Logo image
Evaluating the Accuracy of Google Translate in Emergency Department Discharge Instructions
Journal article   Peer reviewed

Evaluating the Accuracy of Google Translate in Emergency Department Discharge Instructions

Jossie A. Carreras Tartak, Adrian D. Haimovich, Caroline E. Thomas, Emily Bivens Moore, Anna Slebonick, Nabeeha Haque and Stephen H. Thomas
The Journal of emergency medicine, v 79, pp 132-142
01 Dec 2025
PMID: 41129839

Abstract

Discharge instructions Google Translate Language Limited English proficiency Machine Translation
Most emergency departments (EDs) don’t have interpreter services bandwidth to translate written discharge instructions in a timely manner for patients with limited English proficiency (LEP). Prior studies evaluating the accuracy of Google Translate have been variable across languages. We aimed to evaluate if translations of ED discharge (DC) instructions by Google Translate are acceptable to patients and if there are language characteristics that can predict accuracy. We prepared DC instructions for three hypothetical ED visits. Reviewers with self-identified fluency in English and a non-English language were asked to indicate “Yes” or “No” to instruction-specific queries about translation accuracy. Accuracy was analyzed across different DC instruction domains and across different language categories. Comparisons between language groups were executed using nonparametric Kruskal–Wallis testing. We identified a 95% overall quantitative success rate for ED DC instruction translation by Google Translate across 26 languages. In 16 of 26 languages (61.5%, 95% CI 40.6–79.8%), at least one of the four DC instruction arenas scored 90 or less. The total score was less than 90% in 5 (Amharic, French, Oromo, Portuguese, and Yoruba) of 26 languages (19.2%, 95% CI 6.6–39.4%). There were no differences by DC instruction domain or across different language categories. Using Google Translate for DC instructions found an overall error rate of 5% across all languages. Specific languages for which Google Translate functions better are not easily identified. Further investigation into the safety of routine use of Google Translate in medical contexts with limited interpreter availability is warranted.

Metrics

Details

UN Sustainable Development Goals (SDGs)

This publication has contributed to the advancement of the following goals:

#3 Good Health and Well-Being
#10 Reduced Inequalities

Source: SDGs in the Output

InCites Highlights

Data related to this publication, from InCites Benchmarking & Analytics tool:

Collaboration types
Domestic collaboration
International collaboration
Web of Science research areas
Emergency Medicine
Logo image