Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/can-ai-ml-reform-us-medical.20255/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Can AI/ML reform US medical

Arthur Hanson

Well-known member
The US has among the world's most expensive medical taking twenty cents of every dollar while delivering 37th in quality of care. Where tested AI/ML has already beat the best doctors in diagnostics by 20% margin (in detecting breast cancer, in did better than the best doctors). This is a market of staggering size that is set for real reform for the benefit of everyone and the country. Combined with advancing robotics, this also offers a chance for the US to become a leader instead of a distant follower. The market for these possible technologies is worldwide and staggering in size and benefits for all. It would be nice for the US to become a leader instead of a follower when it comes to delivering value and quality to medical.
 
Time to call in John Edwards for a discussion about "greedy trial lawyers", lol! What I mean is that without indemnification at an individual provider level, or a new BAA structure to the healthcare providers, this is not implementable. The costs of indemnification/BAA would be bundled into the service cost, and make it impractical to turn a profit at the AI provider. AI/ML Diagnostic tools are already used in practice, so I am presuming you mean a standalone/automated process in my response.
 
Related https://www.theinformation.com/articles/tempus-ai-shows-ipos-are-finally-getting-weird-again

"Its IPO filing increasingly has leaned harder on this story. A Tempus IPO has been nearly three years in the making after the market for listings froze, leaving a paper trail for how the company’s story has evolved. Tempus used the word “AI” 78 times when it first filed draft IPO paperwork in 2021, but the latest version of the document says “AI” 228 times.

But the actual AI business is unproven. AI applications represented less than 1% of its revenue last quarter. That largely involves selling a machine-learning application that tries to notify oncology doctors “when a guideline or standard is not being followed,” Lefkofsky said in a release.

Tempus declined to comment beyond what’s in its S-1 filing."
 
To reform medical, you need to do what Uber did to taxis. Have take not prisoners execs and carpet bomb the government with lawyers to eliminate the obstacles, which are not technical.

It is a $4T market with (benchmarked at GDP fraction relative to better results in high income G20 nations with private health systems) about 40% waste. That waste alone is roughly equal to the sum of all income taxes.

The reasons it is not subject to market forces are regulatory capture and oligarchic vertically integrated incumbents, not technology.
 
@Tanj, appreciated. To understand the depth of the carpet bombing needed, spend a day reading up on FHIR, HL7, QHINs, RCEs. The 40% waste (an underestimate) is all in the "value-seeking" administration we've foisted on top of providers. "Data Bureaucrats" sense blood and adhere themselves like vampire leeches.

Quietly and without fanfare, "concierge medicine" is winning in USA because it avoids the payer-payee (plus leech adhesion) dynamics.
"Keep Yo' Business Clean (like a wafer)" - Unattributed future rap lyric from the 2030 Semi Wars
 
Back
Top