REFS for a different kind of judgment call

GNS Healthcare, Covance to create computer models to predict drug development success

Kelsey Kaustinen
Register for free to listen to this article
Listen with Speechify
0:00
5:00
CAMBRIDGE, Mass.—In hopes of driving more productive drugdevelopment, leading healthcare data analytics company GNS Healthcare andCovance, one of the world's largest contract research organizations (CRO), haveannounced the establishment of a new collaboration. The companies will seek todevelop novel data-driven models that can aid pharmaceutical companies inoptimizing and streamlining the efficiency and cost effectiveness of their drugdevelopment activities. Financial details for the collaboration were notdisclosed.
 
"Our collaboration with Covance combines our uniquecollective resources and capabilities to tackle what has previously been anintractable challenge—improving dismal clinical drug development successrates," Colin Hill, president, CEO and co-founder of GNS Healthcare, said in apress release. "The predictive computer models arising from our collaborationwill address this problem and will in turn lead to better treatment options forpatients. The collaboration with Covance is a key part of our expandingportfolio of relationships with customers and partners employing the REFSplatform to learn what drugs work for whom in healthcare." 
 
The collaboration will combine GNS' supercomputer-drivenReverse Engineering and Forward Simulation (REFS) big data analytics platformand Covance's data assets, which have been drawn from the CRO's significantclinical and scientific support of drug development. Together, the companieswill develop computer models that are capable of predicting the likelihood ofsuccessful development for a drug candidate given available safety and efficacydata across a range of patient characteristics. The collaboration will seek todevelop these models in a variety of diseases, beginning with type 2 diabetes.
 
The REFS platform is a "scalable, supercomputer-enabledframework for discovering new knowledge from real world data," GNS notes on itswebsite, adding that the platform "automates the extraction of causal networkmodels directly from observational data and uses high-throughput simulations togenerate new knowledge." The REFS process consists of two steps, reverseengineering and forward simulation. REFS uses machine learning to extractunderlying structure from a data set, which is then encoded in the form ofcausal network models, which represent causal relationships, not justcorrelations.
 
 
Multiple different data types and multi-layered data are allincorporated. As many companies continue to face issues with late-stage failureof drugs that have cost millions to develop—a 2010 study by Tufts Universityestimated current approval rates, from Phase I to U.S. Food and DrugAdministration approval, at 16 percent—forecasts about success probability canhelp guide development decisions before clinical failure and investment losses.
 
Covance deferred additional questions about the deal to GNS,which did not respond to contact requests by press time.
 
 
The collaboration is the first of two GNS announced for themonth. A week after publicizing the deal with Covance, the company announcedthat it has established a new program with Aetna that will also make use ofGNS' REFS platform. Through the program, the companies will seek to treat andprevent metabolic syndrome by using REFS to identify Aetna members who are atrisk for metabolic syndrome sooner than current options, evaluate members whoalready have metabolic syndrome and determine effective treatment for membersindividually.
 
Metabolic syndrome consists of a compilation of three ormore of five health conditions, which include high blood pressure, hightriglycerides, low HDL ("good") cholesterol, high blood sugar and large waistsize. Those with metabolic syndrome are at increased risk for diabetes, heartdisease or stroke.
 
 
Having one or two of the conditions places a person at riskfor the syndrome. The computer models generated by REFS will extrapolate anAetna member's health information, looking at which conditions they havecurrently and which condition they are likely to develop next, and how quickly.Based on their risk factors, each member is then matched with specific healthinterventions that would be most effective for their condition. Depending onhow successful the program is in the indication of metabolic syndrome, GNS andAetna expect to move forward into additional major disease areas.

Kelsey Kaustinen

Subscribe to Newsletter
Subscribe to our eNewsletters

Stay connected with all of the latest from Drug Discovery News.

March 2024 Issue Front Cover

Latest Issue  

• Volume 20 • Issue 2 • March 2024

March 2024

March 2024 Issue