The availability and accessibility of healthcare and molecular data continues to increase, with new sources emerging on a seemingly ongoing basisAccessing data is only the first step to gaining valuable knowledge and insights. Powerful analytics are an essential component to deriving value from data assets. GNS is a leader in developing custom analytic solutions tailored to organizations’ specific needs to identify meaningful insights and optimize the value generated from this wealth of information.

GNS has partnered with many of the world’s largest and most innovative biopharmaceutical companies, foundations, academic medical centers, and health plans to create custom models that extract new information and value from health-related data.

In collaborations with its partners for the past 16 years, GNS has applied our causal machine learning platform, REFS, to 100+ projects that are applicable to all stages of the product development life-cycle to advance understanding of disease progression, drug response, and the commercialization of new products. The learnings derived from these projects have helped guide decisions that advance the entire value-chain of the pharmaceutical industry – from early-stage R&D through commercial operations.

 Custom Modeling Offerings from GNS include:

dataDisease Progression Models, Causal Drug Targets, & Combination Therapies

Uncover novel drivers of disease progression and discover the drugs to manage it.

Current medical knowledge and analytic tools struggle to identify drivers of disease in an efficient manner. More data doesn’t necessarily translate to more discoveries. In order to find value in data, advanced analytical methods are often needed. GNS Healthcare’s machine learning platform discovers the underpinnings of disease through the creation of causal models. These models have been applied to numerous areas including neurology, oncology, cardiovascular, and autoimmune diseases. Furthermore, the causal models enable the identification of therapeutic targets, molecular characterization of patient subpopulations, and fundamental disease biology, lending rapid and valuable insights to pre-clinical and clinical development programs.  

progressionBiomarkers and Diagnostics for Patient Stratification in Clinical Trials and the Real World 

Stratifying patient subpopulations based on disease progression risk or expected response to therapy is an essential frontier of personalized medicine. The GNS causal machine learning and simulation platform discovers the most accurate and robust markers of disease risk and treatment response. These have been applied to both clinical trial and real-world settings, across diseases from various cancers to arthritis to depression to diabetes. This discovery of patient sub-populations for both efficacy and safety simultaneously improves evidence for optimal value demonstration and design of clinical development programs 

cogsEfficacy to Effectiveness

Predict real-world performance of investigational drugs.

The Efficacy-Effectiveness gap creates uncertainty around the translatability of clinically measured outcomes of drug therapies to the real world. By using the REFS platform to analyze Phase II/ Phase III trial data, GNS allows clients to make projections of real-world outcomes that anticipate future safety and effectiveness of newly developed therapies in real world populations and subpopulations.

scaleValue-Based Care and Comparative Effectiveness

Defining the value of treatments for individuals and populations.

Quantifying the value of interventions to safely treat disease in the real world remains a challenging problem for health economists, payers and pharmaceutical companies. The root of this problem is establishing the relative value of an intervention for different patient subpopulations that may experience different effectiveness or safety profiles of the therapy. Using our causal machine learning approach, GNS can identify the value of each therapy for a given patient and its immediate impact on short and long-term healthcare costs. The sum of these valuations can be then be forecasted within a patient population to give an overall value for a given therapy.


Discover what works. For whom.