EPtalk by Dr. Jayne 7/18/24

Former US Food and Drug Administration Commissioner Scott Gottlieb, MD published a call to action this week in JAMA Health Forum that asks Congress to update FDA regulations for medical AI. He begins the piece by summarizing the events leading up to the FDA’s approval of the Apple Watch in 2018 for identifying irregular heart rhythms, noting that the FDA cleared the device largely based on its developer’s validation and quality approaches rather than on a review of the hardware itself. He states that “this same concept is uniquely suited to the regulation of artificial intelligence (AI) medical devices that can augment patient care.” Bills are pending in both the US Senate and the House with the so-called Verifying Accurate Leading-edge IVCT Development Act (VALID Act) creating laws around this regulatory approach.

Gottlieb says that change will allow the FDA “to oversee the methods used to develop a technology and validate its reliability, rather than trying to decouple the product’s construction” and draws parallels between device regulation and the need to regulate medical AI, especially with regard to rapid innovation and development cycles during product development. He goes on to discuss developers’ approaches to mitigating any FDA uncertainty, including avoiding having their solutions be classified as devices. Clinical decision support software isn’t subject to the same level of scrutiny as medical devices, which allows a faster go-to-market approach for developers. It will be interesting to see if Congress passes the VALID Act and if they then in turn move forward with policies to address AI technologies.

Bad news for night shift workers. A recent study that was published in The Lancet suggests a higher risk of diabetes for individuals who were exposed to the most light between 12:30 a.m. and 6:00 a.m. Study participants wore light sensors to capture personal light exposure, which strengthened the reliability of this study compared to its predecessors. The sensors captured light in all forms, such as the sun, lamps, or screens. After eight years of tracking, researchers found that those with lower overnight light exposure had a lower risk of type 2 diabetes. Those with the highest exposures had a risk increase that was similar to that for patients with a family history of the condition. It’s suspected that atypical light exposure alters the body’s circadian rhythm, which can have an impact on how it handles sugar. I guess I need to get more sunlight during the day to counterbalance the late night monitor light that I’m exposed to while writing for HIStalk.

I’m playing catch up with my journal reading, so I’m just now seeing this piece from the March Journal of the American Board of Family Medicine that looked at the differences in hospital readmission rates for patients who received their follow-up care in person compared to telemedicine. The authors found comparable readmission rates regardless of the follow-up modality, concluding that “telemedicine poses little threat of negatively impacting HEDIS performance” and may be as effective as traditional in-office transition of care visits. The authors note some limitations in the study, including reliance on provider accuracy to capture discharge follow-up codes and the inability to capture the information patients who had follow-up visits outside the EHR whose data was used for the study. They also noted that the telemedicine sample size was small and had a younger population. Larger multi-site studies that incorporate intentional use of telehealth would be of benefit to create stronger evidence.

I consume a lot of study write-ups as part of my regular reading, so I’m familiar with how to critically appraise data and determine if the authors of a particular piece are trying to lead readers to a conclusion that might not fully correlate with the data. I was skeptical when I saw headlines this week about the physician burnout rate falling below 50% for the first time in four years. The AMA is claiming this result from their “exclusive survey data” that compares record-high data from 2021, where 63% of physicians reported burnout, to more recent data collected in 2023. Data was collected as part of what the AMA calls its “Organizational Biopsy” and represented 12,000 physician responses across 31 states.

Since this is proprietary AMA data and not a peer-reviewed publication, it is unclear whether or how it was controlled against previous data. Were the respective physician panels representative as far as specialty, age, and gender? What about practice setting or full-time status? How about employment status and the stratification of academic physicians against private practice or those in an organization that is owned by private equity?

I’m not a burnout expert, but I’ve talked to hundreds of physicians in the last several years, and here is my private hypothesis. The most burned out physicians have retired early, cut back, or otherwise left direct patient care. I receive at least a dozen requests each month from physician contacts who want to learn more about “how to get off the hamster wheel” and whether they can just make the jump to clinical informatics or a technical role. (Spoiler alert: it’s not as easy as you think.) Many of them get pulled into unsavory arrangements that essentially amounts to their renting their medical licenses to companies that are looking to make a buck. I wonder how or if those physicians have been represented in the AMA’s data gathering efforts.

Based on an email I received today, HIMSS must be desperate for revenue, because they’re promoting sales of the recordings from the HIMSS24 conference earlier this year. On top of the 150 recordings from this year, they’re throwing in bonus recordings from HIMSS22 and HIMSS23. I can’t imagine that many attendees who are thinking back to those conferences and wishing they had a recording of a particular session. If I’m seeking deeper information about a conference presentation or topic, I’m likely to just reach out to the presenters, who are generally excited to correspond about their pet projects. If you’ve got cash to burn and time on your hands it might be for you, but to me it feels like a sad attempt to squeeze revenue out of former attendees.

CMS has issued an invitation to its Leadership National Call Update on August 1 at 3:30 p.m. ET. Administrator Chiquita Brooks-LaSure and her team will be updating attendees on advancements related to the CMS Strategic Plan. I’ve never attended one of these calls and was surprised to learn that the registration link leads to a special Zoom for Government site. I wonder what features are different from a corporate Zoom account or even a paid individual account? Inquiring minds want to know, so if you have the details, leave a comment or email me.

Email Dr. Jayne.