The article examines whether ambient AI scribes, which document speech, miss clinicians’ nonverbal observations and unarticulated judgments, and whether concurrent verbalization could affect clinical reasoning.
Approach:
Clinical context: The article contrasts speech-based documentation with clinical information that clinicians observe or recognize before they can put it into words.
Reasoning and cognition: It reviews analytical and tacit reasoning, research on verbal overshadowing, and the role of documentation in clinical cognition.
Key Findings:
Ambient scribes capture spoken content but may omit observations such as appearance, affect, psychomotor activity, and a clinician’s initial sense that something is wrong.
Tacit pattern recognition can prompt more deliberate analysis, but it can also be erroneous or biased.
Verbal-overshadowing research raises a possible concern about concurrent verbalization; the article notes that direct evidence that ambient scribes impair clinical reasoning is not available.
Interpretation:
The article raises concern that speech-based ambient documentation may fail to capture nonverbal observations and judgments that precede articulation. It discusses verbal-overshadowing research as grounds for caution while noting that direct evidence of cognitive harm from ambient scribes is not yet available.
Limitations:
Direct evidence that ambient scribes impair expert clinical reasoning is not yet available.
Conclusion:
Ambient scribes document speech, but some clinical observations and judgments arise before they can be articulated. Research on verbal overshadowing raises a possible concern about concurrent verbalization, though direct evidence of harm to clinical reasoning from ambient scribes is not yet available.
