Investigation of Pressure Injuries With Visual ChatGPT Integration: A Descriptive Cross-Sectional Study.
Pelin Karaçay, Polat Goktas, Özgen Yaşar and 4 others
PMID 40084802WHAT IT FOUND
Visual ChatGPT staged pressure injuries less accurately than expert nurses overall, especially unstageable and deep tissue injuries.
Adding wound details helped slightly, but the tool did not match nurse assessment.
Key findings
01The expert nurse showed high agreement in pressure injury staging (kappa = 0.731), higher than Visual ChatGPT (kappa = 0.310 with image-only and kappa = 0.366 with image plus wound characteristics).
02For unstageable pressure injuries, Visual ChatGPT sensitivity was 15.8% with image-only and 36.8% with wound characteristics, compared with 84.2% for the expert nurse.
03For deep tissue pressure injuries, Visual ChatGPT sensitivity was 2.3%, compared with 70.5% for the expert nurse.
STILL TO COME
How it was doneWhat they foundWhat it means for RNs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study used 155 pressure injury images from a hospital database, so the results may not apply to other settings. The number of Stage III and Stage IV images was small. Wound characteristics used for Visual ChatGPT were recorded by wound care nurses, and expert nurses found it difficult to assess size, depth, undermining and tunnelling from images alone. The study was retrospective and used images already stored in the hospital system.
Declared interests
The authors declared no conflicts of interest.
The easy way to misread this
Do not conclude that Visual ChatGPT can replace nurse staging. For deep tissue pressure injuries its sensitivity was 2.3%, and for unstageable pressure injuries its image-only sensitivity was 15.8%.