PERCEIVED ADEQUACY AND INFORMATIONAL QUALITY OF AN ARTIFICIAL INTELLIGENCEGENERATED PRE-ANAMNESIS IN EMERGENCY DEPARTMENTS

Artificial intelligence has been increasingly incorporated into emergency care to support documentation and clinical workflow, but its real-world value depends on physician acceptance. This cross-sectional descriptive study evaluated physicians’ perceptions of a pre-anamnesis generated by an artificial intelligence system used in the adult and paediatric emergency departments of a private tertiary-care hospital in Manaus, Brazil. SAMIA, an institutional artificial intelligence assistant integrated into the electronic health record, generates pre-anamnesis summaries before physician evaluation. All physicians working in these settings were invited to complete an anonymous online survey. A total of 74 physicians participated. Regarding contextual adequacy, 70 (94.6%) considered the generated pre-anamnesis at least partially adequate in their clinical encounters, whereas 4 (5.4%) reported that it was rarely adequate. Regarding informational content, 48 (64.8%) rated it as adequate or excellent, whereas 26 (35.2%) considered it less than optimal. These findings suggest substantial physician acceptability, while highlighting contextual precision and informational calibration as priorities for further refinement.