10/27/2025
ChatGPT demonstrates efficiency in clinical trial recruitment https://buff.ly/cMlROFK
AI-assisted trial recruitment: promising, but not perfect.
A new study in Machine Learning: Health evaluated how ChatGPT (GPT-3.5 and GPT-4) performs in identifying eligible patients for a head and neck cancer trial.
🔹 GPT-4 outperformed GPT-3.5 (AUC 0.838 vs 0.761) in identifying eligible patients
🔹 Structured prompts + expert guidance improved accuracy
🔹 GPT-4 was slower and costlier (7.9–12.4 min, ~$0.15–$0.27 per screen), but still efficient at scale
🔹 Best use case: AI to narrow candidates → clinician to confirm eligibility
Dr. Michael Dohopolski (UT Southwestern) noted that while LLMs can’t yet interpret nuanced clinical criteria, they can dramatically reduce manual workload in screening electronic health records.
As AI evolves, these tools could become valuable adjuncts for research teams facing limited personnel and time pressures, provided models remain HIPAA-compliant and clinician oversight stays central.
Multiple versions of ChatGPT demonstrated efficacy in identifying eligible patients for clinical trial recruitment, showing potential to complement manual chart reviews, according to a study published in Machine Learning: Health.Improving clinical trial recruitment has been an ongoing challenge in t...