Pharmacogenetics, cancer genetics, cardiac genetics, WGS… AI-aided genetic counseling for healthcare organizations and genetic testing companies
Artificial Intelligence (AI) is gaining in the popular consciousness. Last week, the White House announced The American AI Initiative with new investment, government AI R&D resources opened up, establishing AI governance and emphasizing education and training.
Perhaps by coincidence (or perhaps not) the UK’s National Health Service (NHS) Health Education England (HEE) published The Topol Review with the subtitle “Preparing the healthcare workforce to deliver the digital future”. The report “makes recommendations that will enable NHS staff to make the most of innovative technologies such as genomics, digital medicine, artificial intelligence and robotics to improve services.”
Over a year in preparation, it has been called an ‘unprecedented vision of medicine’s future’ and ‘acknowledges the critical role of a literate health professional’. Wherever you are among this list of innovations as it relates to healthcare, the way healthcare is delivered is set to change, and AI will be a part of that.
The areas needed for genetic counseling
There is a shortage of genetic counselors available. In a 2017 JAMA study of 2451 women diagnosed with Stage 0 – Stage II breast cancer, the following results were reported:
“In this large, population-based study, most patients reported wanting genetic testing and 29% reported having it. Yet only 39.6% of all high-risk women and 61.7% of tested high-risk women reported having a genetic counseling session. This suggests a gap between need and availability of genetic counseling.”
JAMA. 2017;317(5):531-534. doi:10.1001/jama.2016.16918
It was only this week that the American Academy of Breast Surgeons recommended multi-gene genetic testing of individuals with breast cancer to assess hereditary risk. With lower cost and wider availability of tests for breast cancer (in the US, NCI’s SEER estimates 266,000 new cases in 2018) how will all these individuals get the needed guidance from a qualified medical professional (i.e. a genetic counselor)?
This is only for cancer genetics; the list for other genetic conditions and disorders is much longer.
What about pharmacogenetics? For those unaware, the costs of gene testing for metabolism have plummeted in recent years widening adoption, but gaps in implementation remain; groups like the Clinical Pharmacogenetics Implementation Consortium seek to ‘facilitate use of pharmacogenetic tests for patient care’.
What about cardiac genetics? You’ve undoubtedly heard of young people dying suddenly of an undiagnosed inherited cardiac condition – screening for inherited heart conditions is becoming more widely available. Dr. Birget Funke of Massachusetts General Hospital, Boston US has said in CAP Today “These are very prevalent disorders with a strong genetic basis and a potentially severe outcome. So genetic testing for both diagnostic and predictive purposes is important.”
The list goes on: reproductive genetics (and pre-implantation genetics), rare-disease genetics, pediatric genetics: how can healthcare professionals provide better care of patients through better access to genetic counselling?
How AI helps human judgement
The answer may be in Artificial Intelligence. There are services that provide phone-based health expertise (Teladoc is one of many that come to mind) along with genetic counselling, however these are appointment-based. Surveys show that live chat is a preferred method of communication over the telephone, and email (41%, versus 32% for telephone and 23% for email).
AI may be deployed best as an augment to human judgement. This is preferred over a wholesale replacement of humans via automated ‘chatbots’. Remember the famous Microsoft Twitter chatbot ‘Tay’ from 2016, designed to interact with millennials only to devolve into racist ranting. Rather than unsupervised, automated AI interaction, AI can be leveraged as an assistant, in real-time, to qualified medical professionals. In DNAfeed’s case, the qualified medical professional is a licensed genetic counsellor.
DNAfeed’s scalable business in volume, subject matter expertise, and even language
DNAfeed is a private company headquartered in San Diego, and offers AI-assisted genetic counseling as an on-demand service. They offer live chat interactions for counselling (appointment-based phone consultations naturally are also offered) to genetic test providers (for genetic counseling) and are expanding into pharmacogenetics to offer both genetic test consultations as well as clinical pharmacist consultations for healthcare organizations that need pharmacogenetics consultations.
The AI-assistance provided to their counselors occurs in real-time, invisible to the patient receiving the counseling. The advice is from a licensed counselor – while the AI-engine on the back-end provides suggested responses that a counselor can choose to use, whether a phrase or a complete answer. DNAfeed’s staff is virtual, offering remote highly-skilled work on a flexible schedule (often the chat sessions are asynchronous, functioning more like email).
As mentioned with pharmacogenetics, a licensed pharmacist can also be enabled on the same platform to consult directly with patients on questions related to specific drug metabolism genes as a result of a genetic test.
DNAfeed also has genetic testing companies and healthcare organizations located abroad who use their services; their chat sessions are translated in real-time. Isn’t technology great?
While not the main emphasis of DNAfeed’s business, individuals who have 23andMe, whole-exome (WES) or whole-genome sequencing (WGS) performed can also use DNAfeed’s consultants and their on-demand service. Much, much better than a DIY approach; a human will be able to interpret nuances that an automated database simply cannot match. However for the intrepid who have this data and give these four databases a whirl, nothing is stopping you. For the rest of us, $79 for 30 days of online chat is quite a lot of value from a licensed professional. https://www.dnafeed.com/