To the Editor:
I would like to comment on the article, “AI advances vision science, identifies disease patterns faster,” in the August 2018 issue of Primary Care Optometry News.
Karen F. Perry, OD, FAAO, was quoted as saying that it is widely believed that AI will improve clinical outcomes, increase clinician efficiency and decrease health care costs.
“I can see where, especially in optometry, if we aren’t embracing this technology from the outset, we may risk the profession being bypassed by legislation and insurance companies,” she said.
The legislation is here. The most developed artificial intelligence systems have worked with Google Brain’s AI and deep learning system to hone in on diabetic retinopathy and diabetic macular edema. In April, the FDA approved the IDx-DR, an AI device, “to detect more than a mild level of diabetic retinopathy in adults with diabetes,” as reported on Healio.com.
Basically, AI is becoming a version of telemedicine. A new system in development, DeepMind, will be able to make clinical referral suggestions classified as “urgent,” “semi-urgent,” “routine” and “observation only” (De Fauw et al.).
Joseph HallakSuch systems can be placed in primary care physician offices, in malls and airports or will find their way in smart phone apps. Given today’s technological advances, we would see this in the not-so-distant future.
It’s important to consider that widespread screening could increase referrals to eye care providers. However, consumers must know that it will not supplant the professional consultation.
A parallel to this situation, historically speaking, is the autorefractor. It is so widespread and an essential element in the examination, but it still did not replace a good manual refraction despite the relatively recent launch of smart phone apps that provide refractions.
The issue of legislation and insurance companies is more complicated. Legislation invites litigation. Can you litigate AI? For that we need to find a way of virtual/artificial litigation and compensate it with Bitcoins! It will be a long, arduous and bumpy road fraught with real hurdles.
- De Fauw J, et al. Nature Medicine. 2018;doi.org/10.1038/s41591-018-0107-6.
- Gulshan V, et al. JAMA. 2016;doi10.1001/jama.2016.17216.
- Healio.com. FDA allows marketing of IDx-DR artificial intelligence device for diabetic retinopathy detection. Posted April 11, 2018. https://www.healio.com/ophthalmology/regulatory-legislative/news/online/%7b41474ac4-1903-46ea-9464-4f448b881b64%7d/fda-allows-marketing-of-idx-dr-artificial-intelligence-device-for-diabetic-retinopathy-detection.
- Ross C. Opening the ‘black box,’ Google DeepMind AI system diagnoses eye diseases and shows its work. STAT Magazine. Posted Aug 13, 2017. https://www.statnews.com/2018/08/13/google-deepmind-ai-diagnoses-eye-diseases/.
Joseph Hallak, OD, PhD, FAAO
Disclosures:Hallak reports no relevant financial disclosures.