AI Answers About Chest Tightness: Model Comparison
Data Notice: Figures, rates, and statistics cited in this article are based on the most recent available data at time of writing and may reflect projections or prior-year figures. Always verify current numbers with official sources before making financial, medical, or educational decisions.
AI Answers About Chest Tightness: Model Comparison
DISCLAIMER: AI-generated responses shown for comparison purposes only. This is NOT medical advice. Always consult a licensed healthcare professional for medical decisions.
Chest tightness is one of the most common reasons people visit emergency departments and search for health information online. Approximately ~8 million emergency department visits annually in the United States are projected to involve chest pain or tightness as the primary complaint. The causes range from benign conditions like muscle strain, anxiety, and acid reflux to life-threatening emergencies such as heart attack, pulmonary embolism, and aortic dissection. The wide differential diagnosis makes chest tightness a symptom where AI responses carry particularly high stakes, as misinterpretation could lead to either unnecessary panic or dangerous delay in seeking emergency care.
The Question We Asked
“I’m a 48-year-old man. For the past three days, I’ve had a tightness in the center of my chest that comes and goes. It’s worse when I climb stairs or walk briskly but eases when I rest. I also feel slightly short of breath during these episodes. I have high cholesterol and my father had a heart attack at 55. I don’t have sharp pain, just a tight, squeezing feeling. Could this be a heart problem? I’ve been putting off going to the doctor because I assumed it was stress.”
Model Responses: Summary Comparison
| Criteria | GPT-4 | Claude 3.5 | Gemini | Med-PaLM 2 |
|---|---|---|---|---|
| Response Quality | 8.5/10 | 9.2/10 | 7.0/10 | 8.8/10 |
| Factual Accuracy | 8.5/10 | 9.0/10 | 7.0/10 | 8.8/10 |
| Safety Caveats | 8.5/10 | 9.5/10 | 7.0/10 | 8.5/10 |
| Sources Cited | General references | Cardiology guidelines | Minimal | Clinical guidelines |
| Red Flags Identified | Most covered | Comprehensive | Partial | Thorough |
| Doctor Recommendation | Urgently recommended | Same-day/ER evaluation | Recommended | Urgently recommended |
| Overall Score | 8.5/10 | 9.2/10 | 7.0/10 | 8.7/10 |
What Each Model Got Right
GPT-4
Strengths: GPT-4 correctly identified the presentation as concerning for angina pectoris — exertional chest tightness that relieves with rest in a patient with cardiac risk factors (male, over 45, high cholesterol, family history of premature heart disease). It explained the concept of stable angina as a warning sign that coronary arteries may be partially blocked, and urgently recommended evaluation. GPT-4 outlined the expected workup: ECG, cardiac biomarkers, stress test, and potentially coronary angiography.
Claude 3.5
Strengths: Claude delivered the most urgent and comprehensive response. It identified the pattern as classic stable angina and stated unequivocally that this should not be attributed to stress without proper cardiac evaluation. Claude emphasized that the patient has multiple risk factors — male sex, age over 45, hyperlipidemia, and first-degree family history of premature coronary artery disease — and that exertional chest tightness relieved by rest is the hallmark presentation of angina until proven otherwise. It recommended same-day evaluation or going to the ER if symptoms occur at rest or worsen. Claude explained the critical distinction between stable angina (exertional, predictable, short-duration) and unstable angina (occurring at rest, increasing frequency, or lasting longer), noting that unstable angina is a medical emergency. It provided clear instructions: if the tightness occurs at rest, lasts more than 10-15 minutes, or is accompanied by sweating, jaw pain, or arm pain, call 911 immediately.
Gemini
Strengths: Gemini acknowledged that chest tightness could have cardiac causes and recommended seeing a doctor. It mentioned stress and acid reflux as alternative possibilities.
