Jaunt Logo

    AI's Impact on the Future of Healthcare

    AI's Impact on the Future of Healthcare

    Z
    @ZakTheK
    4 Followers
    4 months ago 347

    Artificial Intelligence is rapidly transforming healthcare, stirring both excitement and caution. As innovations emerge, challenges like the primary care shortage and data utilization become pressing. This exploration dives into AI's potential benefits and the looming questions surrounding authority, access, and the ethical implications of integrating AI in medicine's future.

    AI and Healthcare
THE MUDDY FUTURE
Zak Kohane
Department of Biomedical Informatics 
Harvard Medical School
October 4th, 2023
    1/24
    Highlights
• Why the excitement?
• Reminder of a few of the potholes.
• Interesting ways forward for infectious disease.
    2/24
    UDN1
3
Worsening dystonia
(not walking, not speaking)
L-Dopa
Folinic Acid
5-hydroytryptophan
GTP cyclohydrolase I deficiency
Walking, talking!
    3/24
    Present
    4/24
    5
DEPARTMENT OF
Biomedical Informatics
Trend in USLME over 2 years
    5/24
    6
DEPARTMENT OF
Biomedical Informatics
Trial Path
Trial
    6/24
    7
DEPARTMENT OF
Biomedical Informatics
Trainee path
Trainee
    7/24
    8
DEPARTMENT OF
Biomedical Informatics
Trochbearer Dr. House
Torchbearer
    8/24
    9
DEPARTMENT OF
Biomedical Informatics
9
    9/24
    10
DEPARTMENT OF
Biomedical Informatics
When does Medicine succumb to 
AlphaZero?
• Deterministic.
• Fully observed.
• The action space is discrete.
• Access to a 
perfect simulator (the game 
itself), so the effects of any 
action are known exactly. 
• Each episode/game is 
relatively short.
• Evaluation is clear, fast and 
allows a lot of trial-anderror experience.
• Huge datasets of human play.
• NOT physiology
• NOT disease course
• NOT drug response
• NOT surgery
• So…. What might succumb…? Which 
medical ‘game’ fits the criteria….?
•
    10/24
    11
DEPARTMENT OF
Biomedical Informatics
25% of US Healthcare Costs
11
NBER Economics of Artificial Intelligence Conference, September 2022. 
Nikhil R. Sahni, George Stein, Rodney Zemmel, and David Cutler
    11/24
    12
DEPARTMENT OF
Biomedical Informatics
Primary care shortage between 
17,800 and 48,000 
• AAMC report
• Retirements 
getting worse 
with COVID and 
burnout.
    12/24
    13
DEPARTMENT OF
Biomedical Informatics
Not just a 
US Problem
    13/24
    14
DEPARTMENT OF
Biomedical Informatics
+
14
AI?
    14/24
    15
DEPARTMENT OF
Biomedical Informatics
15
    15/24
    16
DEPARTMENT OF
Biomedical Informatics
    16/24
    17
DEPARTMENT OF
Biomedical Informatics
4 weeks later: CovidClinical.net & 
medRxiv…Nature Digitial Medicine
    17/24
    18
DEPARTMENT OF
Biomedical Informatics
• Early evidence of major 
pathologies
• Countries more similar than 
sites within them
Meanwhile…
    18/24
    19
DEPARTMENT OF
Biomedical Informatics
    19/24
    20
DEPARTMENT OF
Biomedical Informatics
Fundamental antecedent questions 
to inform policy
20
Whose utilities are 
maximized?
Whose medicine is 
authoritative?
Dyad to Triad: Distinct 
New Party ?
Who controls 
incorporation of patient 
data into model?
Should consumers have 
access to medical advice 
directly?
Who funds?
    20/24
    21
DEPARTMENT OF
Biomedical Informatics
21
    21/24
    22
DEPARTMENT OF
Biomedical Informatics
22
    22/24
    23
DEPARTMENT OF
Biomedical Informatics
Implications
• Sea-change in healthcare with AI
º Faster than regulation
º Faster than education
• Acceleration of clinical and basic research
• Data + computational resources are obvious king.
• Globalization paths
• Increases dangers of non-quant data exploration.
• Urgency of exploration by educators and science 
leaders
23
    23/24
    24
DEPARTMENT OF
Biomedical Informatics
NEJM AI
    24/24

