Shauna O’Brien and her husband Julian Millikan are the kind of people the health system is supposed to work for: good jobs, great health insurance, a solid hospital nearby. But the weeks before the holiday break, they still found themselves fighting for Julian’s life.
It started as a nagging pain low in Julian’s leg, near his Achilles and calf, the kind of ache that feels like a workout injury. He rested. The next week, the pain moved into his knee and got worse until he could not put weight on it. He hobbled on crutches and ended up sleeping on the couch because the stairs in their home were too much.
During their first visit to the ER, clinicians processed Julian quickly and sent him home with pain medication and a plan to wait a week or two before pursuing an MRI. The urgent care doctor they visited days later diagnosed Julian with a gout flare, prescribed gout medication, and again suggested waiting. The MRI she ordered, and which might have given them clarity, was immediately declined by insurance.
He began to feel nausea and he became so short of breath he couldn’t take more than two steps on crutches without stopping. With the visits to the ER and urgent care still in her mind, Shauna wondered if returning again seemed crazy, and calling 911 felt like a move that would scare the couple’s two young kids.
But just to be sure, Shauna typed Julian’s symptoms and the earlier diagnoses into ChatGPT, and Chat told her to “go back to the ER immediately,” warning that the pattern could indicate a septic joint infection or deep vein thrombosis (DVT), both medical emergencies. Paramedics arrived, took his vitals, and agreed something bigger was happening: elevated heart rate and low blood oxygen saturation.
This time, Shauna arrived at the ER armed with specific knowledge to raise at triage. In the crowded waiting room, Julian was fast-tracked. Ultrasound revealed a large DVT in his leg, and scans showed bilateral pulmonary emboli in his lungs, obstructing his breathing and straining his heart. Doctors told Shauna that Julian likely would not have survived another night at home. Either his heart or breathing would have stopped. Shauna thought: “I would have woken up that morning next to a dead body.” Julian was admitted to critical care and immediately started on oxygen and blood thinners.
During his four-day hospital stay, ChatGPT helped Shauna understand high-stakes treatment tradeoffs, interpret monitor readouts, and convince staff to put Julian back on supplemental oxygen when his levels dipped. As discharge approached, it helped her assemble the questions that clarified next steps.
Back home, ChatGPT helped the family plan Julian’s recovery within the constraints of staircases and his limited mobility, from sleep arrangements to what his monthslong healing timeline might look like. Even in a well-resourced community and supported by “really good insurance,” Shauna says “it’s still such a fight” to get lifesaving care.
