Accepted for/Published in: JMIR Diabetes
Date Submitted: Nov 3, 2021
Date Accepted: Apr 8, 2022
(closed for review but you can still tweet)
Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
Identifying Patients with Hypoglycemia Using Natural Language Processing: A Systematic Literature Review
ABSTRACT
Background:
Approximately 34 million (~13.0%) U.S. adults have diabetes . [1] Worldwide, a total of 387-million persons have diabetes, a number that is expected to rise to 592 million by 2035. [2] In 2017, direct and indirect costs attributed to diabetes in the U.S. were estimated to $327 billion. [3] Optimal glycemic control (HbA1c < 7%) can be achieved with comprehensive antidiabetic treatment; however, the risk of hypoglycemia increases. Hypoglycemia is defined as level 1: glucose level 54-70 mg/dL; level 2: glucose level <54 mg/dL; level 3: a severe event requiring external assistance for recovery. [4] In T2D after experiencing hypoglycemia, the 3- year incidence of cardiovascular (CV) events was 35.1% and mortality 28.3%-31.9%.
Objective:
Accurately identifying patients with hypoglycemia is key in preventing adverse events and mortality. Natural language processing (NLP) is a scalable, efficient, and quick method to extract hypoglycemia related information when using electronic health records resources (EHRs) from a large population. The objective of this systematic review was to synthesize the literature on the application of NLP to extract hypoglycemia from the EHR clinical notes.
Methods:
Literature searches were conducted electronically from Ovid PsycINFO (Ovid), PubMed, CINAHL (EBSCO), and Web of Science Core Collection. Keywords included hypoglycemia, low blood glucose, NLP, and machine learning. Inclusion criteria included studies that applied NLP to identify hypoglycemia, reported the outcomes related to hypoglycemia, and were published in English as full papers.
Results:
This review (N=7 studies) revealed heterogeneity of the reported results related to hypoglycemia. Four studies reported that the prevalence of any level of hypoglycemia was 3.4 - 46.2%. The use of NLP to analyze clinical notes improved the capture of undocumented or missed hypoglycemic events using ICD-9/10 codes or lab testing. The combination of NLP and ICD-9/10 codes significantly increased the identification of hypoglycemic events compared with individual methods, for example the prevalence of hypoglycemia was 12.4% for ICD codes, 25.1% for NLP algorithm, and 32.2% for combined algorithms. All 7 reviewed studies applied rule-based NLP algorithms to identify hypoglycemia.
Conclusions:
The findings provided evidence that the application of NLP to analyze clinical notes improved the capture of hypoglycemic events, particularly when combined with the ICD-9/10 codes and lab testing.
Citation
