I am a pediatric cardiologist and Chief Intelligence and Innovation Officer of CHOC Sharon Disney Lund Medical Intelligence, Information, Investigation and Innovation Institute (Mi4). I am the founder of Artificial Intelligence in Medicine (AI-Med), the Medical Intelligence Society (MIS), the American Board of Artificial Intelligence in Medicine (ABAIM), the Alliance of Centers in Artificial Intelligence in Medicine (ACAIM) and the Pediatric Centers of Artificial Intelligence in Medicine (PCAIM). My book, Intelligence-Based Medicine is the first of its kind textbook on AI in medicine and is used at colleges and universities around the globe.
“As far as laws of mathematics (artificial intelligence) refer to reality, they are not certain; and as far as they are certain, they do not refer to reality”
The aforementioned quote is one of my very favorite quotes from the physics genius Einstein on the dynamic between mathematics and the duality of reality and certainty. With the rapid ascent of the capabilities of deep learning and transformers, perhaps clinician cognition is more important than ever before to reduce uncertainty and increase reality in decisions in healthcare. Several key concepts will be discussed in the following weeks:
These biases, initially introduced in Science in 1972 by Tversky and Kahneman, are systematic errors in thinking due to either the processing or the interpreting of the information available for the decision. While cognitive biases increase our thinking efficiency by enabling us to expedite decisions (perhaps evolving from our ancestors in threatening situations), these shortcuts can often lead to suboptimal and even erroneous decisions. Kahneman described this type of thinking as system 1 (a fast and intuitive thinking that is efficient but error-prone) and system 2 (a slower and deliberate thinking that is more mistake-proof.)
A major contributing factor in cognitive biases is the mental shortcuts that humans tend to rely on called heuristics. Additional causes of cognitive biases include inadequate attention (hence the current interest in “self-attention” mechanism of transformers) and/or memory, individual motivations and beliefs as well as emotional and physical state, and group and social influences.
The following are a few of the more common cognitive biases in our day-to-day situations including in clinical medicine and healthcare:
- Anchoring bias: over-reliance on the initial piece of information
- Availability heuristic: over-emphasis on information that is readily available
- Confirmation bias: focus on information that conforms to your current theory
- Correspondence bias: attributing behavior to personality or traits as a social judgment
- Hindsight bias: an incorrect perception that the outcome prediction was known all along
Not all cognitive biases lead to undesired outcomes in healthcare, but some do result in poor judgment and outcomes in clinical medicine. There can be neutralizing measures for unfavorable cognitive biases in healthcare, including self-awareness and countermeasures for biases, but perhaps artificial intelligence can be in a position to mitigate bias altogether. It is likely that artificial intelligence, if deployed strategically as a “system 2” resource, can help provide a balanced dyadic partnership to this human-driven cognitive process and its cognitive biases.
The importance of human cognition will be one of the topics of discussion at the in-person AIMed Global Summit 2023 scheduled for June 4-7th of 2023 in San Diego. The remainder of the week will be other exciting AI in medicine events like the Stanford AIMI Symposium on June 8th.
Hope to see you soon in California!
We at AIMed believe in changing healthcare one connection at a time. If you are interested in discussing the contents of this article or connecting, please drop me a line – [email protected]