More in Health & Medical.
Decide to go back to the ER for worsening leg pain and breathlessness with ChatGPT
Beginner-friendlyNo logins, payments or coding needed. Low risk: paste the prompt and you should get something useful.
Copy the prompt below and paste it into the AI tool named on this page.
Prompt Recreated prompt
The source didn't share its exact prompt, so we wrote this from what it describes.
My husband's leg pain started near the calf and moved to the knee, and now there's nausea and so much shortness of breath that two steps on crutches are impossible. The ER sent my husband home and urgent care said gout. Should we go back to the ER?
Try this in:
Sends the Recreated prompt shown above. ChatGPT and Grok open with it filled in, and ChatGPT may send it straight away. Gemini opens empty, so we copy the prompt for you to paste. Every button also copies it.
Opens in a new tab. We don't see what you send.
These links are only a shortcut. We're not affiliated with OpenAI, Google or xAI, and your results may not match the source.
Optional stronger version
Our rewrite with extra checks. The poster didn't test it.
Before you start, ask me 1–3 short questions about the details below and wait for my answers. Once I've answered, confirm the plan in one line and run the whole task start to finish. Goal: Decide quickly whether someone's worsening symptoms need emergency care, even after earlier visits said it was minor. Questions to ask me first: - What are the symptoms now, and how have they changed since the last visit? - What did each doctor or clinic diagnose and prescribe? - Is the person having trouble breathing, chest pain or confusion right now? Inputs (fill in or confirm): - Symptoms and timeline: [SYMPTOMS / DATES] - Earlier visits and diagnoses: [VISITS] - Medicines taken: [MEDICINES] - Any readings we have (heart rate, oxygen, temperature): [READINGS] Rules: - If any symptom could be life-threatening, say first and plainly to call emergency services or go to the ER now. - Name the serious conditions the pattern could fit, in plain words, so I can raise them at triage. - Don't tell me it's safe to wait at home. - Give me a short summary I can read out to paramedics or the triage nurse. Checkpoint: Give me the go-now-or-not answer and the triage summary first, before anything else. If you get stuck: If you aren't sure, treat it as urgent and tell me to get the person seen now. Verify before you finish: Check that the triage summary includes the timeline, earlier diagnoses and current symptoms. Output: A clear urgency answer, a few serious conditions to mention and a short triage summary. Done when: The person is being seen by emergency care or a clinician has told us what to do.
AI tool
Outcome
Julian Millikan's leg pain had already been seen once in the ER and once at urgent care, where it was diagnosed as a gout flare, and the MRI that was ordered was declined by insurance. When Julian became nauseated and too short of breath to manage two steps on crutches, Julian's wife, Shauna O'Brien, typed the symptoms and earlier diagnoses into ChatGPT. It told O'Brien to 'go back to the ER immediately,' warning the pattern could mean deep vein thrombosis (DVT) or a joint infection. Paramedics found an elevated heart rate and low blood oxygen, Julian was fast-tracked at triage, and scans showed a large DVT and bilateral pulmonary emboli. Doctors said Julian likely would not have survived another night at home. During the four-day stay, ChatGPT also helped O'Brien understand treatment choices and prepare questions for discharge.
Evidence: Article or interviewA newsletter, journalist or interview account of what happened.
Source
- Posted
- by OpenAI Academy (story about Shauna O'Brien)
- Last checked
- Method:
- How we verify
Context: This story is published by OpenAI, which makes ChatGPT. Paramedics and ER doctors made the diagnosis; ChatGPT's role was urging the return visit and helping the family follow the care.
More in Health & Medical
Decode a scary dental insurance denial and confirm the real bill with Claude Code
John Sambrook, 66, woke at 4:30 a.m. to an Aetna Explanation of Benefits for some dental fillings: six of the eight line items said DENIED, and the 'Your Share' column added up to $1,316. Sambrook had spent about 15 minutes setting up a small Claude Code workspace for personal insurance, with the policy documents in it, and asked it to explain the EOB. About three minutes later, it showed that the plan excludes 'comprehensive dental services', so the denials were expected, but that the dentist is in network and bills at a contracted rate. The actual amount owed was $723.83, the figure already being paid, so nothing more was due.
PersonalHealth & medicalClaude CodeWorkedRecreated promptEvidence: Screenshots or logsView CaseNarrow down sudden vertigo to likely BPPV before an ENT visit with Grok
After a workout, a swim and a massage, founder Violetta Bonenkamp had extreme vertigo and nausea, and the paramedics who came noted high blood pressure. At home, Bonenkamp gave Grok the symptoms, blood pressure readings and Apple Watch stats. Grok put benign paroxysmal positional vertigo (BPPV) at the top, explained how a vertigo attack can raise blood pressure, and suggested the Dix-Hallpike test and a follow-up with an ENT specialist. The home test wasn't clear-cut and Grok sounded less sure, but the ENT confirmed BPPV and treated it, and it took two different maneuvers to get the vertigo under control.
