Show Notes
Her relative said it was nothing. Two weeks of symptom history in ChatGPT said get checked. She was diagnosed at stage one.
In early 2023, Alvina Nadeem was 36, a mother of two young boys, and tired in a way she put down to burnout or perimenopause. She was already a heavy ChatGPT user through her work in change management, so she tried something she had never done: she described her symptoms to it. When the model refused to answer first-person health questions, she reframed herself as a clinical case for a medical resident and asked it to explain its reasoning. Ovarian cancer appeared as one of four possibilities. What followed — a post-it note of symptoms to present together, a GP who ordered imaging, a stage one diagnosis — is the starting point for a wider conversation about how patients actually use AI: to navigate the health system, to filter misinformation, and to carry their own story between specialists who cannot see each other's notes.
This episode describes one patient's experience and is not medical advice. If you have persistent symptoms, speak to a clinician.
GUEST
Alvina Nadeem — Professional engineer (P.Eng.), change management advisor and leadership coach; Board member, Ovarian Cancer Canada; Co-director, Cancer and Work; Coalition Priorité Cancer au Québec
Host: Tjaša Zajc
THE AGENTIC PATIENT
A Faces of Digital Health series on how patients and caregivers actually use AI — which tools, which prompts, which guardrails — and what that does to the relationship with clinicians.
Series hub: https://www.facesofdigitalhealth.com/agentic-patient
WHAT THE CONVERSATION COVERS
• "No-lution": why dismissing patient AI use hides how people already use it
• Shadow AI in healthcare: patients were using ChatGPT long before health systems noticed
• Using ChatGPT for health questions in 2023, when the model refused first-person medical queries
• Role-play prompting: presenting yourself as a clinical case and asking the model to show its reasoning
• Prompting for possibilities rather than confirmation — and why that limits sycophancy
• Ovarian cancer symptoms: bloating, early satiety, fatigue, cramping and abnormal bleeding as a constellation
• Why ovarian cancer is diagnosed late, and why age 36 put her outside anyone's differential
• How to describe symptoms to your doctor so they lead to the right tests
• CA-125 testing and the routine pelvic pain workup
• Sharing health data with ChatGPT: privacy, identifiers and the risk-reward of a crisis
• AI as a health system navigation tool: which specialist, which questions, is this wait normal
• The worried well, wellness misinformation and AI as a filter
• Practical advice: approach the model as a tester, and end every session with questions for your doctor
• Emergency versus chronic illness: where general chatbots lose context
• Patient-built AI tools for specific conditions
• Continuity of care: the patient as the relay between family doctor and oncologist
CHAPTERS
03:00 The Agentic Patient: how patients use AI in their care
04:14 "No-lution" and the shadow AI already inside healthcare
07:38 ChatGPT in early 2023: a cautious, restricted tool
10:10 Learning to prompt: role, context and the role-play workaround
16:54 Challenging the model: sycophancy then and now
19:36 Presenting a symptom constellation to get the right tests
25:09 The post-it note, the pelvic exam and the imaging referral
27:49 CA-125 and the routine pelvic pain workup
29:00 What patients use AI for: health system navigation
33:31 The worried well, misinformation and AI as a filter
37:21 Practical advice: test the tool, end with questions for your doctor
39:45 Emergency versus chronic: where general chatbots fall short
46:28 AI as the keeper of a patient's story between specialists
MENTIONED
Ovarian Cancer Canada • Cancer and Work • Coalition Priorité Cancer au Québec • CA-125 (cancer antigen 125) test • ChatGPT
Discussion with Russ Read Barrow: https://podcasts.apple.com
More from Faces of Digital Health
Three in ten AI pilots at Sheba Medical Center never scale. Eyal Zimlichman on why, and what Europe's regulation holds back.
When prevention works, nothing happens. That's why it gets cut. Joanna Lane on financing prevention as long-term investment.
Nearly €200m, twenty years of EU health data projects. Henrique Martins on what they built and what EHDS needs next.
One ultrasound tech, four scanning rooms. NVIDIA's David Niewolny on what "augmented autonomy" asks of clinicians.
Most interoperability debates focus on exchange formats. In this discussion with Dr Nils Hellrung, medical informatician and CEO Strategy & Operations at vitagroup, Nils focuses on the need for shaping hospital IT with governance and financing.