RKendallPMR
3.5K posts

RKendallPMR
@mtnfreek
Rehabilitation Spine/MSK physician, Outdoor, endurance athlete. Believer in Potential. Opinions my own.
SLC Katılım Temmuz 2012
720 Takip Edilen704 Takipçiler

Seriously? Talk about marketing tactics. Does this guy actually think we don’t know about the incredible of benefits of exercise and nutrition, or pass those on to our patients? I can’t speak for all physicians, of course, but I, for one, would certainly prefer to see a patient take ownership of their health and remain as prescription-drug-free as possible.
Mike Matthews@muscleforlife
If doctors prescribed diet and exercise as zealously as they prescribe drugs, there would be a lot less chronic disease.
English

@dereckwpaul A BIG problem with this is the extensive use of cloned notes which say almost NOTHING new over multiple visits. Garbage in….Garbage out.
English

AI is about to fix one of medicine's most well-intentioned but potentially dangerous practices. The chart biopsy.
Before seeing a patient, a doctor is going to do their best to review the chart so that they have the context needed to handle whatever the patient's chief complaint is at the time effectively. Oftentimes, doctors do not have time to review the entire chart or review the chart at all, secondary to the ongoing pressures to see more and more patients.
The very best doctors I've known have, at the very least, tried to do a chart biopsy. They scan through the available previous documentation in the chart and try to intelligently choose pieces of documentation that are likely to provide them with a good overall picture of the patient. For example, a previous specialist note.
The reality is that they can only choose one or two notes to read—hence the term chart biopsy. Just like a real tissue biopsy, it is only taking a tiny sample of what is going on. A good chart biopsy is better than no chart review, but it leaves a tremendous amount of unreviewed context.
In the course of an inpatient hospitalization, as the hospitalization plays out, a great doctor will review more of the chart through those days and learn more about the patient. They accumulate previous context and present context through their conversations with the patient over time.
But at the time when they are seeing the patient and admitting the patient and making the most critical decisions, they often have the lowest context.
Growing context windows mean that an AI can actually process more text than the average human can. A million token context window is the equivalent of about ten novels. For patient records that are even bigger than this, they can be represented and searched using techniques including context compaction, summarization, and vector embedding.
AI is going to make it so that the chart biopsy is not necessary. We are going to have considered all the available context early on in patient encounters, at the point of care when it really matters. This is going to help doctors be more effective at their job of diagnosis and treatment planning, and lead to downstream positive impacts for patients.
English
RKendallPMR retweetledi

A new study just published. @CzaplewskiLloyd
60% of persistent LBP cases with Modic types 1 & 2 achieved substantial reductions in pain and disability after an antibiotic-infused linezolid intradiscal injection. I have read the paper, and my preliminary report is available on my blog.
drmarklaslett.com/post/
Urge you to acquire and read a copy - it is open access
thelancet.com/journals/eclin…

English
RKendallPMR retweetledi

Up to 30% of low back #pain patients have #SIJ complex pain, but it’s often missed. New consensus #guidelines from @AmerAcadPainMed & @ASRA_Society provide a roadmap for care.
📄 Q&A with lead author Dr. Zachary McCormick (@UofUHealth): bit.ly/3NyChbC
#BackPain

English
RKendallPMR retweetledi

There were 2 big physical activity studies this week that have attracted a lot of attention
bmjmedicine.bmj.com/content/5/1/e0… @BMJMedicine
thelancet.com/journals/lance… @TheLancet


English
RKendallPMR retweetledi

✨Featured: Researchers explore the assoc b/n preop lumbar paraspinal IMF & recovery over a 5y window s/p surgical decomp for LSS. Pts w/ nonsevere IMF in the lum multifidus reported a ⬆️% of successful recovery & surgical success comp w/ the severe group
journals.lww.com/spinejournal/f…

English

@jonhagedornn What watch do you run with? Is this the Garmin + edition of connect?
English

@gretchemaben Maybe you should send out the contact so all of X could contact them about itching.
English


@EricTopol @ouraring @tomeghale I’ve used one for the past year. While some interesting data to be aware of, I won’t continue it past this year.
English

The CEO of @ouraring proposes a new FDA regulatory path for wearable sensors—"digital health screeners" @tomeghale
What do you think of this idea?
wsj.com/opinion/with-l…

English

@stevemagness Unfortunately the 99% can’t understand complicated programs are unlikely to get them to the 1% (and I mean reality, not ego).
English
RKendallPMR retweetledi

Our engineers give a bionic hand a mind of its own. #UofU researchers are using AI to finetune robotic prosthesis to improve manual dexterity by finding right balance between human and machine control.
📰 attheu.utah.edu/health-medicin…

English












