Purchase this article with an account.
Samuel I Marks, Grayson W Armstrong, David S Friedman; Devaluing privacy: medicine for AI in the public domain. Invest. Ophthalmol. Vis. Sci. 2022;63(7):2150 – A0178.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Advocate and demonstrate a public domain methodology for the collection, analysis, and use of medical data. This is as opposed to keeping patient data private and siloed; the norm. Although privacy is attractive, we posit insufficient attention is targeted on the advantages of its inverse: the public domain. Public domaining medical information enables the advance of knowledge and understanding (through data-pooling and disparate expert analysis), and its sharing, regardless of financial and physical positions.
Recruit ophthalmologists who are department directors to allow their patients to be screened without the siloed gatekeeper approach. Each individual patient must explicitly give their informed consent, and, following a grace period of 2 months, be unable to withdraw their consent. New public-domain software systems are designed to facilitate this process. The medicolegal and other bureaucratic requirements for running these studies will also be placed in the public domain to aid replication.
Initiating two studies at two departments at a major top-tier eye \& ear hospital in the United States, specifically: its ophthalmic emergency and glaucoma departments. Over 1,000 new patient records are expected to be collected and public-domained every month.
The culmination of patient data from disparate locations is greater than the sum of its parts… being a great boon for everything from epidemiological analyses to automated diagnoses through Machine Learning and Artificial Intelligence. Long term we hope others follow suit with the public domain methodology.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.
This PDF is available to Subscribers Only