In the healthcare industry, it’s considered a given that change is slow. When you’re working in a large, fractured system that is heavily regulated, where maintaining privacy and security are critical, and where every change must be evaluated for its impact on patient outcomes, gradual, prolonged transformation is the expectation.
Clinicians can often be the pacesetters when it comes to implementing change. We’ve seen time and again that handing down legislation or technology does not automatically lead to broad adoption, especially when it’s done without taking into consideration the context of how it will be used or the ripple effects on patient care.
And yet, when it comes to the use of generative artificial intelligence (AI), physicians are surprisingly receptive to embracing this new technology. In fact, a recent survey my company conducted showed 40% of U.S. physicians are ready to use generative AI this year when interacting with patients at the point of care.
To be clear, AI and machine learning (ML) have been used in healthcare for years. What’s new is the rapid emergence of generative AI models such as ChatGPT that are trained to see patterns in data and generate new content. In the last year, institutions have published multiple studies evaluating how generative AI might be used in healthcare. For instance, researchers at Mass General Brigham recently published research on using a large language model to draft patient message responses. While physicians found the responses helpful in reducing their workload, the study demonstrated the importance of having a human in the loop, with a small percent of responses conveying inaccurate information if left unedited.
Healthcare providers are exploring other applications for generative AI as well. Almost 10% of attending physicians at Mass General Brigham are using generative AI to take notes on patient visits, for instance, and the results have been positive, with patients shocked that their doctors aren’t typing on laptops while engaged in conversations on their health.
Given the novelty of this technology, it’s not surprising that physician perspectives on it are evolving rapidly. Our survey found two-thirds of physicians have changed their opinion on generative AI in healthcare in the last year to see it as more beneficial. That is a swift and significant shift, indicating healthcare is ripe for the adoption of generative AI. As one Yale Medicine emergency medicine physician put it, “I am floored by the speed at which all of this has moved.”
CONSUMERS CALL FOR GREATER TRANSPARENCY IN GENERATIVE AI SOURCE CONTENT
While physicians have quickly understood the potential value of generative AI in healthcare, patients have questions they’d like answered before knowing generative AI is being used in their own care. In addition to surveying physicians, we also asked Americans for their perspectives and almost nine out of 10 agreed a problem with generative AI in healthcare is not knowing where it finds source content or how that information is validated.
As a recent Brookings commentary piece points out, technologies such as ChatGPT are trained on publicly available data only, noting that clinicians should “exercise caution” in deploying generative AI unless it is trained on extensive medical datasets and content. For both patients and doctors to trust and rely on generative AI at the point of care, it is critically important that the technology is trained on content provided and vetted by medical professionals.
It’s also important that the content used to train generative AI is current, as some generative AI platforms answer questions based on source materials that may not have been updated for years. This is particularly concerning in the field of medicine, where new clinical information is published regularly to guide patient care and treatments.
Not only does the quality of content matter when it comes to training AI models, but trusted, harmonized content is also important to help reduce care variability as the care continuum continues to evolve. By using technology to bring together clinical, drug, and patient engagement content, we can ensure all members of the care team have access to the same information to help drive consistent care and patient education across the healthcare ecosystem.
PHYSICIANS HAVE A CRITICAL ROLE IN ENGAGING PATIENTS
With research showing trust in doctors is higher than trust in the overall health system, physicians can play an important role helping healthcare realize the potential of generative AI, not just by adopting it for their own use but also by helping to educate their patients.
The importance of patient education is clear when you look at the data. When initially asked how they would feel knowing that their healthcare provider was using generative AI, four out of five respondents to our consumer survey said they would be concerned. But when asked the same question knowing their doctor was using generative AI that came from an established source in healthcare, was created by clinicians, and was up to date, the numbers of respondents who voiced concern dropped to three out of five.
Physicians share in Americans’ concerns on the need for trusted content. Before using generative AI in their own practice, nine out of 10 physicians say they would need to know the material it sourced from was created by doctors and medical experts. By communicating their high standards for using generative AI at the point of care, physicians can help earn patient trust in this new technology.
Physicians have tremendous influence on their patients’ health and wellbeing, and that is no small responsibility. Where physicians see potential for generative AI to improve how healthcare is delivered, it is because they see a benefit for their patients to receive more timely, personalized care. After hearing for years on the need to evolve to meet the growing demands of empowered, digital-savvy consumers of care, physicians now have a critical role to play in helping their patients evolve to prepare for a future of increasingly technology-augmented healthcare.