Portrait - Laetitia Chanas, Manager of the Health data consulting and services team
A trained biologist, with a PhD in bioinformatics, Laetitia Chanas began her career in a start-up which developed a molecular modelling software. She then worked for Inserm at Orphanet, a unit dedicated to referencing rare diseases. In 2004 she joined the U900 bioinformatics unit at Institut Curie where she works on developing a tool to help researchers make queries on the gene sequencing platform.
Today in her position in the Institut Curie Data Department, she is responsible for coordinating the Health data consulting and services team, and for interfacing with users. "I am the contact person for people from the hospital group or from research wishing to use clinical data," she tells us. Her role To facilitate the extraction of patient cohorts and gather together needs in terms of clinical data.
Provide a quality service around clinical data Laetitia Chanas' team works on several Institut Curie databases, in particular the Breast base. This base contains no fewer than 70,000 patients treated for breast cancer since 1959. This represents years of follow-up and innumerable clinical data from patient records. Until 2012 this data was up-to-date, but since then the volume has grown to such an extent that data is updated according to projects. This is the mainstay of the data managers' work, namely updating the base manually, according to requests, and checking that data is consistent. In 2018, 2,110 patients were added to or updated in the base. In concrete terms, data mangers use medical records to gather useful data. "But when we have a cohort of 600 patients to update, this can take months", explains Laetitia Chanas. Hence the advantage of working with the team led by Julien Guérin in the Data Department, to use their BIOMEDICS base, which contains data from the hospital group databases and the ConSoRe software. It is a search engine dedicated to big data, and lets the user explore and translate in real time all medical reports and therapeutic protocols administered to patients. "It is able to structure medical data, in other words categorize it into groups." So we can obtain a list of patients with a specific criterion in order to create a cohort. These tools facilitate the work of the data managers in updating the databases. However, manual validation of the data remains essential and even useful for improving the software learning system. And this data specialist goes on to explain, "structuring the information is complicated since each physician has their own way of speaking and the interpretation of the data varies according to the context. I suspect that manual validation will be needed for a long time to come." In short, "the idea is to collaborate to create a virtuous circle; automatic structuring helps data managers in their manual work. And furthermore, we try to improve automatic structuring through feedback from data managers." The ultimate aim is to update and structure as much patient data as possible, and in as much detail as possible.
Although it would seem difficult to eliminate human intervention, another alternative could be to structure medical reports at source, as Laetitia Chanas suggests. This is part of a project coordinated by Sandra Malak.
It would be up to the health professionals to do the structuring, since they are in the best position to do so. Now we need to know how to structure the information in an appropriate way and ensure that this change in practice can be implemented, but it's an interesting lead.
Obtain compliance of collection and data extraction procedures Because when we use patient data we have to follow rules, Laetitia Chanas works with the DPO (Data Protection Officer) and the Legal department. The aims are to implement compliance procedures (MR004 from the CNIL, GDPR) and ensure security of data. She takes part in the CRI Data, a committee recently created to oversee the use of patient data already collected in the care context for the purposes of research. Regulations are indeed different from those governing clinical trials, where data is collected directly within the context of research.
We help project leaders to save time and complete these essential procedures.