ChatGPT provides mostly accurate responses about cardiovascular disease

ChatGPT, a popular online artificial intelligence (AI) tool, provides largely appropriate responses to simple cardiovascular disease (CVD) prevention questions, according to a research letter published online Feb. 3 in the Journal of the American Medical Association.

Ashish Sarraju, M.D., from the Cleveland Clinic, and colleagues evaluated the appropriateness of AI model responses to simple, fundamental CVD prevention questions. The 25 questions addressed fundamental preventive concepts, including risk factor counseling, test results, and medication information, based on guideline-based prevention topics. Responses were graded on a patient-facing information platform (like a website) and as draft responses to electronic message questions.

The researchers found that AI model responses to 21 of 25 questions (84 percent) were graded as appropriate in both contexts, but four responses had inappropriate content.

For example, the AI model responded to exercise questions by firmly recommending both cardiovascular activity and lifting weights, which may not be appropriate for certain patients. Furthermore, responses interpreting a low-density lipoprotein cholesterol level of 200 mg/dL lacked pertinent details, including details about familial hypercholesterolemia and genetic factors. None of the AI responses were considered unreliable.

“Findings suggest the potential of interactive AI to assist clinical workflows by augmenting patient education and patient-clinician communication around common CVD prevention queries,” the authors write.

Two authors disclosed financial ties to the pharmaceutical industry.

More information:
Ashish Sarraju et al, Appropriateness of Cardiovascular Disease Prevention Recommendations Obtained From a Popular Online Chat-Based Artificial Intelligence Model, JAMA (2023). DOI: 10.1001/jama.2023.1044

Journal information:
Journal of the American Medical Association

Source: Read Full Article