Terms that have the identical linguistic form but different meanings are polysemous. Biomedical polysemy manifests itself in different ways (Roth and Hole, 2000, Liu et al., 2001). Some words in clinical texts have different biomedical meanings or word senses. For instance, the word "discharge'' has two word senses—one word sense meaning a procedure for being released from the hospital, as in "prior to discharge,'' and one word sense meaning a substance that is emitted from the body, as in "purulent discharge.''
Acronyms and abbreviations with more than one meaning may be the most frequently occurring type of biomedical polysemy. A striking example of this is the acronym "APC,'' which has more than thirty unique biomedical definitions, including activated protein c, adenomatosis polyposis coli, antigen-presenting cell, aerobic plate count, advanced pancreatic cancer, age period cohort, and alfalfa protein concentrated. According to one study (Wren and Garner, 2002), 36% of the acronyms in MEDLINE are associated with more than one definition. The number of unique acronyms in MEDLINE is increasing at the rate of 11,000 per year, and the number of definitions associated with unique acronyms is increasing at 44,000 per year. In the sublanguage of patient reports, the type of report is helpful in disambiguating the correct meaning of an acronym or abbreviation, because the report type indicates the type of medical specialty. In this way, "APC'' in a microbiology lab report is more likely to mean aerobic plate count, whereas "APC'' in a discharge summary may be referring to advanced pancreatic cancer.
Triage chief complaints are full of abbreviations created by clerks and triage nurses to keep the complaint short (Travers and Haas, 2003). Some of the abbreviations are standard and are easily understood by physicians, such as "rt'' for "right'' and "h/a'' for headache. But many abbreviations in chief complaints are unique to the sublanguage of chief complaints or perhaps even to a single hospital or registration clerk. For example, "appy'' is commonly used to describe an "appendectomy,'' and in one hospital "gx'' indicates the patient came to the ED by ground transportation.
Depending on the particular clinical variables that we want to extract or encode from text, understanding the meaning or word sense of polysemous words in the patient reports can be critical to success.
Was this article helpful?