EPtalk by Dr. Jayne 12/5/24

It’s that time of year, with cybersecurity firm NordPass releasing its annual list of most used passwords. Topping the Hall of Shame list this year: secret, 123456, password, qwerty123, qwerty1, 123456789, password1, 12345678, 12345, and abc123. I didn’t have to scan too much farther down the list to find ones that were more interesting: iloveyou, baseball, princess, football, monkey, and sunshine all ranked within the top 20.

Come on, people, it’s not that hard to have at least a minimally secure password. The list can be sorted by country, and some of the international options are a bit more entertaining: liverpool, arsenal, and chelsea were popular in the UK, but hockey made the list in Canada.

From Cheer Mom: “Re: prescription drug fraud. Wise advice from Mr. H on physicians remaining vigilant around prescription drug fraud. One of our hospital’s providers recently discovered that his DEA number had been used for a number of fraudulent prescriptions for controlled substances. Too bad the patient in question was another provider at the hospital, who had been calling in her own prescriptions under her colleague’s name. The pharmacy didn’t catch the fact that the alleged prescribing physician sent every single prescription using electronic prescribing except for those called in for one single patient.” As someone who has had fraud committed against their DEA number, it’s a terrible thing when it happens. With the widespread adoption of electronic prescribing, it still amazes me that some states still permit certain levels of controlled substances to be phoned in.

A friend sent me this article from JAMA Network Open and asked my opinion on it since I’ve worked in telehealth for quite some time. It’s an original research article and looks at the rates of so-called low-value care services in primary care practices that use telehealth. The authors looked at care performed between January 2019 and December 2022 and used Medicare fee-for-service claims data for practices in Michigan. Practices were stratified as low, medium, or high users of telehealth and the low-value services were grouped as office-based, laboratory-based, imaging-based, and mixed-modality services. Over 577,000 patients were represented in the claims. Some of the low-value services avoided during telehealth visits included cervical cancer screening, PSA testing, and thyroid testing for patient groups where those tests were not indicated.

Non-clinical readers may ask why these services are considered low-value since at least some of them are marketed as potentially life-saving. In reality, it all depends on the patient, their age, and their risk factors as to whether the tests should be done. Sometimes physicians get in the habit of ordering tests across the board even when they’re not truly indicated, which makes them low value since they provide little to no clinical benefit for patients and can even cause harm or unnecessary follow-up testing. Since they require a physical exam or a blood draw, you can’t exactly conduct them during a telehealth visit, and doing so would require either a follow up-visit in office or a trip to the lab.

The authors found that practices that had high telehealth use had lower rates of low-value services performed in the office. There was no association between telehealth use and other low-value services that were not performed in the office. They concluded that, “our findings suggest the potential for telehealth to help reduce office-based low-value care and could reassure policymakers concerned about telehealth encouraging unnecessary or wasteful care due to added convenience.” One of the limitations of the study is the time period during which it was performed, which overlapped the worst parts of the COVID pandemic, when in-person visits were down across the board simply because primary care offices were closed. It would be interesting to perform a follow up on years post 2022 as well as to look at data from various parts of the country, to determine whether the results hold across time and place.

Still, I look at my own recent visit to my primary care physician. Except for a blood draw, it could have been performed via telehealth. The majority of the visit was spent discussing data gathered from home monitoring devices and updating the physician on a recent visit to a subspecialist who is not on the same EHR and who didn’t send a copy of their visit note. The blood draw wasn’t time sensitive and could have been easily done the next day since I would have to drive past the lab on a planned errand. For the labs that were ordered, it would have been easy for my physician to order a broad spectrum of labs, but fortunately he practices evidence-based medicine and only ordered the ones for which I was truly due. But for every physician who practices like that, there are twice as many who just order larger laboratory panels to “cover everything.”

There is still plenty of low-value care being performed, whether via telehealth or in-person visits. Antibiotics for viral illnesses are at the top of my list, and likely the lists of anyone who has ever worked in a primary care, urgent care, or telehealth urgent care setting during the three days leading up to Thanksgiving in the US. The number of patients who are presenting with what are almost certainly viral upper respiratory infections but who are simply seeking antibiotics is staggering. They come in with requests like, “I just want to get ahead of this because I’m having 20 people for dinner on Thursday” or “I just know this is going to turn into a sinus infection” and often haven’t tried any home care or over the counter remedies.

Frankly, writing an antibiotic prescription is a lot easier than a 20-minute conversation on why antibiotics aren’t indicated and how they can even cause harm, so you can guess how those visits often turn out, especially in practice settings where physicians are graded on patient satisfaction. I’d love to see a national public health campaign on appropriate use of antibiotics and why you don’t need to throw them at a common cold, but I don’t see that coming any time soon.

Like Mr. H, I’m migrating to Bluesky. You can find me there as @Jaynehistalkmd.bsky.social, although I’m slow to get started. I haven’t been much of a user of the platform formerly known as Twitter since its change of ownership, so maybe 2025 will be my year for returning to social media. I’m following Mr. H’s tip sheet for making the transition and looking forward to scrolling again with a more curated feed and hopefully fewer distractions.

A recent article published in Nature Communications looks at the effectiveness of an artificial intelligence system for matching patients with relevant clinical trials. Researchers from the University of Illinois and the National Institutes of Health have developed a solution called TrialGPT that was 87% accurate in matching patients with clinical trial eligibility criteria, which isn’t terribly far off from the performance level of humans. The study was limited by the fact that the system looks at written patient summaries versus lab values and imaging results, but I imagine it wouldn’t take too much work to bring structured data into the mix. I recently enrolled in a clinical trial that I only found out about through a tangential reference from one of my clinicians. It won’t yield results for five to 10 years, so it would be interesting to see what else I might be eligible for.

Have you ever participated in a clinical trial? Was there a technology component or did it involve manual data collection? Leave a comment or email me.

Email Dr. Jayne.