Accuracy of Online Artificial Intelligence Models in Primary Care Settings.

INTRODUCTION: The importance of preventive medicine and primary care in the sphere of public health is expanding, yet a gap exists in the utilization of recommended medical services. As patients increasingly turn to online resources for supplementary advice, the role of artificial intelligence (AI) in providing accurate and reliable information has emerged. The present study aimed to assess ChatGPT-4's and Google Bard's capacity to deliver accurate recommendations in preventive medicine and primary care. METHODS: Fifty-six questions were formulated and presented to ChatGPT-4 in June 2023 and Google Bard in October 2023, and the responses were independently reviewed by two physicians, with each answer being classified as "accurate," "inaccurate," or "accurate with missing information." Disagreements were resolved by a third physician. RESULTS: Initial inter-reviewer agreement on grading was substantial (Cohen's Kappa was 0.76, 95%CI [0.61-0.90] for ChatGPT-4 and 0.89, 95%CI [0.79-0.99] for Bard). After reaching a consensus, 28.6% of ChatGPT-4-generated answers were deemed accurate, 28.6% inaccurate, and 42.8% accurate with missing information. In comparison, 53.6% of Bard-generated answers were deemed accurate, 17.8% inaccurate, and 28.6% accurate with missing information. Responses to CDC and immunization-related questions showed notable inaccuracies (80%) in both models. CONCLUSIONS: ChatGPT-4 and Bard demonstrated potential in offering accurate information in preventive care. It also brought to light the critical need for regular updates, particularly in the rapidly evolving areas of medicine. A significant proportion of the AI models' responses were deemed "accurate with missing information," emphasizing the importance of viewing AI tools as complementary resources when seeking medical information.

Smartphone AF Burden During the Blanking Period Predicts Catheter Ablation Outcomes: Insights From DECAAF II.

BACKGROUND: Atrial fibrillation (AF) recurrence during the blanking period is under investigated. With the rise of smartphone-based electrocardiogram (ECG) monitoring, there's potential for better prediction and understanding of AF recurrence trends. OBJECTIVES: In this study the authors hypothesize that AF burden derived from a single-lead Smartphone ECG during the blanking period predicts recurrence of atrial arrhythmias after ablation. METHODS: 630 patients with persistent AF undergoing ablation were included from the DECAAF II (Effect of MRI-Guided Fibrosis Ablation vs Conventional Catheter Ablation on Atrial Arrhythmia Recurrence in Patients With Persistent Atrial Fibrillation) trial. Patients recorded daily ECG strips using a smartphone device. AF burden was defined as the ratio of ECG strips with AF to the total number of strips submitted. The primary outcome was the recurrence of atrial arrhythmia. RESULTS: Recurrence occurred in 301 patients during the 18-month follow-up period. In patients who developed recurrent arrhythmia after 90 days of follow-up, AF burden during the blanking period was significantly higher when compared with patients who remained in sinus rhythm (31.3% vs 7.5%; P < 0.001). AF burden during the blanking period was an independent predictor of arrhythmia recurrence (HR: 1.41; 95% CI: 1.36-1.47; P < 0.001). Through grid searching, an AF burden of 18% best discriminates between recurrence and no-recurrence groups, yielding a C-index of 0.748. After a follow-up period of 18 months, recurrence occurred in 33.7% of patients (147 of 436) with an AF burden <18% and in 79.4% of patients (154 of 194) with an AF burden >18% (HR: 4.57; 95% CI: 3.63-5.75; P < 0.001). CONCLUSIONS: A high AF burden derived from a smartphone ECG during the blanking period is a strong predictor of atrial arrhythmia recurrences after ablation.