At the Research Center for Emergency Medicine, ringing the bell marks a milestone — and this this week, PhD student Sandra Thun Langsted had every reason to celebrate 🎉
She just published two papers, including her 2nd PhD article

📄 #1: Identifying the optimal thoracentesis training strategy: a randomized non-inferiority study
Link: https://lnkd.in/diAKgKjx

Can we train emergency doctors more effectively without taking them away from the clinic?

This multicenter RCT compared traditional instructor-led thoracentesis courses with a blended, self-directed, spaced learning program integrated into clinical work hours.

🔍 Key findings:
✅ Comparable skill acquisition after 2 weeks
✅ Superior skill retention at 3 months with spaced learning
✅ Higher procedural volume in spaced learning group
✅ Better integration into clinical workflows

💡 This study highlights the potential of spaced and self-directed learning in clinical education – especially in a time where resources, supervision and training time are increasingly limited

🧪 In collaboration with: Bo Løfgren, Kasper G. Lauridsen, Søren Helbo Skaarup & Jesper Weile.

🌍 #2: International collaboration with Royal London Hospital
During her research stay, Sandra joined the ED research team at Royal London Hospital led by Ben Bloom and contributed to the DECIPHER study, published in BMJ Health & Care Informatics.
Link: https://lnkd.in/dNJ7gcpW

Goal: Use AI to automate classification of head CT reports for intracranial bleeds (ICB).

Three methods were tested: Text classification program, a commercial NLP tool, and a custom GPT-based LLM

🧠 Highlights:
✔️ 100% sensitivity and 97.4% specificity for the LLM
✔️ 86% reduction in manual review workload
✔️ Outperformed both competing methods

💥 A promising tool for scaling up research and improving data workflows in emergency medicine.