Week of February 19, 2024
My patients came with textbook's worth of communication and comprehension barriers last week. A lack of hearing aids, the cognitive interference of upsetting news, and one patient's tendency to hijack the conversation with extraneous detail made the pace of the back-to-back appointments slower and more tedious than I wanted. Like parenting, healthcare often requires ingenuity and mental Jiu-Jitsu. Relaying essential information and encouraging the appropriate next steps while working within the patient's limitations is an art. And some days in the office, it feels like I'm finger-painting.
P.S. To be clear - patients also offer many moments of joy. Last week, an 85-year-old patient described herself as "a cougar" when telling me about her second husband, who is only 82.
Here is a quick read offering a framework for communication barriers in healthcare. There are tons of other references.
https://flearningstudio.com/communication-barriers-in-healthcare/
---
This week, I found Dr. Michael Hoerger's Pandemic Mitigation Collaborative (PMC) website, which uses wastewater levels to forecast 4-week predictions of COVID rates. While the Biobot data is helpful (and used in the PMC), the PMC dashboard is a more useful analysis. This website uses multiple models to offer a 4-week composite forecast like weather aggregators.
If you click on the multi-page .pdf report, look at page 9, which compares the seasonality of COVID over the last few years. The "3 seasons" depicted by the wastewater data are striking.
https://pmc19.com/data/PMC_Report_Feb12_2024.pdf
The N.Y. Times COVID Tracker reflects only CDC-gathered hospital data. Hospitalization data are a (lagging) indicator.
https://www.nytimes.com/interactive/2023/us/covid-cases.html
Wastewater monitoring is more of a LEADING indicator.
------
COVID articles
U.K. researchers looking at Brazilian health system data found an association between Tamiflu (Oseltamivir) use, the antiviral sometimes used for patients with viral influenza, and a reduced risk of death from COVID-19 (as compared to patients who received no antiviral medication). The study is a retrospective cohort analysis, yet it offers an interesting finding in a large population (there were hundreds of thousands of cases). Tamiflu does not directly inhibit coronavirus but may blunt immune response to viral infection. While this is far from enough evidence to advise treating COVID with Tamiflu, it raises the possibility of an additional, low-risk means of preventing the most severe outcomes.
https://www.sciencedirect.com/science/article/pii/S0924857924000293
and
https://twitter.com/enemyinastate/status/1758254777199128822
European scientists manufactured a "lite" version of the coronavirus (a virion) to better understand the viral spike protein. (The spike protein is used to bind to and enter cells and is often the target of the immune system's response to infection). Using their manufactured virus, they could more clearly see how the spike protein moves - how it changes configurations to hide away - making the virus stealthier to the infected organism's immune system. While spike protein research is not novel, per se, the use of engineered biological models is.
A related older article on spike proteins with graphics: https://answers.childrenshospital.org/sars-cov-2-spike-protein/
Here is a related article from a loyal reader. Researchers have discovered a naturally occurring, oddly shaped virion. Like the previous article, virions are small, virus-like RNA bundles that infect bacteria and replicate. There are two novel findings - the shape of the virion and the fact that they infect the bacteria of the human microbiome.
As trite as it is, all these articles reinforce that Michael Crichton had it right, "Life breaks free. Life expands to new territories. Painfully, perhaps even dangerously. But Life finds a way."
Medical Trends and Technology
A meta-analysis (including 218 studies) found that intense exercise (defined as walking, jogging, yoga, or resistance training) offered significant improvement in depression symptoms as compared to medication alone. Not only is this worth reading for the findings (prescribe exercise + medication), but the BMJ does a fantastic job with summary infographics.
