Table 2.

Medical interpreter–rated translation accuracy and lay participant–rated translation satisfaction scores for ChatGPT, Google Translate, and UD Talk^a.

		ChatGPT (0-5), median (IQR)	Google Translate (0-5), median (IQR)	UD Talk (0-5), median (IQR)	P value
Accuracy
	Overall (n=320)	5.0 (4.0-5.0)	2.0 (1.0-3.0)	2.0 (1.0-3.0)	<.001b
	Cardiology outpatient clinic (n=200)	5.0 (4.0-5.0)	2.0 (0.0-3.0)	2.0 (0.0-3.0)	<.001b
	Pulmonology outpatient clinic (n=120)	5.0 (4.0-5.0)	3.0 (1.0-4.0)	2.0 (1.0-4.0)	<.001b
Satisfaction
	Overall (n=350)	2.0 (1.0-3.0)	2.0 (1.0-3.0)	5.0 (4.0-5.0)	<.001b
	Cardiology outpatient clinic (n=210)	2.0 (1.0-3.0)	2.0 (1.0-3.0)	5.0 (4.0-5.0)	<.001b
	Pulmonology outpatient clinic (n=140)	3.0 (1.0-4.0)	3.0 (2.0-4.0)	5.0 (4.0-5.0)	<.001b

^aEight medical interpreters evaluated 40 dialogue exchanges (cardiology: n=25, 62.5%; pulmonology: n=15, 37.5%) for translation accuracy, and 14 lay participants evaluated 25 dialogue exchanges (cardiology: n=15, 60%; pulmonology: n=10, 40%) for translation satisfaction.

^bP<.001.