دورية أكاديمية
Human visual explanations mitigate bias in AI-based assessment of surgeon skills.
| العنوان: | Human visual explanations mitigate bias in AI-based assessment of surgeon skills. |
|---|---|
| المؤلفون: | Kiyasseh D; Department of Computing and Mathematical Sciences, California Institute of Technology, California, CA, USA. danikiy@hotmail.com., Laca J; Center for Robotic Simulation and Education, Catherine & Joseph Aresty Department of Urology, University of Southern California, California, CA, USA., Haque TF; Center for Robotic Simulation and Education, Catherine & Joseph Aresty Department of Urology, University of Southern California, California, CA, USA., Otiato M; Center for Robotic Simulation and Education, Catherine & Joseph Aresty Department of Urology, University of Southern California, California, CA, USA., Miles BJ; Department of Urology, Houston Methodist Hospital, Texas, TX, USA., Wagner C; Department of Urology, Pediatric Urology and Uro-Oncology, Prostate Center Northwest, St. Antonius-Hospital, Gronau, Germany., Donoho DA; Division of Neurosurgery, Center for Neuroscience, Children's National Hospital, Washington DC, WA, USA., Trinh QD; Center for Surgery & Public Health, Department of Surgery, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA., Anandkumar A; Department of Computing and Mathematical Sciences, California Institute of Technology, California, CA, USA., Hung AJ; Center for Robotic Simulation and Education, Catherine & Joseph Aresty Department of Urology, University of Southern California, California, CA, USA. ajhung@gmail.com. |
| المصدر: | NPJ digital medicine [NPJ Digit Med] 2023 Mar 30; Vol. 6 (1), pp. 54. Date of Electronic Publication: 2023 Mar 30. |
| نوع المنشور: | Journal Article |
| اللغة: | English |
| بيانات الدورية: | Publisher: Nature Publishing Group Country of Publication: England NLM ID: 101731738 Publication Model: Electronic Cited Medium: Internet ISSN: 2398-6352 (Electronic) Linking ISSN: 23986352 NLM ISO Abbreviation: NPJ Digit Med Subsets: PubMed not MEDLINE |
| أسماء مطبوعة: | Original Publication: [London] ; New York : Nature Publishing Group, [2018]- |
| مستخلص: | Artificial intelligence (AI) systems can now reliably assess surgeon skills through videos of intraoperative surgical activity. With such systems informing future high-stakes decisions such as whether to credential surgeons and grant them the privilege to operate on patients, it is critical that they treat all surgeons fairly. However, it remains an open question whether surgical AI systems exhibit bias against surgeon sub-cohorts, and, if so, whether such bias can be mitigated. Here, we examine and mitigate the bias exhibited by a family of surgical AI systems-SAIS-deployed on videos of robotic surgeries from three geographically-diverse hospitals (USA and EU). We show that SAIS exhibits an underskilling bias, erroneously downgrading surgical performance, and an overskilling bias, erroneously upgrading surgical performance, at different rates across surgeon sub-cohorts. To mitigate such bias, we leverage a strategy -TWIX-which teaches an AI system to provide a visual explanation for its skill assessment that otherwise would have been provided by human experts. We show that whereas baseline strategies inconsistently mitigate algorithmic bias, TWIX can effectively mitigate the underskilling and overskilling bias while simultaneously improving the performance of these AI systems across hospitals. We discovered that these findings carry over to the training environment where we assess medical students' skills today. Our study is a critical prerequisite to the eventual implementation of AI-augmented global surgeon credentialing programs, ensuring that all surgeons are treated fairly. (© 2023. The Author(s).) |
| References: | Nat Med. 2021 Jan;27(1):136-140. (PMID: 33442014) Nat Commun. 2021 Jul 9;12(1):4221. (PMID: 34244504) J Urol. 2011 Apr;185(4):1191-7. (PMID: 21334030) Nat Commun. 2022 Aug 6;13(1):4581. (PMID: 35933408) J Surg Oncol. 2021 Aug;124(2):221-230. (PMID: 34245578) Am J Surg. 2001 Jun;181(6):484-6. (PMID: 11513770) Nat Med. 2022 Jun;28(6):1159-1160. (PMID: 35710994) Int J Comput Assist Radiol Surg. 2018 Dec;13(12):1959-1970. (PMID: 30255463) J Urol. 2018 Jan;199(1):296-304. (PMID: 28765067) J Urol. 2009 Sep;182(3):1126-32. (PMID: 19625032) J Surg Educ. 2020 May - Jun;77(3):491-494. (PMID: 31954662) Commun Med (Lond). 2021 Aug 23;1:25. (PMID: 34522916) Prostate Cancer Prostatic Dis. 2012 Mar;15(1):1-7. (PMID: 21844888) Chirurg. 2022 Mar;93(3):223-233. (PMID: 35147728) J Urol. 2011 Sep;186(3):1019-24. (PMID: 21784469) NPJ Digit Med. 2022 Jul 19;5(1):100. (PMID: 35854145) Sci Adv. 2022 Aug 12;8(32):eabq6147. (PMID: 35960806) Surg Endosc. 2021 May;35(5):2104-2109. (PMID: 32377839) Med Image Anal. 2022 Feb;76:102306. (PMID: 34879287) J Endourol. 2010 Feb;24(2):261-6. (PMID: 20073551) Nat Med. 2022 Jul;28(7):1455-1460. (PMID: 35864252) Eur J Cardiothorac Surg. 2022 Jan 24;61(2):239-248. (PMID: 34601587) Urol Pract. 2022 Nov;9(6):532-539. (PMID: 36844996) Nat Biomed Eng. 2023 Mar 30;:. (PMID: 36997732) Science. 2019 Oct 25;366(6464):447-453. (PMID: 31649194) Eur Urol Focus. 2022 Mar;8(2):613-622. (PMID: 33941503) Commun Med (Lond). 2023 Mar 30;3(1):42. (PMID: 36997578) Ann Surg. 2005 Feb;241(2):364-72. (PMID: 15650649) JAMA Netw Open. 2020 Mar 2;3(3):e201664. (PMID: 32227178) J Surg Educ. 2019 Nov - Dec;76(6):e1-e14. (PMID: 31601487) BMJ. 2003 Nov 1;327(7422):1032-7. (PMID: 14593041) Nat Med. 2021 Dec;27(12):2176-2182. (PMID: 34893776) BJU Int. 2012 Sep;110(6):870-4. (PMID: 22313582) J Gen Intern Med. 2019 May;34(5):712-719. (PMID: 30993611) Acad Med. 2015 May;90(5):549-52. (PMID: 25674911) Br J Surg. 2012 Dec;99(12):1610-21. (PMID: 23034658) |
| معلومات مُعتمدة: | K23 EB034110 United States EB NIBIB NIH HHS |
| تواريخ الأحداث: | Date Created: 20230330 Latest Revision: 20230509 |
| رمز التحديث: | 20230509 |
| مُعرف محوري في PubMed: | PMC10063676 |
| DOI: | 10.1038/s41746-023-00766-2 |
| PMID: | 36997642 |
| قاعدة البيانات: | MEDLINE |
Find this article in full text from Springer Verlag. 
Full Text Finder كن أول من يترك تعليقا!