What is Causal Machine Learning?

With the rise of big data, machine learning emerged as an important predictive tool in modern healthcare analytics and bioinformatics. The idea of machines learning from data they consume to make predictions isn’t new. What is new is the inference of causal mechanisms from patient data that enables “what if?” simulations of a variety of treatments and interventions across individual patients to determine optimal therapies.

GNS moves beyond analytical and machine learning approaches that rely on data correlations to match treatments to patients. We reverse engineer the complex causal mechanisms that determine which therapies will produce the best outcomes for each patient. We call this approach causal machine learning and it helps us unlock the full potential of the data and get at the underlying causal mechanisms of the “system” that produced the data.

The Importance of Data Integration

We know data – from traditional healthcare data to consumer, mobile, and emerging data. And we know the value of using all the data, all at once. Our use of massive, multi-modal health datasets is foundational to revealing cause-and-effect relationships in data – and to creating new knowledge of “what works for whom” directly from data.
We’ve been working with diverse and complex medical data since 2000, before the emergence of big data in healthcare. We routinely integrate disparate, multi-modal data types, into coherent, normalized data frames, including data types such as:
  • Genetic
  • Genomic and other -omic [gene expression, epigenomic, proteomic, metabolomics]
  • Molecular and laboratory
  • Electronic medical records (EMR)
  • Patient registry
  • Consumer demographic and behavioral
  • Pharmacy and medical claim
  • Features extracted from imaging and natural language
  • Emerging, novel, real-world, such as mobile health

No one knows healthcare data more intimately than we do. To get the full value from all your data, there simply is no better partner.

With GNS, you’re not just predicting the future. You’re finding ways to change it.

Learn More




The Power of “What If?gns-pyramid-03


Source: “Competing on Analytics,” by Tom Davenport.

Discover what works. For whom.