HL7’s 3rd annual conference on genomics, February 20-21 in Washington, DC, combines professionals from multiple areas – technologists, academic research and commercial vendors, while mixing in policy and government, all from a global perspective. This year’s theme is Global Clinical Genomics, Artificial Intelligence and Innovation.
The meeting will feature working sessions resulting in a conference position statement and recommendation paper to jumpstart the conversation that will continue at the annual HIMSS convention.
The ‘Innovation’ Part of Genomics and AI
An informal, pre-conference survey was sent out asking opinions considering the top three ‘innovations’ and top three ‘issues’ of the conference theme. Following is a sample of the responses received.
Research and Discovery
Machine learning holds promise to accelerate gene discovery. “There are genomic data on thousands of unsolved rare disease patients. We need to be able to match to other unsolved cases when we don’t have a clear gene candidate,” says Zornitza Stark of Australian Genomics. Artificial intelligence (AI) could speed up and increase the accurate analysis of genomic data and the data analysis process “which is currently heavily manual and subjective and is the biggest bottleneck in clinical genomics,” she says. Many survey respondents added they expect machine learning to advance clinical variant interpretation to the point where there are no longer variants classified as ‘unknown significance’.
In clinical, the innovation list includes “pulling the right information at the right time to improve clinician workflow,” says Gillian Hooker of Concert Genetics. She continues, “for example, predicting when a genetic test/new interpretation of existing sequence might be relevant for a particular patient, guiding clinicians to the right test, factoring in insurance coverage and policies, extracting the necessary information to allow insurers to make coverage decisions, and guiding clinicians to the right follow up care plan once results come in.”
“Extracting clinically-relevant or actionable information from the diversity of EHRs/EMRs” is another advancement, according to Jacques Beckmann from the Université de Lausanne in Switzerland.
Data Use and Analysis
“Personalized treatment recommendations augmented by low-cost genomic data for a broad range of clinical conditions” is a plus from Lonnie Northrup of Intermountain Healthcare. Applying machine learning to clinical decision support (CDS) and CDS algorithm development is another innovation according to Stark.
Scott Kahn of Precision Medicine Now added that “thriving knowledgebase networks that effectively share data for the diagnosis of the next new patient” would be innovative. A cardiologist, Dr. Jeff Anderson, also at Intermountain Healthcare, said the “ability to do whole genome sequencing (WGS) in cases with an uncertain phenotype/genotype followed by the ability to pull out likely or potential pathogenic variants of many variants found in the WGS” would be a great advance.
New Processes and Knowledge
New on the horizon could be “the integration of real world data from wearables such that the data can be included in clinical decision making,” suggests Kahn. Lynn Dressler at Mission Health brought forward these three points to make clinical genomics smarter:
- develop personal apps for professionals and patients
- have discrete results in the EMR over the lifetime of the patient
- integrate outcome and cost effectiveness
Add Your Ideas
Do you have other thoughts on genomics and AI innovation? Join the discussion by emailing me at firstname.lastname@example.org. You can also register for the conference and join us in Washington, DC! Look for the next blog post which will cover the ‘issues’ to solve.