GNS Unveils Platform to Predict New Therapy’s Likely Success in Real World at ISPOR

Patricia Silva, PhD avatar

by Patricia Silva, PhD |

Share this article:

Share article via email
causal machine learning

GNS Healthcare is presenting a data-driven causal machine learning solution, called Efficacy to Effectiveness, designed to predict how potential therapies will actually perform in distinct populations. The data, being released today at ISPOR 2016, used pre-launch data from a study comparing Gilenya (fingolimod) and other multiple sclerosis (MS) therapies to build and validate causal models that estimate Gilenya’s likely performance on the market.

The platform, demonstrated in collaboration with Novartis (the developer of Gilenya) and Harvard University, is designed to help inform pharmaceutical makers’ pre-launch market strategies, and to create evidence supporting the benefits of a new therapy to specific groups of patients.

“Clinical trials are the gold standard for assessing efficacy, but the real world — where a drug is in competition with other therapies and is no longer constrained to a well-defined trial population — is ultimately where new therapeutics have to perform,” Iya Khalil, GNS’ chief commercial officer, executive vice president and co-founder, said in a press release.

GNS applies causal machine learning technology —  data integration architecture,  causal inference and a simulation engine — to predict which treatments will work best for given groups of patients before the treatments are actually released. According to the company, this “Efficacy to Effectiveness” approach helps to improve both patient outcomes and reduce costs.

The poster presentation, “Using Clinical Trial and Real World Data to Bridge Efficacy to Effectiveness of Fingolimod in Multiple Sclerosis Patients,” describes how Gilenya’s real-world outcomes might be predicted using pre-launch data from a retrospective study of Gilenya clinical trial data and observational administrative claims data.

“This work shows that, by leveraging a combination of causal machine learning and pre-launch data, launching a new therapeutic without visibility into its real-world performance can be a thing of the past,” Khalil said.

ISPOR 2016, the International Society for Pharmacoeconimcs and Outcomes Research 19th Annual European Congress, opened in Vienna, Austria, on Oct. 29 and runs through Nov. 2.