On the Creation of a Corpus-Derived Medical Multi-Word Term List

Although several studies have succeeded in creating medical word lists using corpus analysis methods, there is currently a shortage of comprehensive lists containing medical multi-word terms (MWTs). This study attempts to fill this gap by identifying medical MWTs using a large corpus of English language medical textbooks (28,384,681 running words). The term extraction function in Sketch Engine was used to extract high-frequency MWTs and to calculate keyness and dispersion data for each MWT. The validity of the resulting list and of specific subsets was tested using a different medical corpus and a general English corpus. The resulting list comprises 3307 MWTs with 63.83% (2111 MWTs) occurring at comparable frequencies in the different medical corpus and only 0.97% (32 MWTs) occurring at comparable frequencies in the general English corpus. The study also revealed clear differences in replicability between semantic subsets, with MWTs from the Anatomy and the Disorders semantic groups displaying high replicability, while MWTs from the Concepts and Ideas semantic group showed low to moderate replicability. The list may be used to develop evidence-based materials in English for Medical Purposes courses and to further explore how information is packaged in healthcare communication settings.