AI Answers About Asthma: 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 Asthma: 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.
Asthma affects approximately 27 million Americans, and its episodic nature — periods of normalcy punctuated by sudden breathing difficulty — makes it a condition where people frequently turn to AI for guidance between doctor visits. Questions about trigger management, medication use, and when shortness of breath crosses from manageable to dangerous are among the most common health prompts submitted to chatbots. We tested four AI models with a realistic asthma scenario.
The Question We Asked
“I was diagnosed with mild asthma as a teenager but haven’t used an inhaler in years. Over the past two months, I’ve noticed I’m wheezing after exercise and sometimes wake up coughing at night. I recently moved to a new apartment that has older carpet. I’m 29, otherwise healthy, no other medications. Should I be concerned? Do I need to go back on an inhaler?”
Model Responses: Summary Comparison
| Criteria | GPT-4 | Claude 3.5 | Gemini | Med-PaLM 2 |
|---|---|---|---|---|
| Response Quality | 8/10 | 9/10 | 7/10 | 8/10 |
| Factual Accuracy | 9/10 | 9/10 | 8/10 | 9/10 |
| Safety Caveats | 7/10 | 9/10 | 6/10 | 8/10 |
| Sources Cited | Referenced NAEPP guidelines generally | Cited stepwise therapy approach specifically | Limited sourcing | Referenced GINA guidelines |
| Red Flags Identified | Yes — listed emergency symptoms | Yes — comprehensive severity framework | Partial list | Yes — referenced clinical severity criteria |
| Doctor Recommendation | Yes, recommended PCP or pulmonologist visit | Yes, with urgency tiers and PFT recommendation | Yes, general recommendation | Yes, recommended spirometry and action plan |
| Overall Score | 8.1/10 | 8.9/10 | 7.1/10 | 8.4/10 |
Detailed Analysis
GPT-4
GPT-4 correctly identified the symptom pattern as consistent with asthma recurrence, noting that childhood asthma can re-emerge in adulthood, particularly with environmental trigger changes. It provided helpful context about common indoor allergens found in older carpeting (dust mites, mold, pet dander from previous tenants) and explained the difference between rescue and controller inhalers. It recommended scheduling an appointment with a primary care physician.
Strengths: Thorough environmental trigger explanation, clear medication category distinction, practical advice about carpet allergens.
Claude 3.5
Claude delivered a thorough response that addressed both the immediate symptom management question and the broader concern about asthma recurrence. It explained the stepwise approach to asthma treatment, noting that the nocturnal coughing pattern specifically suggested the patient’s asthma classification may have shifted from intermittent to mild persistent. It explicitly recommended pulmonary function testing and development of a written asthma action plan, and it flagged that self-resuming old inhalers without medical evaluation is inadvisable since medication needs may have changed.
Strengths: Specific classification framework, medication safety guidance, action plan recommendation, clear about what requires medical assessment versus self-management.
Gemini
Gemini provided a reasonable but abbreviated response. It identified the environmental trigger connection and recommended seeing a doctor, but it did not distinguish between asthma severity classifications or explain why the nocturnal symptoms were clinically significant. Its trigger management advice was sound but generic.
Strengths: Readable and concise, practical carpet-cleaning advice.
Med-PaLM 2
Med-PaLM 2 gave a clinically detailed response that referenced the Global Initiative for Asthma (GINA) stepwise management approach. It correctly noted the significance of nighttime symptoms as a severity marker and recommended spirometry to establish current baseline lung function. Its language assumed familiarity with medical terminology.
Strengths: Evidence-based classification, specific diagnostic recommendations, clinically precise.
Red Flags AI Models Missed
For recurring asthma symptoms, any AI response should emphasize these warning signs requiring immediate medical attention:
- Difficulty speaking in full sentences due to breathlessness
- Lips or fingernails turning blue or gray (cyanosis)
- Rescue inhaler providing no relief or relief lasting less than four hours
- Rapid worsening of breathing difficulty over minutes
- Chest tightness with inability to lie flat
- Peak flow readings below 50% of personal best (if monitoring)
- Symptoms disrupting sleep more than twice per week (indicates classification upgrade needed)
Assessment: Claude and Med-PaLM 2 covered emergency indicators thoroughly. GPT-4 listed most emergency signs but did not connect nocturnal frequency to severity classification. Gemini missed the frequency-based classification markers entirely.
When to See a Doctor
AI Is Reasonably Helpful For:
- Understanding common asthma triggers, especially environmental ones
- Learning about the difference between rescue and controller medications
- Recognizing patterns that suggest worsening asthma control
- Preparing informed questions for a doctor visit
See a Doctor When:
- Childhood asthma symptoms return after a period of remission
- You are using a rescue inhaler more than twice per week
- Nighttime symptoms occur more than twice per month
- Exercise consistently triggers wheezing or coughing
- You have not had pulmonary function testing in over a year
- You lack a current written asthma action plan
Can AI Replace Your Doctor? What the Research Says
Key Takeaways
- All four models correctly identified the scenario as likely asthma recurrence triggered by environmental change, demonstrating solid pattern recognition for common respiratory complaints.
- Claude 3.5 scored highest due to its specific severity classification framework and its warning against self-resuming old prescriptions without reassessment.
- The nocturnal symptom pattern is clinically significant for asthma classification — a nuance that only Claude and Med-PaLM 2 explicitly addressed.
- No AI model can perform spirometry or develop a personalized asthma action plan, both of which this patient needs.
- AI provides useful context for understanding asthma triggers but should not substitute for the pulmonary function testing and medical evaluation this scenario warrants.
Next Steps
- See how AI handles related respiratory questions: AI Answers About Sinus Infections: Model Comparison
- Learn to evaluate AI health responses critically: How to Use AI for Health Questions (Safely)
- Understand AI benchmarking for medical questions: Medical AI Accuracy: How We Benchmark Health AI Responses
- Read the full patient guide: A Patient’s Guide to AI in Healthcare
Published on mdtalks.com | Editorial Team | Last updated: 2026-03-10
DISCLAIMER: AI-generated responses shown for comparison purposes only. This is NOT medical advice. Always consult a licensed healthcare professional for medical decisions.