OpenAI continues its push into healthcare with the launch of ChatGPT Health, a new feature that connects its artificial intelligence chatbot with users' medical records and wellness apps for more personalized answers to medical questions.
People already are using publicly available AI chatbots to ask healthcare-related questions. More than 800 million regular users of ChatGPT, 1 in 4, submits a prompt about healthcare every week, according to OpenAI. More than 40 million turn to ChatGPT every day with healthcare questions, according to an OpenAI report.
OpenaI says ChatGPT Health builds on this so the AI chatbot's responses are informed by users' health information and context, the company said in an announcement.
Users can now securely connect medical records and wellness apps—like Apple Health, Function and MyFitnessPal—so ChatGPT can help them understand recent test results, prepare for appointments with their doctor, get advice on how to approach diet and workout routines or understand the trade-offs of different insurance options based on healthcare patterns, the company said.
The new feature has additional, layered protections designed specifically for health, including purpose-built encryption and isolation to keep health conversations protected and compartmentalized, OpenAI said. Conversations in Health are not used to train OpenAI's foundation models, the company said.
The company said it was designed in close collaboration with physicians. ChatGPT Health is designed to support, not replace, medical care, and it is not intended for diagnosis or treatment, the company said.
OpenAI will initially provide access to a small group of early users—users with ChatGPT Free, Go, Plus and Pro plans outside of the European Economic Area, Switzerland and the U.K. are eligible. The company plans to expand access and make Health available to all users on web and iOS in the coming weeks.
Medical record integrations and some apps are available in the U.S. only, and connecting Apple Health requires iOS.
OpenAI tapped digital health company b.well Connected Health to power the underlying health data connectivity infrastructure.
Through this collaboration, users can explicitly authorize access within ChatGPT to securely connect their health data and medical records from U.S. healthcare providers. B.well operates a health data network designed to support consumer-mediated access to live clinical data across more than 2.2 million providers and 320 health plans, labs and other sources. Built on FHIR-based application programming interfaces (APIs) and trusted healthcare exchange frameworks, the network maintains strict controls around identity verification, consent management, data security and auditability, according to the company.
"One of the most common and meaningful ways people use ChatGPT today is for health and wellness information," said Ashley Alexander, vice president of health products at OpenAI, in a statement. "By working with trusted brands like b.well, we're able to make the user experience even more useful and personalized."
"Consumers are more and more often using [large language models (LLMs)] to interact with their healthcare, anywhere from 5% to 30% of queries every day are health- or healthcare-related. As consumers are starting to interact and engage with their health in a way that is comfortable to them, and I would argue, the natural digital front door, we wanted them to have a seamless and easy way to integrate their medical records, which they are entitled to under the information blocking and interoperability rules," Kristen Valdes, CEO and founder of b.well, told Fierce Healthcare in an interview.
"We're incredibly excited that consumers will now be able to have a secure and private mechanism where they can choose to connect their medical records to get more contextually aware and relevant information as they start to query these LLMs for their health," Valdes said. "We believe that consumers are educated and smart and need to be empowered to seek out answers to questions that they have, or even to help prepare them for upcoming doctors' appointments and visits where they can really engage in healthcare outside of just the four walls of a doctor's office."
OpenAI is leveraging b.well's SDK for Health AI, the first SDK (software development kit) built to power healthcare AI assistants.
"We created a semantic interoperability layer that we call the Data Refinery that lets users' data be more usable by AI," Valdes said.
A key component of the SDK is b.well's proprietary 13-step Data Refinery, which cleanses, reconciles, standardizes, summarizes, embeds and enriches fragmented data. The Health AI SDK unifies real-world health information into a single AI-ready dataset. The SDK supports secure use of both structured and unstructured clinical data—including patient summaries, lab and vital trends, clinical notes such as progress notes and discharge summaries, diagnostic interpretations, care plans and coverage information, according to the company. This approach allows AI systems to generate answers in seconds, even when working with years of longitudinal patient records.
B.well's Health AI SDK is HITRUST-certified and aligned with Centers for Medicare & Medicaid Services (CMS) and National Institute of Standards and Technology guidelines for responsible AI, the company said.
"Today, AI is only able to work across a small part of the patient record. The b.well Health SDK for AI finally enables AI to work across the full patient record—spanning multiple EMRs and with complete access to both structured data and unstructured data," said Imran Qureshi, chief AI and technology officer at b.well.
To connect medical records or wellness app data, users can log into those applications or their patient portals to integrate that information.
"In the future, as the new CMS Aligned Networks go live and organizations also turn on the network exchange, we will be able to use their biometric identity through CLEAR, who is a deep partner of ours, for us to use something called record location services, where we will be able to allow them to find their data more seamlessly, and using their identity be able to pull their information back," Valdes said.
LLMs and AI chatbots are the "natural digital front door" to help patients get answers to medical questions, organize health information and make medical information more understandable.
"For decades, every healthcare player has wanted to control and be the primary relationship holder with a patient, but that's really not feasible and has created this massive fragmentation problem," Valdes said. "It is secure in that is consumer-consented and controlled. No information is shared without a consumer actively consenting to share that information and for whatever purpose they choose. It really, truly puts the consumer in charge. And, LLMs have these unique opportunities of being able to surface experiences to consumers that now can be contextually relevant with the consumers' consent, things like shopability or the ability to find an appointment with an appropriate doctor. I think we're at the very beginning of being able to give people contextually aware interactions with their healthcare."
She added, "I think all LLMs have this amazing potential to truly, finally deliver on the single launching-off point for a person's healthcare that they're really looking for and have not had."
Emarketer healthcare analyst Rajiv Leventhal noted that the launch of ChatGPT Health marks OpenAI’s shift from positioning ChatGPT as a resourceful health information tool to one that encourages users to upload sensitive medical record data.
"Other tech giants, most notably Apple with Health Records, have developed products that link consumer health apps to personal data, but efforts have largely stalled due to fragmented access to patient records and privacy concerns. ChatGPT Health could be different because it has a built-in user base of 800 million global weekly active users, 25% of whom submit a prompt about healthcare each week," Leventhal said in a statement. "Lots of these folks use AI for quick answers to their health questions on symptoms and medical conditions—but whether they’re willing to share more personal health data will determine ChatGPT Health’s success."
With publicly available LLMs and AI chatbots, there are ongoing concerns about privacy and security, especially with sensitive healthcare information.
The Center for Democracy and Technology's Andrew Crawford noted that the U.S. doesn’t have a general-purpose privacy law, and the Health Insurance Portability and Accountability Act (HIPAA) only protects data held by certain people like healthcare providers and insurance companies.
"AI companies, along with the developers and companies behind health and apps, are typically not covered by HIPAA. The recent announcement by OpenAI introducing ChatGPT Health means that a number of companies not bound by HIPAA’s privacy protections will be collecting, sharing, and using peoples’ health data. And since it's up to each company to set the rules for how health data is collected, used, shared, and stored, inadequate data protections and policies can put sensitive health information in real danger," Crawford said in a statement.
"While OpenAI says that it won’t use information shared with ChatGPT Health in other chats, AI companies are leaning hard into personalization as a value proposition. Especially as OpenAI moves to explore advertising as a business model, it’s crucial that separation between this sort of health data and memories that ChatGPT captures from other conversations is airtight," he said.