Most of us have seen images created by AI, whether we realize it or not. Sometimes they’re easy to spot, such as when people don’t have the correct numbers of digits or have atypical facial expressions.
Those who use stock images in their work rather than AI-generated images may have higher quality options, but a recent article in Hypertension looked specifically at the accuracy of images that are used to educate patients on how to monitor their blood pressures outside the medical environment.
The authors visited the websites of 11 major online stock photo providers and analyzed the first 100 images on each after searching for the term “blood pressure check.” They found that only 14% of the images were accurate as far as the patient being correctly positioned and with an appropriate device, but scores ranged from 7% to 28%.
Some of the most common issues were patients whose backs weren’t supported, feet weren’t on the floor, forearms weren’t resting on a surface or level with the heart, and legs were crossed. The authors call for better education, not only for patients, but for media organizations and website developers.
From Patagonia Sweater: “Re: professional attire. Right after I saw your mention about that and the role of white coats last week, my office administrator shared this article about the potential for medical coats to aid in the spread of drug-resistant infections. Nearly everyone in my office wears a monogrammed jacket courtesy of our academic department. Unlike white coats, the heathered appearance makes it much harder to tell if they’ve been cleaned lately so I think there’s a bit of an “ick” factor there.” The study looked specifically at contamination of healthcare personnel gowns (as they are called in the country where the study originated) by gram-negative bacteria and the relationship of that contamination to growing antimicrobial resistance. The observational study looked at 321 hospital workers and found a contamination rate of 61% for the gowns, with medically important bacteria found more often on those worn in operating suites and intensive care units.
The authors concluded that healthcare personnel gowns are a significant reservoir of pathogenic bacteria at the hospital in question. They stated that “It is essential to implement infection control strategies that include improving the cleaning and laundering of gowns and ideally eliminating them from clothing to reduce the risk of transmission of nosocomial infection.”
There’s some irony to this when you consider the origin of the white coat as mentioned in the article that such attire has “been considered a symbol of authority, respect, cleanliness, neatness, commitment to health, and perceived patient safety” and that it dates back to the 1800s when Joseph Lister promoted its use during surgical procedures as an element to combat the presence of germs.
From AI Skeptic: “Re: AI. I’m one of the curmudgeons sitting in the back row and eating popcorn while waiting for the AI bubble to burst. There have been examples of AI creating bogus citations for scholarly articles and legal filings, but I got a kick out of this piece that looked at how an AI tool flagged journals for ‘questionable’ conduct.” The article proposes that making use of the algorithm “could help scientists avoid publishing in shady titles.” The underlying study looked at 15,000 open access journals to identify those that could negatively impact scholarly work by prioritizing profits over scientific integrity. More than 1,000 journals were flagged as potentially problematic.
The work is receiving praise from organizations that promote quality and transparency in scientific publishing. The article mentions shifts in publishing business models, where authors pay a fee so that their articles are free to read, and notes that such a model has created incentives to publish high volumes of papers fast at the expense of ensuring quality.
That last sentence really resonated with me. I’ve seen too many examples lately where quality is being devalued in favor of incentivizing other factors, such as patient reviews, facility aesthetics, and speed of treatment. No one wants to spend more time waiting around for healthcare services than they have to, and inefficient processes are maddening to me whether I’m wearing my patient hat or my physician hat. However, I also don’t want to be rushed through the care process by an organization or care team that’s cutting corners because they’re trying to meet an outsized metric.
The article mentioned that the algorithm isn’t as strong as it could be, noting that there were 1,700 false negatives, but also some false positives. Attempts to tune the model weren’t as effective as they hoped. Researchers in the field note the need for ongoing development of the model in order to combat unscrupulous publishers who change their titles or processes in an attempt to avoid being identified. The article notes that such publications will persist as long as research institutions base tenure and promotions on the number of papers published.
I’m always on the lookout for articles about wearable health devices. This one brought up a point that I hadn’t thought of previously. One of the physicians interviewed commented that, “When you become too dependent on what you perceive to be objective data … you lose a certain relationship with your body, such that it becomes hard for you to discern how you feel apart from what a device is telling you.”
We as physicians are always counseled to treat the patient, not the numbers, as a way of reminding us that we need to look at the entire picture of a patient’s history, symptoms, and exam and not just lab values. Maybe we need to incorporate some similar messaging into the conversations we have with patients about the best way to use health trackers in the real world.
Speaking of wearables, I’ve written previously about the Oura Ring, but a recent item mentions that the new facility in Forth Worth, TX is needed to support the company’s expanding relationship with the US Department of Defense, which is referred to as the company’s largest enterprise customer. The partnership has been around since 2019 and focuses on stress management, resilience training, fitness optimization, fatigue risk management, and early illness detection. The Texas facility is being purpose-built to fulfill defense orders and will have additional security.
I reached out to some active-duty military personnel, including ones for whom disrupted sleep is the norm. It sounds like the decision to purchase them is handled at the unit level. One officer mentioned that although he has heard about it, he’s never seen one in the wild.
Another mentioned that some human performance staffers showed interest in obtaining them, but it was determined to be a questionable use of funding, noting that “people who work in a SCIF (Sensitive Compartmented Information Facility) are probably most likely to be stressed, but can’t wear one. Nor can those at high risk of fatigue, such as air crew and special operations personnel.”
I would be interested if other military folks would like to weigh in. If you’re an Oura user and you feel that the device is making a difference for your health and well-being, give us a shout.
Do clinicians in your organization recommend fitness trackers or other wearables for patients? Have you used one to help manage your health? Leave a comment or email me.
Email Dr. Jayne.
"A valid concern..." Oh please. Everyone picks the software they like and the origin of that software is an afterthought.…