Skip to content 
On 23 February 2026, researchers at Mount Sinai published a study in Nature Medicine showing that ChatGPT Health failed to direct users to emergency care in 52% of gold-standard medical emergencies.
A few weeks earlier, STRAT7 Health conducted research with 400 consumers and 347 primary care physicians across the US and Europe to understand perceptions and anticipated use of ChatGPT Health. We found that 45.5% of consumers plan to use ChatGPT Health when it is available to them. Most said they plan to use it within six months of launch.
There is a growing tension between consumer demand for AI health tools and emerging evidence of their limitations. This presents both a commercial opportunity and clinical risk for pharmaceutical companies navigating this new paradigm.
The challenge for pharma is no longer whether patients will use AI, but how these behaviours influence healthcare dynamics and decisions.
Key statistics
- More than 60% of both consumers and primary care physicians were aware of ChatGPT Health to some extent before its full launch.
- 5% of consumers plan to use ChatGPT Health when it is fully available.
- 5% of consumers expect to be using ChatGPT Health within the first month of full launch, 47% within six months.
- 41% of consumers expect to use ChatGPT Health to understand medical notes or test results, the most-cited use.
- 28% of consumers expect to use ChatGPT Health to seek a diagnosis.
- 35% of primary care physicians are uncomfortable or very uncomfortable with their patients using ChatGPT Health.
- 52% of gold-standard medical emergencies were not directed to emergency care by ChatGPT Health in the Mount Sinai study (Nature Medicine, 24 February 2026).
The rise of the AI-primed patient
The use of AI in healthcare is increasing dramatically, with growing numbers of patients turning to AI-powered tools to access health information, understand symptoms, and support healthcare decisions.
The use of AI is reshaping when patients seek support from healthcare professionals, the assumptions they bring into consultations, and the discussions that follow. Healthcare professionals are already encountering the AI-primed patient – individuals whose understanding, expectations, and healthcare decisions have been influenced by AI before they enter the consultation room.
This trend is only accelerating, positioning AI as an increasingly influential stakeholder in the patient journey and a growing force in shaping healthcare decisions.
More than 60% of consumers and PCPs in our study told us they were aware of ChatGPT Health. Whilst the depth of awareness varies, the key takeaway is that a tool that has only been soft-launched to a limited US cohort is already widely on consumers’ radar.
Adoption is likely to be rapid. 21.5% of consumers expect to use it within the first month, 47% within six months.
Will consumers use ChatGPT Health the way OpenAI intends?
ChatGPT Health markets itself as a wellness companion, not a substitute doctor. The product is not designed, on its surface, to offer a diagnosis or recommend a treatment.
Consumers describe something different. 28% of consumers in our study expect to use ChatGPT Health to seek a diagnosis. 45% to identify when to consult a healthcare professional. 47% to prepare for an appointment. 35% to support a treatment decision. Understanding medical notes or test results, the most-cited use, sits at 41%.
That is well past ‘wellness companion’ and steadily moving into territory traditionally occupied by healthcare professionals.
Unlike traditional search engines, AI tools provide personalised, authoritative-sounding answers that can shape patient expectations before the consultation even begins.
PCPs are still working out where AI fits
35% of the primary care physicians in our study told us they were uncomfortable or very uncomfortable with the idea of their patients using ChatGPT Health. While some clinicians have strong views, many are still working out where AI fits into patient care and the consultation process.
That creates an opportunity for pharma to provide practical support for clinicians encountering AI primed patients – with patient communication frameworks and tools, objection handlers, educational materials and byeond
How will the AI-primed patient impact the patient journey?
The first major shift is in when patients seek care. AI may bring some patients into the healthcare system sooner by highlighting symptoms that warrant attention. However it can also delay presentation by providing reassurance where further investigation is needed. Both outcomes have clinical consequences, yet neither is reflected in most patient-journey models today.
The second shift is around expectations. The AI-primed patient now arrives with a working hypothesis and, often, a treatment preference. That makes the PCP’s first job to manage an existing belief, before getting to the diagnosis. Most pharma communication assets are not built for that conversation.
The third shift is in the consultation dynamic. As patients arrive with pre-formed hypotheses and AI-shaped expectations, the consultation increasingly shifts toward shared decision-making.
What pharma should do now
For pharma brands, there are three key points that should be considered urgently:
- Help HCPs handle the AI-primed patient. Practical training, not theory. Built around the language and tools they can actually use in a fifteen-minute consultation.
- Remap the patient journey for your brand’s therapy area. Understanding how the steps within it, timelines, stakeholders and dynamics may shift.
- Build credibility beyond your owned channels. AI tools increasingly favour information that appears authoritative, consistent, and well-supported across the wider digital ecosystem. Building credibility beyond your owned channels is therefore becoming a prerequisite for relevance.
The AI-primed patient is already becoming a feature of clinical practice. Brands that adapt early will be better positioned to engage both AI-influenced patients and the healthcare professionals involved in their care.
Be sure to read our follow up blog ‘How the patient journey is about to be redrawn’ where we reveal two things that pharma brands can do now to make a measurable difference quickly.
If you need help navigating these shifts and prioritising what your brand should do next, get in touch with the STRAT7 Health team.
About the author
- Dr Sarah Rosen
- Director, STRAT7 Jigsaw
Dr Sarah Rosen is a Director in the Health team at STRAT7 Incite, specialising in oncology, women’s health, and rare diseases. She helps pharmaceutical companies understand evolving patient and healthcare professional behaviours to inform strategy and decision-making. Her recent work focuses on the impact of AI in healthcare, exploring how it is reshaping diagnosis, treatment decisions, and the patient journey.
FAQs
What is ChatGPT Health?
ChatGPT Health is OpenAI’s health-focused product. It is positioned as helping people take a more active role in understanding their health, though our research finds consumers plan to use it for diagnosis and treatment decisions too.
When is ChatGPT Health launching fully?
ChatGPT Health has been soft-launched in the US to a small cohort as of early 2026. A wider full launch is expected within the next twelve months. Awareness is already above 60% among both consumers and primary care physicians, ahead of access.
What is an ‘AI-primed patient’?
An AI-primed patient is one who arrives at the consultation having already used an AI tool to form a hypothesis about their condition, look up treatments, or prepare questions.
Where does STRAT7’s data come from?
The figures in this post come from the paired study STRAT7 Health ran in February 2026. The Mount Sinai figure is from a paper in Nature Medicine, published 23 February 2026.