Artificial intelligence (AI) will profoundly change the patient experience of healthcare. Many of the changes will be “under the hood” and not immediately visible to the public. As with most technological disruptions, many commentators will overestimate the change that will happen very soon but underestimate the level of change that will be evident in a decade.
01
Most medical practices haven't used AI in their patient-facing operations, although AI can optimize appointment scheduling and provide 24/7 customer service. We'll see increased use of patient-facing AI as more practices use electronic health records for scheduling and standardizing their administrative practices.
02
Many patients already see a change in their provider office visits, where providers using “ambient scribing” use AI to create their notes. This means clinicians don’t have to be focused on the computer screen — and could mean more useful office visit notes. There are some reports of AI hallucinating information from an office visit, so clinicians must still review AI office visit output. As a practical matter, providers often make errors in their medical documentation without any use of AI.
03
Doctors who use AI can also become better diagnosticians. AI systems can digest massive quantities of unstructured data, including history, physical exam, imaging and laboratory tests. Sometimes, AI can detect the “fingerprints” of a rare disease that even a seasoned physician would miss. That’s because doctors are more subject than AI to certain biases, such as selectively remembering recent cases and putting information together in a way that confirms their initial hypothesis.
Clinicians are already using AI based tools to answer their questions about care for patients. For example, Google searches for the AI tool OpenEvidence.ai, which licenses content from the Journal of the American Medical Association, the New England Journal of Medicine and others, are approaching searches for uptodate , which has long been the go-to source for practicing clinicians.
04
Patients will likely see more personalized recommendations for preventive care and chronic disease treatment based on AI. For example, some women would get the most benefit from mammography at an interval other than two years. The availability of AI to segment risk would allow hyperpersonalization. Radiologists are already using AI programs to interpret mammograms and arterial plaque in CT coronary angiography to determine which patients should have a cardiac catheterization, and which can be safely treated and monitored without that procedure.
05
Patients also benefit from a host of new medications being approved, as AI can predict protein folding and could speed up the process of drug invention. As these new drugs preserve health, we’ll need new approaches to maintaining affordability.
06
Some have suggested patients could see cost savings due to the use of AI instead of expensive medical professionals. Right now, there's an AI arms race to raise (or lower) medical costs. Some health plans use AI to downcode office visits, and many providers use AI to increase revenue capture. Our current healthcare finance system allows higher charges when new technology is introduced. The Center for Medicare and Medicaid Services just approved a $1,000 fee for the AI interpretation of CT coronary angiograms. So, I’m not convinced that AI will be the solution to skyrocketing healthcare costs.
Each patient’s healthcare experience isn't likely to be totally transformed by AI this year or next. But from our current vantage point, the healthcare experience of 2035 would likely be unrecognizable.
Our video series boosts employees' mental health knowledge, equipping them to support themselves, their loved ones and foster a supportive workplace community.
