Clinical Errors From Acronym Use in Electronic Health Record: A Review of NLP-Based Disambiguation Techniques

The adoption of Electronic Health Record (EHR) and other e-health infrastructures over the years has been characterized by an increase in medical errors. This is primarily a result of the widespread usage of medical acronyms and abbreviations with multiple possible senses (i.e., ambiguous acronyms). The advent of Artificial Intelligence (AI) technology, specifically Natural Language Processing (NLP), has presented a promising avenue for tackling the intricate issue of automatic sense resolution of acronyms. Notably, the application of Machine Learning (ML) techniques has proven to be highly effective in the development of systems aimed at this objective, garnering significant attention and interest within the research and industry domains in recent years. The significance of automating the resolution of medical acronym senses cannot be overstated, especially in the context of modern healthcare delivery with the widespread use of EHR. However, it is disheartening to note that comprehensive studies examining the global adoption of EHR, assessing the impact of acronym usage on medical errors within EHR systems, and reporting on the latest trends and advancements in ML-based NLP solutions for disambiguating medical acronyms remain severely limited. In this current study, we present a detailed overview on medical error, its origins, unintended effects, and EHR-related errors as a subclass of clinical error. Furthermore, this paper investigates the adoption of EHR systems in developed and developing nations, as well as the review concludes with an examination of various artificial intelligence techniques, particularly machine learning algorithms for medical acronym and abbreviation disambiguation in EHRs.