BE PART OF THE REVOLUTION TRANSFORMING HEALTHCARE WITH AI - - PowerPoint PPT Presentation

Insert Logo, text, URLs, etc here 1000 ATTENDEES 80 SPEAKERS 10 WORKSHOPS 2 SOCIAL EVENTS 1 AIMed19

www.aimed.events/northamerica-2019/ #AIMed19

BE PART OF THE REVOLUTION

TRANSFORMING HEALTHCARE WITH AI

CALIFORNIA — THE RITZ-CARLTON, LAGUNA NIGUEL 11–14 DECEMBER 2019

Edward H. Shortliffe, MD, PhD; Professor (Adjunct, Retired) Columbia University Arizona State University Weill Cornell Medical College Senior Executive Consultant, IBM Watson Health AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019 www.aimed.events/northamerica-2019

Insert Photo Here

tedshortliffe @tshortliffe

AI AI i in M Med edicine: e: T The e Evolution of the Sc Science and Its Pra ractice

www.aimed.events/northamerica-2019/

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019

Go Goals f s for To Toda day’s Pr s Prese sent ntation

• Show that today’s state of the art in medical AI is

part of a 50-year scientific trajectory that is still evolving

• Emphasize how the evolution of digital technology

has affected our ability to field systems for routine use

• Summarize where we are today in that evolution

www.aimed.events/northamerica-2019/

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019

De Decisi sion Ma Making as s an In Intri rinsic c Part of

• f Med

edici cine

• Early assumption that “computers in medicine”

meant computers as diagnostic aids

• Subsequent focus on broader set of decision-

support needs – and on importance of information and knowledge representation

• Most physicians acknowledge that some “support”

would be useful to them, if “properly” delivered

Decision Making vs Decision Support

www.aimed.events/northamerica-2019/

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019

19 1950 50’s

• Earliest formal recognition of statistical issues in

diagnosis and the potential role of computers

Real-World Use of Bayes?

X

www.aimed.events/northamerica-2019/

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019

19 1960’s-70 70’s

• Substantial AI work on problems familiar to us today:
• Machine learning, neural networks
• Natural language processing
• Simulation of human problem solving
• Emergence of applications of artificial intelligence in

medicine (knowledge-based systems, also often dubbed “expert systems”)

www.aimed.events/northamerica-2019/

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019

Ar Artifi ficial Intelligence (AI AI)

The study of ideas that enable computers to do the things that make human beings seem intelligent: The ability to reason symbolically The ability to acquire and apply knowledge The ability to manipulate and communicate ideas

Taken from Parick Winston’s AI Textbook from the 1970s-80s

www.aimed.events/northamerica-2019/

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019

Exa xample AI Programs from 1970s

• Internist-1 / QMR

Task: Diagnosis in internal medicine and neurology

• The MYCIN System

Task: Advise clinicians on the selection of antimicrobial regimens in critical settings (especially before culture results)

• Approaches/methods broadly adopted outside of medicine

www.aimed.events/northamerica-2019/

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019

Exp xplosive Interest in AI

• Expert systems heavily covered in lay press, with

medical examples prominent

• Cover stories in leading magazines and

newspapers

• Investment by companies in learning about AI and

its potential application in their businesses

www.aimed.events/northamerica-2019/

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019

19 1980’s a s and 9 nd 90’s

• Disappointment that applied AI had not rapidly lived

up to predictions for impact

• Perceived “overselling” of artificial

intelligence

• Avoidance of use of the term by

AI researchers AI Winter

AI in Medicine 1993;5:93-106

www.aimed.events/northamerica-2019/

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019

20 2000-20 2010

• Integration of decision support with workflow viewed

as a central requirement

• Further stimulated by concerns regarding patient

safety and error reduction (IOM: “To Err is Human”)

• Increasing incorporation of decision-support

functionalities (mostly warnings and alerts) in commercial products

AI in Medicine 2009;46:5-17

www.aimed.events/northamerica-2019/

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019

20 2000-20 2010

• New issues regarding relationships between vendors and

hospital IT staffs

• Locally built systems give way to vendor-supplied

electronic health records

• Challenges in the incorporation of decision support and

knowledge-management tools

• Institutional barriers (beyond technology) that limit the

ability to share rules and guidelines between hospitals and health systems

www.aimed.events/northamerica-2019/

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019

Th The C Current D t Decade de

• Explosion in interest in AI broadly, and medical AI in particular
• Confusion in terms: “big data”, machine learning, data analytics,

data science

• How do these terms relate to AI and, in medicine and biology,

to biomedical informatics?

• Commercial products in the clinical AI and decision-support area
• Many of the early challenges persist (integration, interoperability,

terminology standardization, data ownership, and many others)

Th Thank k You

te ted@ d@sho hortlif tliffe.net