Optra Health Launches 'Digital Genetic Assistant' for Explaining Genetic Tests
February 22, 2018
WASHINGTON (GenomeWeb) – This week, Optra Health, maker of the iPhronesis artificial intelligence platform for automated review of genetics literature and data, stepped into the consumer realm by introducing OptraGuru, an AI-based clinical report interpretation and guidance system.
San Jose, California-based Optra Health calls OptraGuru a "digital genetic assistant" that allows consumers and healthcare professionals alike to verbally query genetic data via Amazon Alexa and Microsoft's Cortana.
The OptraGuru system puts these consumer-grade digital assistants on the front end of iPhrenesis to help patients, genetic counselors, and bioinformaticians make sense of genetic test results — particularly those from direct-to-consumer testing companies Ancestry.com and 23andMe — in a conversational manner.
"One of the differences that we are trying to make in the industry is [in] how easily you can understand the genetic data and how can that be helpful to you," Optra Health Cofounder and Executive Chairman Abhi Gholap told GenomeWeb shortly after speaking at Health Level Seven International's annual Genomics Conference in Washington, DC.
There, Gholap gave the first public demonstration of the technology by speaking into an Alexa device attached to his laptop computer. The system seemed to work properly, explaining in lay terms the implications of specific gene variations from an Ancestry test.
Gholap said that there are "three clear applications" for OptraGuru: the direct-to-consumer tests; family genetic screening that might take place in a prenatal or neonatal environment; and tests ordered by clinicians in the context of diagnosing cancer or rare diseases. For now, though, Optra plans on concentrating on the first two.
"We have not gone much into the clinical test as such because we still [need to tune] our systems to that level," Gholap said. Plus, there are regulatory considerations for moving into the professional side of genomic interpretation.
Consumer genomics interests Optra because so many patients already struggle with basic health literacy, according to Gholap.
"We have chosen genomics as the field to begin with because I think there are more complexities," Gholap said.
"With genetic data, patients are becoming even more confused," he added. "What they are trying to understand from the [genetic] report is based on their own understanding of the medical sciences." That often is not very extensive.
"They will never be able to match the physician level" in terms of medical knowledge, Gholap said of typical consumers. But physicians — particularly those in primary care — often do not have the time during a patient visit to explain the health implications of a genetic test.
"That's why we were trying to build kind of a conduit where maybe the first level of questioning can be handled by the machine," Gholap said. In genetics and genomics, medical professionals often are asked many of the same questions repeatedly, and an automated database can take some of the pressure off them, he suggested.
Plus, the relatively young field of genomics is rapidly evolving and the body of knowledge is constantly changing. "Maybe there is a researcher in Spain who published something just a couple of hours back," Gholap said. "In our system, that is an advantage that you get. You are going to get very up-to-date information."
When Optra released its iPhronesis natural language processing app in March 2017, it only included a keyboard interface. But the company found that that strategy had limitations. "People feel more comfortable if they feel they are interacting with someone. That's where we extended to also interface with voice devices," Gholap said.
Instead of, for example, setting up a call center to augment the original system, Optra chose to leverage the voice-controlled smart speakers that are proliferating, in part because customer service representatives may not even be fully up to date, Gholap said.
Meanwhile, patients regularly search the internet to try to make sense of their test results. Google and other search engines, of course, do not always return trustworthy medical information, and even if they do, users do not always click on the most reliable sites.
"That's lacking the context," Gholap said. Some sound medical sites may be written in scientific rather than lay language, too, leaving patients with a partial understanding of their results.
"Half knowledge is always riskier," Gholap said. "You have to create a context, and that's why we created this."
Google ranks results by popularity, which is not necessarily the most medically relevant. Optra has about 75 different web crawlers in its Amazon Web Services cloud to look for and parse new, relevant information, according to Gholap. "Less human intervention is a key here," he said.
For quality control, Optra has a global team of about 40 geneticists to curate and verify data. The company has built an app for these professionals to notify them of new knowledge and give them a curation platform.
"They don't do every step of the curation," Gholap said. Some algorithms help weed out unreliable bits of information. The geneticists can choose to accept, reject, or modify reports that the app presents to them.
"Based on that algorithm, we are pulling the data, we are curating the data, we are parsing the data, we are verifying the data," Gholap said. Humans only have to intervene for quality control on some of the more technical points.
"Any information that has been given to any user [by OptraGuru] is not generally based on the number of hits or something, but is based on the relevance. It is something that we are parsing, so your answer becomes more contextual," Gholap said.
Gholap said that Optra is looking to interface the technology with additional digital assistants, including Google Assistant and Apple's Siri. "We would like to be agnostic of device commands," he said. With Siri, users would be able to tap into OptraGuru on their iPhones rather than needing an external device, he noted.
Optra, which has been working on its database of genomic knowledge for about two years, has some big names on its team. Scott Kahn, a former chief information officer of Illumina, serves as chief strategy advisor.
The chairman of Optra's science advisory board is Personal Genome Project founder and CRISPR pioneer George Church. "Dr. Church essentially brings this new concept to look at the [consumer] genomic data," Gholap said.
Patrick Soon-Shiong personally invested in Optra Health and became part of the company's board about a year ago. In 2016, Soon-Shiong's NantOmics invested in telepathology company OptraScan, a sister organization to Optra Health under the Optra Group umbrella.