Table 3. Patient confidence and satisfaction with the original radiology reports and ChatGPT-generated summaries (N=118).

Measure and response	Original report, n (%)	Summaries, n (%)	P value
Confidence in understanding			<.001
Very confident/somewhat confident	31 (26)	116 (98)
Neutral	14 (12)	1 (1)
Very unconfident/somewhat unconfident	73 (62)	1 (1)
Satisfaction with medical jargon/terminology			<.001
Very satisfied/somewhat satisfied	9 (8)	107 (91)
Neutral	18 (15)	9 (8)
Very dissatisfied/somewhat dissatisfied	91 (77)	2 (2)
Satisfaction with level of detail			<.001
Very satisfied/somewhat satisfied	51 (43)	116 (98)
Neutral	43 (36)	1 (1)
Very dissatisfied/somewhat dissatisfied	24 (20)	1 (1)
Satisfaction with time to comprehension			<.001
Very satisfied/somewhat satisfied	27 (23)	115 (97)
Neutral	31 (26)	2 (2)
Very dissatisfied/somewhat dissatisfied	60 (51)	1 (1)