Big data, machine learning, and precision medicine come together with GNS Healthcare’s Causal Machine Learning. In the GNS model, big data provides the means, machine learning is the method, and precision medicine is the goal. The first two elements are absolutely essential to have any chance for the third. And note the use of the term “infers” in the headline. Causal machine learning models go outside or “beyond” correlations, but there’s no guesswork involved and nothing random about how the model works.
Precision medicine is “precise” in identifying treatment plans tailored for specific patients. In order to approach patient singularity, the causal machine needs access to data points of all types including EMR, genetic, proteomic, genomic, claims, consumer, laboratory, prescription, mobile health, sociodemographic and more. There are no limits to the variety of data points that can be included in the model.
The GNS machine learning AI platform is REFS (Reverse Engineering Forward Simulation). REFS reconstructs diseases into computer models that are then used to simulate a range of real-world scenarios to discover the most effective treatments.
GNS uses aggregated data to build disease models that reveal the root causes of disease progression and patient response in a diverse patient universe. The strength of the disease models depends on the quality and quantity of patient data. When the patient population for a specific disease grows, the model gets stronger. To date, GNS has built more than 40 disease models based on more than 4 billion data points. Most of the models are in areas of oncology, cardiovascular and metabolic, neurology, and autoimmune diseases.
Precision medicine’s primary real-world application — at least for now — is identifying the pharmaceuticals that will have the best chance of improved individual patient health outcomes and reduced total cost of care. The information can include dosage and duration plus any other indicated therapeutics, procedures, and care management interventions.
When a specialist suggested a medication dosage a few years ago, he said to me, “In my experience, men your age seem to have the best results with this amount.” In a sense, he was using a form of precision medicine based on his several decades working with patients who had a specific condition. This system scales that concept to encompass much more than any single physician’s experience.
GNS has worked with pharmaceutical companies, health plans, and patient foundations for more than 17 years. Precision medicine goes much wider and deeper than any single practitioner and the GNS model for matching the right treatment to the individual has great promise for personalized, precision medicine.