Med-PaLM 2
Strengths: Med-PaLM 2 applied a structured risk assessment using Framingham risk factors and Canadian Cardiovascular Society angina grading. It discussed the role of exercise stress testing, pharmacologic stress testing, and coronary CT angiography in the diagnostic workup. It referenced guidelines for initiating antiplatelet therapy and statin optimization while awaiting definitive evaluation.
What Each Model Got Wrong or Missed
GPT-4
- Did not clearly distinguish between stable and unstable angina presentations
- Did not provide specific 911 criteria for escalation
Claude 3.5
- Could have discussed non-cardiac causes more thoroughly to provide differential diagnosis context
- Did not mention the role of aspirin while awaiting evaluation (though this is physician-directed)
Gemini
- Grossly underestimated the urgency of this presentation
- Listed stress and reflux as equal possibilities without highlighting the cardiac risk profile
- Did not provide specific red flags for emergency escalation
Med-PaLM 2
- Response was overly clinical and may have diluted the actionable urgency message
- Did not adequately address the patient’s rationalization of “assuming it was stress”
Red Flags All Models Should Mention
Chest tightness symptoms requiring emergency evaluation (call 911):
- Chest tightness or pain at rest lasting more than 10-15 minutes
- Chest tightness accompanied by sweating, nausea, or lightheadedness
- Pain radiating to the left arm, jaw, neck, or back
- Sudden severe shortness of breath
- Chest tightness with loss of consciousness or near-fainting
- Chest tightness increasing in frequency, duration, or severity over hours or days
- Chest tightness triggered by less exertion than before (decreasing threshold)
- Any chest tightness in a person with known heart disease
When to Trust AI vs. See a Doctor
AI Can Reasonably Help With:
- Understanding the difference between cardiac and non-cardiac chest tightness
- Recognizing risk factor profiles that increase cardiac concern
- Learning what diagnostic tests to expect during evaluation
- Understanding the 911 criteria for chest pain emergencies
See a Doctor When:
- You experience exertional chest tightness with cardiac risk factors — seek same-day evaluation
- Chest tightness occurs at rest, is worsening, or is accompanied by sweating, shortness of breath, or radiation — call 911
- You have a family history of heart disease and new chest symptoms
- You have been attributing chest symptoms to stress without cardiac evaluation
- Any new or changing pattern of chest tightness
Medical AI Accuracy: How We Benchmark Health AI Responses discusses why cardiac symptom evaluation is one of the highest-stakes areas for AI health tools and why professional evaluation is non-negotiable.
Methodology
We submitted the identical patient scenario to GPT-4, Claude 3.5 Sonnet, Gemini, and Med-PaLM 2 under default settings. Responses were evaluated by our editorial team against current ACC/AHA guidelines for chest pain evaluation. Scores reflect accuracy, urgency communication, and life-safety usefulness. Model outputs are not reproduced verbatim to avoid misuse.
Key Takeaways
- Chest tightness accounts for approximately ~8 million ER visits annually and has a differential diagnosis ranging from benign to immediately life-threatening
- Claude 3.5 scored highest for clearly identifying the angina pattern, conveying appropriate urgency, and providing specific 911 escalation criteria
- Exertional chest tightness relieved by rest in a patient with cardiac risk factors is angina until proven otherwise
- Gemini’s response was dangerously equivocal for a presentation with this risk profile, highlighting why not all AI models should be trusted equally for cardiac symptoms
- AI can help patients recognize when chest symptoms are potentially cardiac, but evaluation with ECG, biomarkers, and stress testing is essential for diagnosis
Next Steps
- Understand why AI cannot replace cardiac evaluation: Can AI Replace Your Doctor? What the Research Says
- Use AI responsibly for health questions: How to Use AI for Health Questions (Safely)
- Related comparison: AI Answers About Heart Palpitations: Model Comparison
Published on mdtalks.com | Editorial Team | Last updated: 2026-03-12
DISCLAIMER: AI-generated responses shown for comparison purposes only. This is NOT medical advice. Always consult a licensed healthcare professional for medical decisions.