The Ambiguous Medical Abbreviation (AMA) Study ‐ Challenges and Opportunities

BACKGROUND: Healthcare workers often abbreviate for convenience, but ambiguous abbreviations may cause miscommunication, which jeopardises patient care. Robust large-scale research to quantify abbreviation frequency and ambiguity in medical documents is lacking. AIMS: To calculate the frequency of abbreviations used in discharge summaries, the proportion of these abbreviations that are ambiguous and the potential utility of auto-expansion software. METHODS: We designed a software program to extract all instances of abbreviations from every General Medical Unit discharge summary from the Royal Melbourne Hospital in 2015. We manually expanded abbreviations using published inventories and clinical experience, logging multiple expansions for any abbreviation if identified. Abbreviations were classified based on well-defined criteria as standardised and likely to be well-understood, or ambiguous. Outcome measures included the range and frequency of standardised and ambiguous abbreviations, and the feasibility of electronic auto-expansion software based on these measures. RESULTS: Of the 1 551 537 words analysed from 2336 documents, 137 997 (8.9%) were abbreviations with 1741 distinct abbreviations identified. Most abbreviations (88.7%) had a single expansion. The most common abbreviation was PO (per os/orally), followed by BD (bis in die/twice daily), and 68.1% of abbreviations were standardised, largely pertaining to pathology/chemicals. This meant, however, that a large proportion (31.9%) of abbreviations (2.8% of all words) were ambiguous. The most common ambiguous abbreviation was Pt (patient/physiotherapy), followed by LFT (liver function test/lung function test). CONCLUSIONS: Close to one-third of abbreviations used in general medical discharge summaries were ambiguous. Electronic auto-expansion of ambiguous abbreviations is likely to reduce miscommunication and improve patient safety. This article is protected by copyright. All rights reserved.