EPtalk by Dr. Jayne 6/12/25

From Boomer Sooner: “Re: Stanford’s EHR summary tool. The Department of Defense also recently launched an AI summary tool to help with the review of applicant records.” I know a thing or two about the process that military applicants go through, especially those who are applying to the military service academies or are going through the selection processes for highly selective fields. The onus of trying to get all the records to the right place is on the applicant, and it can be tricky when a practice doesn’t release records quickly. One of my favorite candidates said that in that process, the applicants who were military dependents had a bit of an advantage because their records were more easily accessible by reviewers.

The new tool, which was developed by the Innovation Facilitation Team at the US Military Entrance Processing Command (USMEPCOM), creates AI-enabled summaries of medical documents, reducing the time required for provider review. The summary can be seen in the MHS Genesis system as an encounter summary.

I was excited to learn about a recently enacted Arizona law that is aimed at protecting physicians and patients from unintended consequences that are related to AI. House Bill 2175 is designed to keep health insurance companies from using AI as the ultimate decision maker as they review claims and deal with medical necessity appeals and denials. It also applies to prior authorization requests and recognizes that cases that require medical judgment should be reviewed by licensed medical professionals with the appropriate training, experience, and ethical responsibility that is needed for clinical decision making. The law was introduced with the support of the Arizona Medical Association and various care delivery organizations and advocacy groups and goes into effect in 2026.

Nebraska is also addressing hot button healthcare issues with the Ensuring Transparency in Prior Authorization Act, which requires insurers to make their prior authorization requirements visible on their websites. Similar to the Arizona law, it prevents AI from being the sole basis for a denial of coverage. It also requires a 60-day notice period before payers can add new requirements. We often think about healthcare IT in terms of provider side organizations, but plenty of tech folks are working on the payer side. It will be interesting to see how much work is done on websites and how quickly it happens. I’m betting that payers drag it out until the last minute, knowing that it doesn’t go into effect until January 2026.

One more state wading into the healthcare fray is Indiana, which recently enacted a bill that requires non-profit hospitals to either lower their prices or lose their tax advantaged status by 2029. Hospitals will be required to submit audited financial statements that show a decrease in their prices to match or be less than the statewide average. Failure to submit the audited statements can result in a $10,000 per day penalty. The bill has other interesting features, namely creating a state directed payment program for hospitals as well as a managed care assessment fee. A provision requires insurers and health maintenance organizations to submit specified data to the all-payer claims database and another one to reduce drug costs for the state employee health plan.

I wasn’t aware of Guidehealth until the company announced this week that it had received a $10 million investment from Emory Healthcare. As one would expect, the solution has an AI-enabled component. It advertises “AI-driven intelligence with human-centered care” using medical assistants that are “trained in data science and empathy.” They are branded with the trademarked Healthguides moniker. The company plans to use the additional investment to add AI-powered virtual care navigation to support analysis of patient-reported data and with interventions that target fall risk or depression screenings.

Guidehealth was already working with Emory’s Population Health Collaborative to boost quality scores under a Medicare Advantage contract. I would be interested to understand the medical assistant training and whether unique hiring algorithms are being used to find individuals with a particular level of empathy. In my experience, that’s not only hard to find at times, but difficult to enhance with training.

Speaking of AI, over the last year a couple of articles looked at AI-generated messages to patients and found that those with an AI origin were more empathetic. A new study that looked at medical queries across the US and Australia found the opposite. The AI-enabled responses were more accurate and professional than human responses, but lacked emotional depth and also raised concerns of data bias. I’m sure we’re not done with this one, and many more research efforts will be looking at the phenomenon.

While many organizations are looking at technology solutions to close gaps in care, particularly in preventive services, a recent study showed that for cervical cancer screening, lower tech interventions can still drive the needle. Researchers looked at patients in a safety net care setting and compared rates of cervical cancer screening. Patients who received a mailed self-collection kit along with a telephone reminder had greater participation (41%) than those who received a telephone reminder alone (17%). It just goes to show that nudges aren’t enough. We need to make it easy for patients to get the recommended services rather than just telling them they need to do it.

From Weird Al: “Re: earwax as the newest precision medicine tool I wonder how much these tests will cost?” A BBC article notes that wax could contain biomarkers for cancer, metabolic disorders, and even Alzheimer’s disease. Since ear wax is relatively stable, it might be able to show longer-term trends with various chemicals. There’s a team at Hospital Amaral Carvalho in Sao Paulo that is looking at cerumen for cancer diagnosis and monitoring, and several other institutions are conducting research.

Having spent many long hours in the emergency department and urgent care centers, I feel like worked with more than my share of ear wax. Running tests on it isn’t as cool as diagnosing conditions using a Star Trek-style tricorder, but here’s to the next generation of research and seeing if we can develop tests that are not only less invasive, but cost effective.

What healthcare technology advancements do you feel have really changed how we approach patients or conditions? Are they glamorously high tech or startlingly low key? Leave a comment or email me.

Email Dr. Jayne.