Orange Healthcare and Sanoïa endorse the use of machine learning to monitor chronic inflammatory rheumatism

Share

Orange Healthcare, Sanoïa and the Pitié-Salpêtrière hospital in Paris demonstrate that flare-ups in inflammatory rheumatism (rheumatoid arthritis and axial spondyloarthritis) can be detected using an activity tracker combined with machine learning technology.

At the Annual Meeting of the American College of Rheumatology which took place in San Diego from November 3–8, 2017, Sanoïa and Orange Healthcare took part in a scientific presentation in collaboration with Professor Laure Gossec, from the rheumatology department at the Pitié-Salpêtrière hospital in Paris: an analysis of 15 million information points from a cohort of 170 patients monitored over 3 months, conducted by machine learning (*) (Act-Connect study).

Data scientists at Orange Labs used an in-house machine learning tool (Khiops ©) to develop a model that detects flare-ups in the condition with a reliability rate of 96%.

A promising trial

Using and analyzing anonymized data gathered from connected medical objects, the results of this trial were very promising. According to Élie Lobel, CEO of Orange Healthcare: “The conclusions of this study are the result of cross-fertilization of expertise from industry, clinical research (CRO) and health professionals, demonstrating our ability to accelerate the development of services adapted to the monitoring of chronic diseases.” This technique illustrates how artificial intelligence can be used in the healthcare domain. It can contribute to:

  • The care system: it enables closer monitoring of the patient, through telemedicine or the scheduling of appointments around the activity of the disease,
  • Clinical research: it offers continuous and real-time access to certain patient data. This data indicates the frequency of flare-ups and acts as a measure of the effectiveness of drugs in rheumatology.

Professor Laure Gossec says: “putting the patient at the center of their care is our priority. Having access to digital tools that are easy to use, which can quantify a patient’s everyday experience and transform them into clinical indicators is extremely innovative.”

“The healthcare ecosystem has long sought to rely on connected objects, so as to be able to take full advantage of the data they provide in a medical context, as a source of information and predictability. This trial, which combines agility with scientific rigor, demonstrates how this is feasible in practical terms. We will now incorporate these outcomes in the Digital CRO we offer to sponsors of research,” concludes Hervé Servy, CEO of Sanoïa.

(*) “Machine learning” is a field of study within artificial intelligence that provides a computer or a machine with a method of automated learning, enabling it to carry out a number of difficult or demanding tasks. The goal is to make the machine or computer capable of providing solutions to complex problems, by processing a huge volume of information. This offer thus makes it possible to analyze and identify correlations between two or more specific situations, and predict their various outcomes.

Find out more about machine learning at: http://www.lebigdata.fr/machine-learning-et-big-data