Citation

Klock M, Kang H, Gong Y. Stud. Health Technol. Inform. 2019; 264: 639-643.

Affiliation

School of Biomedical Informatics, The University of Texas Health Science Center at Houston, Houston, TX, USA.

Copyright

(Copyright © 2019, IOS Press)

DOI

10.3233/SHTI190301

PMID

31438002

Abstract

Patient falls, a subcategory of patient safety events, cause further harm and anxiety to patients in healthcare systems. Patient fall reports are a valuable resource to identify safety issues that demand further attention. Still, the main challenge for patient fall reports is the lack of quality and detail in writing. A method of evaluating patient fall reports would help us better understand the root causes of falls and prevent their recurrence to improve patient safety. Employing the Agency for Healthcare and Quality rubric for assessing the quality of fall reports, we compared three different machine-learning models and identified the most effective method for scoring fall reports using AHRQ's rubric. The results of this study are intended to be applicable in healthcare facilities to score reports during reporting for reporters to improve report quality. The ultimate goal is to increase learning from fall reports for better prevention of patient falls.

Language: en

Keywords

Falls; Machine Learning; Patient Safety