https://www.bmj.com/content/384/bmj-2023-075847
Infographic
https://www.bmj.com/content/bmj/384/bmj-2023-075847/F1.large.jpg
I wonder if my anti-science patients (and other anti-vaxers) would be willing to receive the Ebola vaccine in the event of an outbreak. First developed in 2020 (link here), recently published data demonstrated the first vaccine for the Ebola virus offered remarkable results:
"The research, based on data from the massive 2018-2020 Ebola Zaire outbreak in the Democratic Republic of the Congo, showed that the risk of dying of Ebola was halved among people who had been vaccinated with a single dose of Merck's Ervebo before developing symptoms — including those who had only received the vaccine a day or two before becoming ill. While that is not enough time for the immune system to develop a robust response to a vaccine, there was clearly a benefit. The case fatality among people who had been vaccinated two or fewer days prior to becoming ill was 27%, compared to 56% among people who were unvaccinated."
https://www.statnews.com/2024/02/12/ebola-vaccine-new-study/
andhttps://www.thelancet.com/journals/laninf/article/PIIS1473-3099(23)00819-8/abstract
Infographics
This XKCD infographic made me laugh - "Large spheres that taste mediocre" and the intellectual potholes of conclusions based on projecting trend data.
https://twitter.com/EikoFried/status/1758142001822736580/photo/1
from
https://www.explainxkcd.com/wiki/index.php/2893:_Sphere_Tastiness
Things I learned this week
In the early 2010s, when my kids started losing their baby teeth, I attempted to set up a bidding opportunity for my family - auctioning off the right to be the tooth fairy. My kids' grandparents seemed uninterested in my market-based approach, but I was WAY ahead of my time.
"What's the Tooth Fairy Leaving These Days? $100 Bills and Louis Vuitton Bracelets."
I found some (very entertaining) evidence that peer review alone does not reflect meaningful intellectual activity, data, or honesty. Though the journal retracted the article, the paper's nonsensical A.I.-generated images are available on Twitter. The mouse graphic, in particular, is simultaneously impressive, disturbing, and evocative.
https://twitter.com/cliff_swan/status/1758135084069302761
and
https://www.frontiersin.org/articles/10.3389/fcell.2024.1386861/full
Living with A.I.
It has been another week of surreal events (which seems to be the norm in the A.I. world).
I am surprised but not surprised by the Wired article - A.I. Girlfriends Are a Privacy Nightmare. While your A.I. love bot may not care about you, the people reading your intimate data seem to.
"An analysis into 11 so-called romance and companion chatbots, published on Wednesday by the Mozilla Foundation, has found a litany of security and privacy concerns with the bots. Collectively, the apps, which have been downloaded more than 100 million times on Android devices, gather huge amounts of people's data; use trackers that send information to Google, Facebook, and companies in Russia and China; allow users to use weak passwords; and lack transparency about their ownership and the A.I. models that power them."
https://www.wired.com/story/ai-girlfriends-privacy-nightmare/
OpenAI published results from their soon-to-be-released tool Sora, an A.I. text-to-video generator. You write a prompt; it makes a disturbingly realistic video. Even the sharpest critical thinker will have difficulty distinguishing generated video from reality when this tool is misused.
Here are some examples of prompts and output:
https://twitter.com/minchoi/status/1759219716009107936
Here is how it works:
https://x.com/AlphaSignalAI/status/1758523862470320634
And, while we're at it, here is some data on how good ChatGPT is at brute-force hacking databases. Spoiler - LLMs will be better than weaponized troll farms and hacktivists.
https://twitter.com/AlphaSignalAI/status/1758896740990660893
I find it difficult to avoid dystopian brooding about what happens when some of these functions are combined - real-time generated video avatars baiting humans to reveal sensitive information and using the data to a nefarious end.
A.I. art of the week
OpenAI's DALL-E vs. Google's Gemini Pro.
And speaking of dystopian brooding, I generated an unintentional homage to Westworld this week. Here are some variations on the prompt "A man sitting at a coffee table with a robot-appearing woman. They are on a date. She is scanning his head with a device to read his thoughts."
DALL-E
Google's Gemini Pro
Clean hands and sharp minds,
Adam
Comments
Post a Comment