KAID Health Technology Demonstrates the Value of Natural Language Processing to Improve Preoperative Care

NLP engine holds promise to automate preoperative assessment, including at times outperforming clinicians.

KAID Health, an AI-powered health care data analysis and provider engagement platform, announced the results of a study that validates the potential utility of its Natural Language Processing (NLP) technology to improve provider efficiency and care quality. The peer-reviewed “doctor vs. artificial intelligence” paper found that KAID Health’s NLP technology greatly aligned with the clinician reviewers in completing a pre-operative checklist, and further was able to identify 16.6% of instances where the presence or absence of a specific condition was not found by the anesthesiologist. The research study was undertaken at UCSD’s Department of Anesthesiology, Division of Perioperative Informatics. The authors of the manuscript were Harrison S. Suh, BS, Jeffrey L. Tully, MD, Minhthy N. Meineke, MD, Ruth S. Waterman, MD, MS, and Rodney A. Gabriel, MD, MAS.

Recommended AI News: Wipro and Informatica Partner to Bring Cloud Based Data and Analytics to Market With Wipro FullStride Cloud Services Data Platform

“We have demonstrated that NLP technology can help identify critical medical conditions relevant to preanesthetic evaluation. Key to this was KAID Health’s ability to utilize unstructured free-text input from the electronic medical record (EMR) to flag critical medical conditions for anesthesiologists,” researcher and senior author Dr. Rodney Gabriel explained. “This research shows that NLP may be a useful tool to aid preoperative anesthesia providers in screening and evaluation of surgical patients.”

For each of the 93 patients in the study, researchers collected all pertinent free-text notes from the EMR. The free-text notes were then processed by a Named Entity Recognition pipeline, which incorporated an NLP machine learning model developed by KAID Health. The model recognizes and labels spans of text that correspond to medical concepts. Medical concepts were then mapped to a list of medical conditions that were of interest for a preanesthesia evaluation. The most common conditions that the NLP pipeline captured that the anesthesiologist did not include cardiac arrhythmias, angina, anticoagulation, peripheral vascular disease, obstructive sleep apnea, and neuromuscular disease.

Recommended AI News: LTI Achieves Premier Partner Status in the AWS Partner Network

“We are proud that leading academic institutions, such as UCSD, are partnering with KAID to ensure our NLP and AI models meet the demanding accuracy and usability standards required of the industry,” said Kevin Agatstein, CEO of KAID Health. “KAID Health’s NLP technology played a vital role in this study in identifying pertinent preanesthesia history to optimize efficiency for anesthesiologists. Our model demonstrated that NLP holds the potential to reduce clinician workloads, improve profitability, and, most importantly, make surgery safer.”

The International Anesthesia Research Society published the study, “Identification of Preanesthetic History Elements by a Natural Language Processing Engine,” in its journal, Anesthesia & Analgesia, one of the leading anesthesiology journals in the world. To read the research, please c*********.

Recommended AI News: Datamatics Enters in a Long-term Strategic Partnership with PSI Services for Customer Management Solutions

[To share your insights with us, please write to sghosh@martechseries.com]