Commentary: Predicting Discharge Disposition Following Meningioma Resection Using a Multi-Institutional Natural Language Processing Model

Abstract

Machine learning is widely prevalent across multiple sectors in society. In health care, the regulatory process and safety concerns have appropriately slowed the adoption of machine learning into clinical practice. As studies demonstrating safety and efficacy propagate, there is great potential to deploy these algorithms to enhance patient care. Within neurosurgery, machine learning has been used in various ways from radiographic segmentation to outcome prediction.1-3 A major factor that influences the quality of an algorithm's performance is the data set by which it was trained. Robust and detailed data sets generate better predictive algorithms, but rarely are clinical data presented in such a clean manner. Extracting the nuanced data needed to design the best models may be time-consuming and cumbersome. In “Predicting Discharge Disposition Following Meningioma Resection Using A Multi-Institutional Natural Language Processing Model,”4 the authors evaluate the potential of a machine learning algorithm based on easily accessible...

Ähnliche Arbeiten