In order to innovate there is no longer any need to choose between patient privacy and research requirements. Olivier Breillacq founded Octopize in 2018, so as to avoid this dilemma, with which he was confronted on a daily basis with MethodOmics, his first health data analysis business. “There is a certain social consensus that data need to be shared, with the consent of the individuals concerned, but in reality, things are very complicated. What was missing was the right technology.”
This technology is avatarization. It has won over not only pharmaceutical companies, but also laboratories, hospitals, and even the Health Data Hub, the body responsible for managing French health data. It saves project owners between 12 and 18 months.
“Our vision, which respects people’s personal data, answers an ethical challenge facing society, while remaining at the service of all. I am confident that within 10 years we will only be using synthetic data for secondary uses. We are just at the start of this revolution,” Olivier Breillacq, Founder and Director of Octopize.
This technology, named Avatar, is backed up by several patents, with the unique market feature of offering metrics that guarantee both the security and the value of the data throughout the process. These safeguards demonstrate that the avatars comply with the three criteria defined by the EDPS (European Data Protection Committee) – individualization, correlation and inference.
The method of the software marketed by Octopize was approved by the CNIL (Commission nationale de l’informatique et des libertés) in June 2020. This avatar system generates fictitious synthetic profiles from source user or patient profiles, which retain the same qualities and values as the original data, without any aggregation or masking. In accordance with the GDPR (General Data Protection Regulation), these synthetic data can be used without any time limit regarding academic research, preliminary studies prior to the launch of a drug or a medtech solution, to evaluate the effect of a drug, generate statistics, etc. “The avatar is totally relevant from a statistical point of view, but prevents the source patient being identified,” says Olivier Breillacq.
Octopize thus worked with the AP-HP to build a large database of the 20% of Covid 19 patients who trigger respiratory complications. The goal was to promote the research and development of machine learning algorithms to establish a severity score and prevent the most severe cases.
In September 2021, the startup raised €1.5 million to deploy its technology beyond the healthcare sector. Partnerships are already in place with the automotive sector and telecommunications companies. Octopize is also an award winner of the I-Nov program. In addition to half a million euros, the company will receive financial support from the French government to develop its solutions. It hopes to become the European leader in the field of data anonymization.
