Research Analysis for Enterprise AI Adoption
Exploring Physicians' Information-Seeking Behavior in the AI Era: A Survey on LLM and Knowledge Graph Perceptions
This study delves into the evolving information-seeking behaviors of physicians in the AI era, focusing on their perceptions and adoption challenges regarding Large Language Models (LLMs) and Knowledge Graphs (KGs). It highlights the critical need for trustworthy, AI-enhanced medical information retrieval systems tailored to clinical needs.
By RIKI BHARALI, HAMZAH BIN OSOP, ZECHAN WANG, MONICA ANTHONY MARY LAWRENCE in HIKM '25: Proceedings of the 2025 18th Health Informatics Knowledge Management Conference (September 2025)
Executive Impact: Key Findings for Healthcare AI
Understand the critical shifts in physician information access and the strategic opportunities for AI integration, based on direct insights.
0
Physicians Prioritize Accuracy
0
ChatGPT is Most Used LLM
0
Aware of KGs, But Not Using
0
Information Overload is Key Barrier
Deep Analysis & Enterprise Applications
Select a topic to dive deeper into the specific findings from the research, rebuilt as interactive, enterprise-focused modules to inform your AI strategy.
General ISB
GenAI/LLM Usage
KG Perceptions
100%
of physicians rate 'Treatment Plan reference' as the most important motivation for information-seeking.
0
Identify Information Overload & Fragmented Knowledge as primary obstacles.
0
Prioritize 'Accuracy' in clinical information.
0
Rely most on PubMed for information.
| Category | Most Important Factor | Least Important Factor |
|---|---|---|
| Motivation |
|
|
| Information Source |
|
|
| Database Used |
|
|
| Obstacles Encountered |
|
|
| Information Attributes Prioritized |
|
|
Physicians' Enduring Reliance on Authoritative Sources
Despite the rise of AI, physicians consistently prefer established authoritative databases like PubMed and UpToDate. This reflects a deep-seated trust in the credibility and traceability of information, essential for clinical decision-making and patient safety. They are cautious about directly applying AI-generated outputs, favoring their use for information reprocessing tasks like literature summarization. This highlights the critical need for AI tools to deliver transparent, verifiable, and accountable outputs to gain clinical acceptance.
83%
of physicians reported ChatGPT as the most used LLM tool.
0
Physicians express 'Somewhat Trusting' attitude towards GenAI/LLM tools.
0
Identify 'Lack of Transparency in Reasoning' as a major limitation.
0
See 'Literature Summarization' as a key application scenario.
| Category | Most Important Factor | Least Important Factor |
|---|---|---|
| Usage Frequency |
|
|
| Tools Used |
|
|
| Application Scenario |
|
|
| Perceived Advantages |
|
|
| Trust in Tools |
|
|
| Limitations Noted |
|
|
| Suggested Improvements |
|
|
50%
of physicians are aware of KGs but have never used them in practice.
0
Perceive KGs as 'Not or Slightly Helpful'.
0
Have considered KGs but not used them.
0
Express 'Low' to 'Moderate' demand for KGs in practice.
Study Methodology Flow
Cross-sectional, exploratory study (Physicians' ISB in AI era)
→
Semi-structured survey design (quantitative & qualitative)
→
Module 1: General Info-Seeking Behavior (baseline, needs, barriers)
→
Module 2: GenAI/LLM Tool Usage (adoption, perceptions, trust)
→
Module 3: Knowledge Graph Perceptions (awareness, experience, value)
→
Data Analysis (trends, qualitative supplements)
Bridging the KG Adoption Gap
The near-total lack of practical exposure to Knowledge Graphs (KGs) in clinical environments highlights a significant gap. Physicians perceive KGs as 'Not or Slightly Helpful,' indicating that limited understanding and hands-on experience hinder adoption. However, 83% have considered KGs, suggesting conceptual openness. To bridge this, technologies like KGs require tailored onboarding strategies, contextual customization, and institutional support.
Calculate Your Potential AI ROI
Estimate the time and cost savings your enterprise could realize by implementing AI-powered information retrieval and knowledge management solutions.
Estimated Annual Savings
$0
Annual Hours Reclaimed
0
Your Enterprise AI Implementation Roadmap
A phased approach to integrate AI-powered solutions, ensuring alignment with physician needs and clinical workflows for maximum impact.
Phase 1: Needs Assessment & Pilot (3-6 Months)
Conduct detailed interviews and workshops with clinical staff to identify specific information-seeking pain points and opportunities for AI. Select a pilot department and implement an initial AI-driven summarization or knowledge retrieval tool. Establish KPIs for trust, efficiency, and accuracy. Focus on transparent source attribution and explainability.
Phase 2: Customization & Integration (6-12 Months)
Refine AI models based on pilot feedback, incorporating domain-specific knowledge graphs for enhanced accuracy and context. Integrate AI solutions directly into existing EMR/EHR systems and clinical workflows. Develop structured training programs emphasizing practical application and trust-building strategies, addressing hallucination concerns.
Phase 3: Scaled Deployment & Continuous Improvement (12+ Months)
Expand AI solutions across multiple departments, with ongoing monitoring of performance, user adoption, and ROI. Establish a continuous feedback loop for iterative model improvement and adaptation to new medical knowledge. Explore advanced features like diagnostic support with robust validation and explainability. Secure institutional endorsement.
Ready to Transform Physician Information Access?
Book a free consultation with our AI strategists to design a bespoke solution that aligns with your clinical objectives and drives efficiency.
Ready to Get Started?
Book Your Free Consultation.
Let's Discuss Your AI Strategy!
Lets Discuss Your Needs
Select a Date
Sun
Mon
Tue
Wed
Thu
Fri
Sat