Ryan Harvey

Radiologist. Trying to keep up with all this

[Ryan Harvey @ryharvey]

@northwoods1980 LLMs are trash at this. Radiology-trained LVMs, however, are getting good.

[Ryan Harvey @ryharvey]

@fortelabs Typora is markdown native, but has an interface like a typical word processor.

[Ryan Harvey @ryharvey]

@benwhitemd Higher WUs than RVUs: x-rays (50% more), diagnostic mammo (100% more), CT chest, inpatients in general

Lower WUs than RVUs: CT head in ER, screening mammo, non-con brain MRI

[Ryan Harvey @ryharvey]

@benwhitemd The official methodology is to take the total CMS RVUs at a whole-practice level. Then distribute them to every permutation of CPT code and patient status (ER/IP/OP) based on average historical interpretation time for that combo. By definition, zero-sum.

[Ryan Harvey @ryharvey]

@20th_Centurygal Chumbawamba Tubthumping

[Ryan Harvey @ryharvey]

@benwhitemd Been enjoying the Lark M2S lapel mic you recommended

[Ryan Harvey @ryharvey]

@heynavtoor Just used Cowork to take two complex spreadsheets and make me another one synthesizing insights. This feels like a cheat code

[Ryan Harvey @ryharvey]

@drsthanus Right arch?

[Ryan Harvey @ryharvey]

@THErealDVORAK Reactogenicity is what she’s describing. The “sick” you feel is the immune response.

[Ryan Harvey @ryharvey]

@AnonymAce2 @TheChiefNerd Vaccines are highly protective but not 100%. If a vaccinated person is surrounded by people with measles, they have a much higher chance of a breakthrough infection. With vaccination rates around 90%, herd immunity is achieved and Measles goes basically to zero. Look it up.

[Ryan Harvey @ryharvey]

@_The_Prophet__ @unusual_whales AI slop

[Ryan Harvey @ryharvey]

@vukandric How do you make the numbers in the clock tall like that

[Ryan Harvey @ryharvey]

@realEstateTrent To compare apples to apples, I think you need to look at risk-free assets. Paying down the mortgage is a risk-free return. The best comparison is risk-free Treasury bills. Comparing to stocks means you are making assumptions about risk tolerance, which is a different conversation

[Ryan Harvey @ryharvey]

@benwhitemd Agree that the see-think-speak loop is important, and recursive. Speaking helps us form our thoughts, I think. But new report technology offers to take some of the speech burden away from us. We can just utter fragments, and an AI will create crisp, structured prose. Thoughts?

[Ryan Harvey @ryharvey]

@benwhitemd Agree. Within any work unit system, as long as rads have the same type of case mix, the peaks and valleys should wash out in the average. And if there are systematic differences between rads/rotations, then their productivity targets or bonus thresholds could be adjusted.

[Ryan Harvey @ryharvey]

@benwhitemd V2 will be adjustments for complexity, to the extent it can be accurately measured. (Trauma vs oncology. Inpatient vs outpatient. Etc)

[Ryan Harvey @ryharvey]

@benwhitemd Yes! In RP, some practices use work units based on time-spent for each exam code. We’ve got a big database of avg “pick-to-sign” times that level out the RVU mispricings we all know about. Also give credit for time spent on non-RVU citizenship activities.

[Ryan Harvey @ryharvey]

@FJB890 @EmeraldRobinson Hi. I’m here. Covid vaccine and boosters every year. Flu shots too. No probs. My wife too. And just about everyone I work with, since we’re doctors. I do know someone who died—but it was from Covid before the vaccine was available.

[Ryan Harvey @ryharvey]

@bwsears @theblanketdog @DGlaucomflecken Mainly because radiologists have not examined the patient or taken a full history—something the referring physician has done. Need complete context to accurately manage a patient. A single imaging data point is not enough.

[Ryan Harvey @ryharvey]

@jbcarmody The radiologist shortage is not distributional. Radiology is increasingly a remote nationwide market, and it’s quite tight.