PersonalHealth & medicalGrokWorkedRecreated promptEvidence: Described by posterView CaseFind out why two Medicare Advantage plans have different deductibles with ChatGPT
Comparing 2027 Medicare Advantage plans, John Major found two from one company with nearly identical features and premiums and copays within $10, but plan E had a $2,000 deductible on non-routine care and plan S had none. Major couldn't find a benefit to explain it and asked ChatGPT. It admitted it couldn't find an explanation either, but on its own it compared the 2026 versions: then, plan E's deductible was $1,500 with an out-of-pocket maximum $2,750 lower than plan S. Major concluded the company was 'apparently squeezing existing plan E customers', counting on them to renew without looking.
PersonalHealth & medicalChatGPTWorkedRecreated promptEvidence: Described by posterView CaseCompare first-time Medicare options with Gemini alongside a human advisor
Facing a first Medicare sign-up, Jeff Hall used AI (Hall calls it Gemini Notebook) to work out which options to choose, alongside an advisor from LeClair Group. The goal was to minimize out-of-pocket costs. Running simulations on the shortlist, Hall found the options 'almost net out the same or within a few hundred dollars of each other', and the results came out entirely different for Hall's spouse. Hall's takeaway: Medicare plans are like a balloon, and anyone who tries this alone 'is a fool'. The post doesn't say which plan Hall picked.
PersonalHealth & medicalGeminiWorkedRecreated promptEvidence: Described by posterView CaseAppeal a $600 insurance denial and file a state complaint with Claude (partial)
After a routine specialist visit turned into a denied claim, Brett Chereskin had over $600 at stake. Chereskin pasted the policy sections and the explanation of benefits (EOB) into Claude for plain English, built a dated timeline of every call and letter, and got a read of the appeal steps and protections that might apply ('Not legal advice'). Claude helped draft the appeal. When the insurer's internal appeal stalled, Chereskin and Claude drafted a formal complaint to the state department of insurance in 'a focused hour'. The complaint is filed, but 'I don't know yet how it resolves.'
PersonalHealth & medicalClaudePartialRecreated promptEvidence: Described by posterView CaseRead an 18-page blood test report with ChatGPT, then find what it missed
Simran Sukhnani gave an 18-page blood report to ChatGPT. Within seconds it flagged high triglycerides (252 mg/dL), low vitamin D (14.9 ng/mL) and blood in the urine. But it never asked about Sukhnani's life. It suggested walking 30 minutes a day, though Sukhnani walks 15,000 steps and does weight training five days a week. It recommended 60,000 IU of vitamin D weekly without asking about supplements already being taken. And it never asked whether Sukhnani was menstruating when giving the urine sample, though that was the case and is a common cause of blood in urine. Sukhnani went to a doctor, and calls ChatGPT 'a starting point, not a destination'.
PersonalHealth & medicalChatGPTPartialRecreated promptEvidence: Article or interviewView Case
Related cases
Cut four monthly bills by $119.75 with ChatGPT negotiation scripts
McKenzie Day dreaded calling service providers, so Day asked ChatGPT for a strategy for each bill, how to reach a real person fast and a script for negotiating, including what to say if the first offer wasn't good enough. Taking one bill a week, Day got home internet down from $109 to $50 a month (with more data, Peacock streaming and the rate locked for 5 years), switched phone providers to go from $48 to $15 a month for the first year ($30 after that), moved pest control to a DIY approach ($48 to $11 a month) and swapped a budgeting app for a self-hosted Actual Budget ($9 to $3 a month). In about a month that came to $119.75 a month, or $1,437 a year.
PersonalMoney & billsChatGPTWorkedImproved promptEvidence: Described by posterBeginner-friendlyView CaseNegotiate internet and phone bills down by $113.99 a month with ChatGPT call scripts
Allison, a 34-year-old paralegal paying about $612 a month across bills, tested ChatGPT bill negotiation for 30 days with Jake Read. For each call, ChatGPT wrote a 60-second script plus three fallback lines. Spectrum dropped the internet bill from $94.99 to $59.99 a month for 12 months, and AT&T moved two lines from $165 to $110 a month after a second call to the retention department. Geico didn't budge, and ChatGPT's estimate that the car insurance could drop to $140 a month was too optimistic. A gym credited three free months and an unused Disney+ was canceled. Total savings: $113.99 a month, or $1,367.88 a year, for 2 hours and 41 minutes of calls.
PersonalMoney & billsChatGPTPartialImproved promptEvidence: Described by posterView CaseFind a clogged catalytic converter on a road trip in Mexico with ChatGPT
Kevin and Ruth's vehicle, Igor, was losing power in a tiny town in Veracruz, Mexico, and the pollution air pump had started making odd whirring noises. Kevin input the vehicle statistics and symptoms into ChatGPT, which gave three possibilities: a clogged catalytic converter, the fuel pump or the MAP sensor. The fuel pressure read 40 psi and a swapped MAP sensor changed nothing, which left the converter. ChatGPT suggested removing the oxygen sensor in front of the converter so the exhaust could escape; the engine fired right up and ran normally. The mechanics hollowed out the converter instead of replacing it, the check engine light stayed on, and the bill was 1,700 pesos (about $98 US).
PersonalCarsChatGPTWorkedRecreated promptEvidence: Described by posterView Case
