Easing the burden: A pilot study evaluating AI-generated In-Basket message drafts to streamline perioperative...

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Endocrine Surgery

Background:

• Artificial intelligence (AI) shows potential for accurate and empathetic patient communication.
• The impact of AI on reducing health care provider (HCP) clinical burden is not well-measured.
• This study assesses AI's effect on HCP workload for perioperative endocrine surgery inquiries.

Purpose of the Study:

• To evaluate the impact of AI-generated draft messages on HCP workload.
• To measure changes in response time and cognitive load when using AI assistance.
• To assess HCP satisfaction with AI-assisted communication.

Main Methods:

• 11 HCPs responded to 20 In-Basket messages, comparing AI-generated drafts (GPT-4) with blank drafts.
• Response times were timed, and text similarity measured edits.
• Cognitive load was assessed using the NASA Task Load Index; provider satisfaction was surveyed.

Main Results:

• AI drafts reduced average response time from 137s to 49s (P < .001).
• AI drafts required minimal edits (similarity ratio 0.88 ± 0.24) and lowered cognitive load (P < .001).
• 64% of HCPs were satisfied, 82% wanted to integrate AI, noting personalization is key.

Conclusions:

• AI-generated drafts effectively reduce HCP response time and cognitive load for perioperative endocrine surgery questions.
• The use of AI in clinical practice enhances HCP satisfaction and requires minimal edits.
• AI tools show promise for streamlining communication and reducing burnout in surgical specialties.