Development and Evaluation of a GPT4-Based Orofacial Pain Clinical Decision Support System DOI Creative Commons

Charlotte Vueghs,

Hamid Shakeri,

Tara Renton

et al.

Diagnostics, Journal Year: 2024, Volume and Issue: 14(24), P. 2835 - 2835

Published: Dec. 17, 2024

Background: Orofacial pain (OFP) encompasses a complex array of conditions affecting the face, mouth, and jaws, often leading to significant diagnostic challenges high rates misdiagnosis. Artificial intelligence, particularly large language models like GPT4 (OpenAI, San Francisco, CA, USA), offers potential as aid in healthcare settings. Objective: To evaluate accuracy OFP cases clinical decision support system (CDSS) compare its performance against treating clinicians, expert evaluators, medical students, general practitioners. Methods: A total 100 anonymized patient case descriptions involving diverse were collected. was prompted generate primary differential diagnoses for each using International Classification Pain (ICOP) criteria. Diagnoses compared gold-standard established by scoring used assess at three hierarchical ICOP levels. subset 24 also evaluated two experts, final-year practitioners comparative analysis. Diagnostic interrater reliability calculated. Results: achieved highest level (ICOP 3) 38% cases, with an overall score 157 out 300 points (52%). The model provided accurate 80% (400 500 points). In model’s comparable non-expert human evaluators but surpassed who correctly diagnosed 54% 3. demonstrated specific categories, diagnosing 81% trigeminal neuralgia Interrater between low (κ = 0.219, p < 0.001), indicating variability agreement. Conclusions: shows promise CDSS improving offering structured diagnoses. While not yet outperforming can augment workflows, care or educational Effective integration into practice requires adherence rigorous guidelines, thorough validation, ongoing professional oversight ensure safety reliability.

Language: Английский

Supporting parents in autism care DOI

Mihit Kalawatia,

Brandon Lucke-Wold,

Aabhali Mehrunkar

et al.

World Journal of Psychiatry, Journal Year: 2025, Volume and Issue: 15(5)

Published: April 30, 2025

Autism spectrum disorder is a developmental impacting child’s social interactions, behaviors, and communication skills. One of the crucial aspects autism care, which often overlooked, parent’s mental health status while trying to improve overcome challenges faced by their child. A study Lu et al examined effectiveness remote family psychological support courses on parents having children with disorder. It was found that integration these conventional care had positive impact reducing stress levels, leading an increase in competence, hence they could engage effectively child therapy. However, long-term necessary assess whether interventions have sustained effect. The emphasizes need for developing such culturally sensitive intervention models global scale, making them accessible all improving support.

Language: Английский

Citations

0
Development and Evaluation of a GPT4-Based Orofacial Pain Clinical Decision Support System DOI Creative Commons

Charlotte Vueghs,

Hamid Shakeri,

Tara Renton

et al.

Diagnostics, Journal Year: 2024, Volume and Issue: 14(24), P. 2835 - 2835

Published: Dec. 17, 2024

Background: Orofacial pain (OFP) encompasses a complex array of conditions affecting the face, mouth, and jaws, often leading to significant diagnostic challenges high rates misdiagnosis. Artificial intelligence, particularly large language models like GPT4 (OpenAI, San Francisco, CA, USA), offers potential as aid in healthcare settings. Objective: To evaluate accuracy OFP cases clinical decision support system (CDSS) compare its performance against treating clinicians, expert evaluators, medical students, general practitioners. Methods: A total 100 anonymized patient case descriptions involving diverse were collected. was prompted generate primary differential diagnoses for each using International Classification Pain (ICOP) criteria. Diagnoses compared gold-standard established by scoring used assess at three hierarchical ICOP levels. subset 24 also evaluated two experts, final-year practitioners comparative analysis. Diagnostic interrater reliability calculated. Results: achieved highest level (ICOP 3) 38% cases, with an overall score 157 out 300 points (52%). The model provided accurate 80% (400 500 points). In model’s comparable non-expert human evaluators but surpassed who correctly diagnosed 54% 3. demonstrated specific categories, diagnosing 81% trigeminal neuralgia Interrater between low (κ = 0.219, p < 0.001), indicating variability agreement. Conclusions: shows promise CDSS improving offering structured diagnoses. While not yet outperforming can augment workflows, care or educational Effective integration into practice requires adherence rigorous guidelines, thorough validation, ongoing professional oversight ensure safety reliability.

Language: Английский

Citations

1