We have made the following Saudi Arabia Email List observation: patients are more and more numerous, in particular with the development of chronic diseases, while the number of doctors and nurses is constant or even decreasing. There are mechanically fewer doctors per patient and therefore less availability for long-term follow-up. Another effect is added, medicine has become more and more sophisticated, with standards, data management, etc. Caregivers do not have time to perform administrative tasks such as entering data into tracking systems. We therefore wanted to respond to this challenge: how to offer a solution for the monitoring of patients that does not take a long time for caregivers?
On the one hand, the patient’s vital card brings together a very rich set of data which are primarily intended for reimbursement but which can also be used to obtain medical information useful to caregivers. We retreat and structure this flow to extract medical data useful for setting up personalized monitoring. For example, we can extract the data on the last appointments with a radiologist who gives us information on the type of act operated (for example a mammogram) that we will translate into a medical language. Then a prescription prescribed by a general practitioner will allow us to create a follow-up alert on the adverse effects or interactions between the two treatments prescribed by the radiologist
What is the concept developed by Semeia
The main difficulty to overcome is the research and structuring of the data. The data used by our solution comes from two main sources. and the general practitioner. Basically, the information provided by social security is “remedied”. These are reliable and robust data because they are linked to the billing of medical acts. The second main source of data comes from biological analyzes carried out in the laboratory. These are text files sent to the prescribing physician that we standardize and structure in a database. On these two sources, we apply management rules in order to identify processing and follow-up actions.
Finally, only nurses integrated into our network are authorized to access health data. For data scientists, we provide access to anonymized data, in a “secure bubble” where all their actions are recorded and traced. Thus, our teams cannot cross-reference data that would make it possible to identify patients. We have developed a server approved by the CNIL, which was long and expensive but now allows us to create relevant models in complete safety. Actually trained, I started my career in consulting; I carried out a lot of missions for health insurance. This is how I discovered the importance of the subject of data in health. Also, my parents are doctors, which allowed me to be aware of health issues very early on.
Concretely, how does Semeia work
We wish to broaden our scope of intervention to other pathologies; for example, the follow-up for lung transplants dates from October 2021 and for breast cancer from June 2021. In a hospital, the organization is done by departments according to the pathologies. This implies that each department equips itself with a tool adapted to its needs, which increases the number of tools within the hospital. This explains why the market is very fragmented, with competing companies specializing in pathology. Doctors exchange information among themselves and disseminate knowledge on the different solutions available as needed. In the more or less short term, we also want to open our service to city health professionals.
So the idea came to me to work on social security data in order to use the information useful for monitoring. We created the project during a hackathon organized by social security on the subject of the correct use of prescription data. I joined forces with my roommate at the time having an engineer profile and with a third employee with a commercial profile. We started the project in 2017. Today, we are 30 people, spread between Toulouse and Paris. We support 20 hospitals and around 1,000 patients. For this, we will go through the territorial professional health communities (CPTS) funded by the ARS (regional health agencies), the CPAM (Primary Health Insurance Caisses) and local authorities and which aim to structure the organization and the tools available to health professionals in a region.