with structured EHR and text description pairs will be built from existing EHR datasets. Experimental results in terms of both natural language generation and clinical efficacy will prove the effectiveness of automatic generation, allowing the documentation work starts by modifying generated EHRs and text descriptions.
Submission Number: 15
Loading