Marcus Welby was a fictional doctor on television in the early 1970s. He made house calls and knew his patients by name, and presumably their medical history from memory. He represents the kind of doctor most patients want but few will ever experience unless they join an expensive concierge medical practice. What if you had a medical provider who knew your medical records intimately. What if, instead of having to re-explain, you are allergic to penicillin or that cancer runs in your family, it was already in your doctor’s knowledge base? More on that below.
The U.S. health care system is convoluted. The average wait time to see a primary care physician is about 31 days, on average. Face time with your doctor averages 10-to-15 minutes, but that can vary with age and insurance. During that encounter an exchange of information must occur between doctor and patient for the visit to have any value. Diagnostic tests need reviewed, patient experiences need to be heard and medical advice delivered – all in a dozen minutes or so. It is easy to understand why doctors who see 20 to 30 patients a day cannot also spend 10 minutes reviewing patient files before conducting a 10-to-15 minute office visit.
How are people filling in the information gap between physician visits? Data going back nearly 30 years shows patients flocking to the internet for information on diseases and conditions. Dr. Google is a euphemism for people using the search engine Google to query medical conditions. Nowadays patients have an even more powerful health information source: artificial chatbot ChatGPT Health. Rather than merely ask general questions about health and medicine, consumers upload their medical history into a personal ChatGPT Health account and ask specific questions about their own health and medical care. The following is from Health Affairs:
[M]ore than 40 million people use it daily for health care questions accounting for roughly 5 percent of all messages.
ChatGPT Health offers a new kind of convenience by allowing consumers to bring together data from previously siloed sources. This includes medical records (whose connection to the platform is supported by b. well’s “underlying health data connectivity infrastructure”), as well as Apple Health, Function, MyFitnessPal, and other consumer health companies. From there, users are encouraged to ask questions like, “Give me a summary of my overall health,” or “I have my annual physical tomorrow. What should I talk to my doctor about?”
As an aside, perhaps it would be good for your doctor to ask ChatGPT Health, “Give me a summary of Devon Herrick’s overall health. What should I ask him about during his physical?” Yet the author of a Health Affairs Forefront article sees dark clouds rather than a silver lining:
In DTC health care, interoperability has largely been nonexistent. ChatGPT Health could change that, making disparate consumer health products feel coherent. But consolidation may not be neutral.
Which consumer health companies are integrated, and which are not? If ChatGPT Health incorporates some consumer health companies while excluding others, it effectively becomes a gatekeeper, privileging certain companies’ data as more interpretable and useful. Over time, this could steer consumers toward a particular set of “recommended” products, creating a parallel consumer care ecosystem outside the traditional health care system.
Bias is entirely reasonable. Biohackers, for example, would likely want to use ChatGPT Health to enhance their unsanctioned use of research peptides [NAD+, BPC-157, TB-500, CJC-1295, Testamorelin, Sermorelin, PT-141. Etc.]. The author also wonders what will happen when ChaptGPT Health disagrees with one’s doctor on a recommended treatment.
ChatGPT Health intensifies these concerns. Patients may arrive with AI-generated summaries from multiple DTC platforms. This might be data their primary care physician did not order, may not trust, and may not have time to fully evaluate, but must nonetheless address. The result of these trends is not just a potentially more informed patient; it is also a clinician increasingly asked to interpret and contextualize results generated outside the traditional health care system.
When I began researching health information on the internet 30 years ago the medical establishment worried about patients reading bogus medical information and wasting the doctors’ time having to explain why it was wrong. That didn’t happen. Instead, patients arrived better informed.
Read more at Health Affairs: When ChatGPT Health Becomes The Health Record For Direct-To-Consumer Care
See also: Introducing ChatGPT Health | OpenAI