Overview
Pre-Module Survey
Module
Recommendations
Cases
Post-Module Survey
References
About Module
Overview
Pre-Module Survey
Module
Recommendations
Cases
Post-Module Survey
References
About Module
Cases
View each of the 3 cases below and answer the corresponding quesetions in the discussion pages. Each case is modeled after real algorithms that have been developed.
Your browser does not support HTML video.
[Download Transcript for Case1A]
Case1A Question 1: What are some of the pros and cons of implementing the algorithm for predicting COVID-19 severity?
Case1A Question 2: How would you interpret results from the algorithm?
Case1A Question 3: What else would you want to know before deciding whether to use the algorithm?
Case References
Singh K, et al. Evaluating a widely implemented proprietary deterioration index model among hospitalized patients with COVID-19. Ann Am Thorac Soc. 2021 Jul;18(7):1129-1137.
Ross C. Hospitals are using AI to predict the decline of Covid-19 patients - before knowing it works. STAT. 24 Apr 2020. Cited 4 Apr 2022.
Drees J. Epic pays hospitals that use its EHR algorithms, report finds. Beckers Hospital Review. 26 Jul 2021. Cited 4 Apr 2022.
Your browser does not support HTML video.
[Download Transcript for Case1B]
Case1B Question 1: Does this study change your initial impression of how to interpret the results of the algorithm and, if so, how?
Case1B Question 2: Is this algorithm useful in real clinical practice and why or why not?
Case References
Singh K, et al. Evaluating a widely implemented proprietary deterioration index model among hospitalized patients with COVID-19. Ann Am Thorac Soc. 2021 Jul;18(7):1129-1137.
Ross C. Hospitals are using AI to predict the decline of Covid-19 patients - before knowing it works. STAT. 24 Apr 2020. Cited 4 Apr 2022.
Drees J. Epic pays hospitals that use its EHR algorithms, report finds. Beckers Hospital Review. 26 Jul 2021. Cited 4 Apr 2022.
Your browser does not support HTML video.
[Download Transcript for Case2]
Case2 Question 1: Can you point out and describe some positive aspects of this algorithm?
Case2 Question 2: What are some other clinical scenarios where algorithms might be helpful in reducing bias?
Case References
Pierson E, Culter DM, Leskovec J, Mullainathan S, Obermeyer Z. An algorithmic approach to reducing unexplained pain disparities. Nature Medicine 2021;27(1):136-140.
Your browser does not support HTML video.
[Download Transcript for Case3]
Case3 Question 1: How would you handle advertising of commercial AI products?
Case3 Question 2: What are the pros and cons of using VA data to train AI algorithms?
Case References
Robins R. AI systems are worse at diagnosing disease whent training data is skewed by sex. STAT. 25 May 2020. Cited 4 Apr 2022.
Suleyman M, King D. Using AI to give doctors a 48-hour head started on life-threatning illness. DeepMind. 31 Jul 2019. Cited 4 Apr 2022.
TomaĊĦev N et al. A clinically applicable approach to continuous prediction of future acute kidney injury. Nature 2019;572(7767):116-119.