    AI's Impact on the Future of Healthcare

    • 1. AI and Healthcare THE MUDDY FUTURE Zak Kohane Department of Biomedical Informatics Harvard Medical School October 4th, 2023
    • 2. Highlights • Why the excitement? • Reminder of a few of the potholes. • Interesting ways forward for infectious disease.
    • 3. UDN1 3 Worsening dystonia (not walking, not speaking) L-Dopa Folinic Acid 5-hydroytryptophan GTP cyclohydrolase I deficiency Walking, talking!
    • 4. Present
    • 5. 5 DEPARTMENT OF Biomedical Informatics Trend in USLME over 2 years
    • 6. 6 DEPARTMENT OF Biomedical Informatics Trial Path Trial
    • 7. 7 DEPARTMENT OF Biomedical Informatics Trainee path Trainee
    • 8. 8 DEPARTMENT OF Biomedical Informatics Trochbearer Dr. House Torchbearer
    • 9. 9 DEPARTMENT OF Biomedical Informatics 9
    • 10. 10 DEPARTMENT OF Biomedical Informatics When does Medicine succumb to AlphaZero? • Deterministic. • Fully observed. • The action space is discrete. • Access to a perfect simulator (the game itself), so the effects of any action are known exactly. • Each episode/game is relatively short. • Evaluation is clear, fast and allows a lot of trial-anderror experience. • Huge datasets of human play. • NOT physiology • NOT disease course • NOT drug response • NOT surgery • So…. What might succumb…? Which medical ‘game’ fits the criteria….? •
    • 11. 11 DEPARTMENT OF Biomedical Informatics 25% of US Healthcare Costs 11 NBER Economics of Artificial Intelligence Conference, September 2022. Nikhil R. Sahni, George Stein, Rodney Zemmel, and David Cutler
    • 12. 12 DEPARTMENT OF Biomedical Informatics Primary care shortage between 17,800 and 48,000 • AAMC report • Retirements getting worse with COVID and burnout.
    • 13. 13 DEPARTMENT OF Biomedical Informatics Not just a US Problem
    • 14. 14 DEPARTMENT OF Biomedical Informatics + 14 AI?
    • 15. 15 DEPARTMENT OF Biomedical Informatics 15
    • 16. 16 DEPARTMENT OF Biomedical Informatics
    • 17. 17 DEPARTMENT OF Biomedical Informatics 4 weeks later: CovidClinical.net & medRxiv…Nature Digitial Medicine
    • 18. 18 DEPARTMENT OF Biomedical Informatics • Early evidence of major pathologies • Countries more similar than sites within them Meanwhile…
    • 19. 19 DEPARTMENT OF Biomedical Informatics
    • 20. 20 DEPARTMENT OF Biomedical Informatics Fundamental antecedent questions to inform policy 20 Whose utilities are maximized? Whose medicine is authoritative? Dyad to Triad: Distinct New Party ? Who controls incorporation of patient data into model? Should consumers have access to medical advice directly? Who funds?
    • 21. 21 DEPARTMENT OF Biomedical Informatics 21
    • 22. 22 DEPARTMENT OF Biomedical Informatics 22
    • 23. 23 DEPARTMENT OF Biomedical Informatics Implications • Sea-change in healthcare with AI º Faster than regulation º Faster than education • Acceleration of clinical and basic research • Data + computational resources are obvious king. • Globalization paths • Increases dangers of non-quant data exploration. • Urgency of exploration by educators and science leaders 23
    • 24. 24 DEPARTMENT OF Biomedical Informatics NEJM AI


    • Previous
    • Next
    • f Fullscreen
    • esc Exit Fullscreen