Deep learning detects and visualizes bleeding events in electronic health records
| Autor: | Eline Sandvig Andersen, Thiusius Rajeeth Savarimuthu, Lou-Ann Christensen Andersen, Charlotte Gils, Pernille Just Vinholt, Martin Sundahl Laursen, Anne-Sofie Faarvang Thorsen, Cathrine Brødsgaard Nielsen, Kristian Voss Bjerre, Rasmus Søgaard Hansen, Anne Alnor, Ina Mathilde Kjær, Jannik Skyttegaard Pedersen, Søren Andreas Just |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
international classification of diseases
Hemorrhage Health records computer.software_genre Deep Learning International Classification of Diseases Electronic health record Machine learning Text messaging Medicine Diseases of the blood and blood-forming organs Decision support system business.industry Deep learning deep learning Original Articles electronic health record Hematology Predictive value machine learning Test set Electronic Health Record Original Article Diagnosis code Artificial intelligence RC633-647.5 hemorrhage decision support systems (clinical) business computer Sentence Natural language processing |
| Zdroj: | Research and Practice in Thrombosis and Haemostasis Pedersen, J S, Laursen, M S, Savarimuthu, T R, Søgaard Hansen, R, Alnor, A B, Bjerre, K V, Kjær, I M, Gils, C, Thorsen, A-S F, Sandvig Andersen, E, Nielsen, C B, Andersen, L-A C, Just, S A & Vinholt, P J 2021, ' Deep learning detects and visualizes bleeding events in electronic health records ', Research and Practice in Thrombosis and Haemostasis, vol. 5, no. 4, e12505 . https://doi.org/10.1002/rth2.12505 Research and Practice in Thrombosis and Haemostasis, Vol 5, Iss 4, Pp n/a-n/a (2021) |
| ISSN: | 2475-0379 |
| Popis: | Background: Bleeding is associated with a significantly increased morbidity and mortality. Bleeding events are often described in the unstructured text of electronic health records, which makes them difficult to identify by manual inspection. Objectives: To develop a deep learning model that detects and visualizes bleeding events in electronic health records. Patients/Methods: Three hundred electronic health records with International Classification of Diseases, Tenth Revision diagnosis codes for bleeding or leukemia were extracted. Each sentence in the electronic health record was annotated as positive or negative for bleeding. The annotated sentences were used to develop a deep learning model that detects bleeding at sentence and note level. Results: On a balanced test set of 1178 sentences, the best-performing deep learning model achieved a sensitivity of 0.90, specificity of 0.90, and negative predictive value of 0.90. On a test set consisting of 700 notes, of which 49 were positive for bleeding, the model achieved a note-level sensitivity of 1.00, specificity of 0.52, and negative predictive value of 1.00. By using a sentence-level model on a note level, the model can explain its predictions by visualizing the exact sentence in a note that contains information regarding bleeding. Moreover, we found that the model performed consistently well across different types of bleedings. Conclusions: A deep learning model can be used to detect and visualize bleeding events in the free text of electronic health records. The deep learning model can thus facilitate systematic assessment of bleeding risk, and thereby optimize patient care and safety. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Plný text