Natural Language Processing Papers Win International Recognition
March 30, 2017
The International Medical Informatics Association (IMIA) has named two papers by Yizhao Ni, PhD, and his colleagues as among the five best natural language processing (NLP) papers of 2015. Both papers are featured in the IMIA Yearbook of Medical Informatics 2016, and have received other accolades as well.
Yizhao Ni is an instructor in the Division of Biomedical Informatics at Cincinnati Children’s Hospital Medical Center. His research focuses on the development of machine learning, NLP, and information retrieval techniques to assist clinical decision making.
The first article featured describes the use of machine learning and NLP techniques to develop a computerized system for medication reconciliation. The system matches a patient’s discharge prescriptions (structured data) against medications documented in free-text clinical notes (unstructured data) to identify discrepancies. The algorithm achieved a 92.4% precision rate on identifying matched medications and 71.5% on discrepant medications. It promises to significantly reduce the time clinicians would spend reconciling medications in the patient admission and discharge processes.
This article was also selected as a top 10 article from BMC Medical Informatics and Decision Making for 2015.
The second article demonstrates that NLP, when used in combination with machine learning and information retrieval, can successfully identify patients eligible for clinical trials. It describes development of an automated system that uses NLP to analyze free-form clinical notes, and machine learning to refine the list of terms extracted from them in order to streamline the pool of potential candidates requiring screening by clinical trial staff. The automated system showed potential to reduce the workload of clinical trial staff by 92% and increase trial screening efficiency by 450%, compared to cases without automation.
It was also featured in the 2015 Year in Review by the American Medical Informatics Association (AMIA) and cited as an example of using technology to increase speed and reduce cost of operations in a business article from Deloitte.
Both articles demonstrate the effectiveness of NLP and machine learning technologies in optimization of precision decisions among clinical units. The works showcase the synergy between computational scientists, information service technicians, and healthcare providers. Project planning is in progress to integrate both systems into a production environment to adequately assess their practicality.
The IMIA Yearbook summarizes recent research and presents a selection of the best papers published in the fields of medical informatics, including clinical natural language processing. IMIA, based in Geneva, Switzerland since 1967, is the world body for health and biomedical informatics.