ARTIFICIAL INTELLIGENCE
( continued from page 7 ) became unstable . We do have to acknowledge the many limitations - think of ECG false negatives and false positives . Most of us have experienced alarm fatigue from ventilators and telemetry machines , but ML may help with these false alarms . The winning ML model in a 2015 competition to improve alarms suppressed 80 % of false alarms but only 1 % of true alarms .
From the University of Utah : “ Science can tell you how to clone a tyrannosaurus rex . Humanities can tell you why this might be a bad idea .”
Machines , like humans , can make mistakes . As of 2019 , most patients do not trust the idea of medical AI , even when shown evidence of these systems outperforming physicians . Datafication is the “ conversion of qualitative aspects of life into quantitative data .” What do we lose in that conversion ? One study of a ML algorithm found that “ asthma is protective against pneumonia ,” a nonsensical conclusion cleared up when humans noticed that asthmatics were more likely to go to the ICU and therefore receive more comprehensive care . A widely used AI system produced racial bias when modeling decisions on insurance claims ; thus , AI tools could make health inequalities worse . We can also become complacent due to overreliance on a technology , known as automation bias . Think about sending a prescription and assuming the EMR will catch a drug allergy , or how lazily you look behind your car when backing out of a parking spot now that newer cars have cameras and alarms . Computers may cut corners because of incentives to meet data targets , a phenomenon known as reward hacking that can harm patients . This may be analogous to Goodhart ’ s law : “ When a measure becomes a target , it ceases to be a good measure .” Can you program a computer to understand the literary wisdom of Goodhart ’ s law ?
When considering the multifactorial challenges facing EM , AI shows promise to help physicians provide effective care . EM physicians must become familiar with AI tools when , but not if , some become large-scale medical applications . The desired outcome should be a synthesis that preserves real human connection between doctors and patients , not just better numbers on the spreadsheet . Israni and Verghese ask : “ Beyond easing the cognitive load and , at times , the drudgery of a busy practice , can AI help clinicians become better at being human ?” To the extent that AI can remove obstacles to this connection , it could represent a catalyst . In The Catalyst , Jonah Berger notes that chemical reactions often use heat and pressure to bring about change . Special substances speed up the process . “ But rather than upping the heat or adding more pressure , [ catalysts ] provide an alternate route , reducing the amount of energy required for reactions to occur .” Like system changes that prevent physician burnout , AI could remove roadblocks between patients and doctors . Sadly , most technology in medicine does the opposite .
Emergency medicine has been called the most exciting 15 minutes of every medical specialty . We quickly form intense , deeply human bonds with vulnerable patients . AI does not yet have intuition . But we cannot turn away from the future . AI can assist with stable EM processes such as EMS triage and ECG interpretation , but also for unexpected and unfamiliar challenges . The COVID-19 pandemic continues to test the resolve and aggregate well-being of most health care workers . What have we learned from this pandemic ? What have the machines learned , and what can they teach us ? How can we combine forces to be better prepared for future pandemics ? At the intersection of statistics and computer science , AI can analyze giant data sets to help us learn from our history and advance to a better future . Just like the impressive connectivity in the brain , maintained connectivity between humans will be the key to success .
References
1 . Berlyand Y , Raja AS , Dorner SC , et al . How artificial intelligence could transform emergency department operations . The American journal of emergency medicine . 2018 ; 36 ( 8 ): 1515-1517 .
2 . Grant K , McParland A , Mehta S , Ackery AD . Artificial Intelligence in Emergency Medicine : Surmountable Barriers With Revolutionary Potential . Annals of emergency medicine . 2020 ; 75 ( 6 ): 721-726 .
3 . Kirubarajan A , Taher A , Khan S , Masood S . Artificial intelligence in emergency medicine : A scoping review . J Am Coll Emerg Physicians Open . 2020 ; 1 ( 6 ): 1691-1702 .
4 . Lee S , Mohr NM , Street WN , Nadkarni P . Machine Learning in Relation to Emergency Medicine Clinical and Operational Scenarios : An Overview . The western journal of emergency medicine . 2019 ; 20 ( 2 ): 219-227 .
5 . Stewart J , Sprivulis P , Dwivedi G . Artificial intelligence and machine learning in emergency medicine . Emergency medicine Australasia : EMA . 2018 ; 30 ( 6 ): 870-874 .
6 . Taylor RA , Haimovich AD . Machine Learning in Emergency Medicine : Keys to Future Success . Acad Emerg Med . 2021 ; 28 ( 2 ): 263-267 .
7 . Vearrier L , Derse AR , Basford JB , Larkin GL , Moskop JC . Artificial Intelligence in Emergency Medicine : Benefits , Risks , and Recommendations . The Journal of emergency medicine . 2022 .
8 . Israni ST , Verghese A . Humanizing Artificial Intelligence . Jama .
2019 ; 321 ( 1 ): 29-30 . 9 . https :// en . wikipedia . org / wiki / Artificial _ intelligence 10 . https :// en . wikipedia . org / wiki / AI _ effect 11 . https :// nautil . us / why-is-the-human-brain-so-efficient-7216 / 12 . Deo RC . Machine Learning in Medicine . Circulation . 2015 ; 132 ( 20 ): 1920-
1930 . doi : 10.1161 / CIRCULATIONAHA . 115.001593
13 . Connor CW . Artificial Intelligence and Machine Learning in Anesthesiology . Anesthesiology . 2019 ; 131 ( 6 ): 1346-1359 . doi : 10.1097 / ALN . 0000000000002694
14 . Handelman GS , Kok HK , Chandra RV , Razavi AH , Lee MJ , Asadi H . eDoctor : machine learning and the future of medicine . J Intern Med . 2018 ; 284 ( 6 ): 603-619 . doi : 10.1111 / joim . 12822
15 . Watson NF , Fernandez CR . Artificial intelligence and sleep : Advancing sleep medicine . Sleep Med Rev . 2021 ; 59:101512 . doi : 10.1016 / j . smrv . 2021.101512
16 . Miller DD , Brown EW . Artificial Intelligence in Medical Practice : The Question to the Answer ?. Am J Med . 2018 ; 131 ( 2 ): 129-133 . doi : 10.1016 / j . amjmed . 2017.10.035
17 . Hamet P , Tremblay J . Artificial intelligence in medicine . Metabolism .
2017 ; 69S : S36-S40 . doi : 10.1016 / j . metabol . 2017.01.011
Dr . Huecker is a practicing University of Louisville emergency medicine physician and faculty member .
Jacob Shreffler , PhD , is an Assistant Professor in the Department of Emergency Medicine . He has published papers on a variety of topics including substance use , wellness , and creativity .
8 LOUISVILLE MEDICINE