Kidney Week

Abstract: SA-PO010

Role of Artificial Intelligence (AI) in Clinical Interpretation of 24-Hour Ambulatory Blood Pressure Monitoring

Session Information

• Augmented Intelligence, Large Language Models, and Digital Health
  October 26, 2024 | Location: Exhibit Hall, Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: Augmented Intelligence, Digital Health, and Data Science

• 300 Augmented Intelligence, Digital Health, and Data Science

Authors

• Alam, Sreyoshi Fatima, Mayo Clinic Minnesota, Rochester, Minnesota, United States

Sreyoshi Fatima Alam, MBBS

• Gonzalez Suarez, Maria Lourdes, Mayo Clinic Minnesota, Rochester, Minnesota, United States

Maria Lourdes Gonzalez Suarez, MD, PhD, FASN

• Thongprayoon, Charat, Mayo Clinic Minnesota, Rochester, Minnesota, United States

Charat Thongprayoon, MD, FASN

• Miao, Jing, Mayo Clinic Minnesota, Rochester, Minnesota, United States

Jing Miao, MD, PhD, FASN

• Sheikh, M. Salman, Mayo Clinic Minnesota, Rochester, Minnesota, United States

M. Salman Sheikh, MD

• Garcia Valencia, Oscar Alejandro, Mayo Clinic Minnesota, Rochester, Minnesota, United States

Oscar Alejandro Garcia Valencia, MD

• Schwartz, Gary L., Mayo Clinic Minnesota, Rochester, Minnesota, United States

Gary L. Schwartz, MD, FASN

• Craici, Iasmina, Mayo Clinic Minnesota, Rochester, Minnesota, United States

Iasmina Craici, MD

• Cheungpasitporn, Wisit, Mayo Clinic Minnesota, Rochester, Minnesota, United States

Wisit Cheungpasitporn, MD, FASN

Background

The utility of AI in interpreting ambulatory blood pressure monitoring (ABPM) data is increasingly recognized. Evaluating the accuracy of AI models, like ChatGPT 4.0, in clinical settings can inform their integration into healthcare processes. However, limited research has been conducted to validate the performance of such models against expert interpretations in real clinical scenarios.

Methods

This study assessed the performance of ChatGPT 4.0 in interpreting 24-hour ABPM records from 54 cases at Mayo Clinic, Minnesota, in March 2024. The AI's interpretations were compared with results confirmed by two nephrologists, who reached a consensus on the presence or absence of specific conditions based on the American College of Cardiology/American Heart Association (ACC/AHA) guidelines.

Results

ChatGPT 4.0 demonstrated varied accuracy across different conditions: Hypertension (84.91%), Hypotension (94.34%), Nocturnal Hypertension (81.13%), Normal Nocturnal Dip (75.47%), Normal Heart Rate (90.57%), Tachycardia (73.58%), and Bradycardia (71.70%). These outcomes highlight the model's capability to reliably interpret complex clinical data with considerable accuracy. The model's performance in identifying tachycardia and bradycardia, however, was comparatively lower, suggesting room for improvement in these areas.

Conclusion

The findings suggest that ChatGPT 4.0 has the potential to assist in the clinical interpretation of 24-hour ABPM. However, the model's current performance indicates that further improvements in accuracy are necessary before it can be effectively incorporated into clinical practice. Ongoing advancements and training on diverse datasets could enhance the model's diagnostic precision, making AI a promising tool for the routine analysis of ABPM data. While the results are encouraging, further research is needed to refine the model's performance and validate its utility across larger, more diverse patient populations.