Project description

We are recruiting a postdoctoral researcher to work on a medical-scientific project funded by a European Research Council (ERC) Starting Grant in the field of inflammatory bowel diseases (IBD). The FORECAST Project is based on linking three real-world data sources: a national health insurance claims database (the French National Health Data System, SNDS), a database derived from electronic health records (the data warehouse of the Assistance Publique–Hôpitaux de Paris, AP-HP EDS), and SUVIMIC, a prospective cohort of IBD patients followed in AP-HP centers that are part of the Paris IBD University Centers (Paris IBD-U) group.

The SUVIMIC cohort is a prospective cohort of over 10,000 IBD patients followed at AP-HP, with structured data on their therapeutic management.

The AP-HP EDS is a database that hosts electronic medical records of millions of patients treated at the 38 hospitals of AP-HP. These data can be structured (diagnoses, treatments, procedures, hospital laboratory results, demographics, etc.) or unstructured (hospital discharge summaries, prescriptions, consultation notes, etc.). More than 35,000 IBD patients have been identified in the AP-HP EDS.

The SNDS compiles healthcare consumption data for the entire French population and includes more than 250,000 IBD patients.

This project is being conducted as part of a collaboration between the FORECAST team at the Pierre Louis Institute of Epidemiology and Public Health, the Pharmacoepidemiology Center of AP-HP, and the team from the Sorbonne Center for Artificial Intelligence.

Evaluating the benefit-risk balance of treatments is essential in medicine but poses significant challenges, particularly in the context of IBD. Although new therapies are available, their optimal positioning remains uncertain due to the lack of direct comparisons between treatments. Recent emulated trials have highlighted the value of real-world data in assessing IBD therapies. The ERC-funded FORECAST project aims to validate algorithms that identify patient characteristics and disease trajectories using administrative health databases. It will emulate clinical trials comparing advanced therapies for IBD, integrate the results with patient preferences into a clinical decision support system, and evaluate this system through a pragmatic randomized controlled trial. The project will also establish post-marketing drug surveillance methods applicable to future IBD therapies and other chronic diseases.

The FORECAST project aims to pave the way for a new framework for post-marketing drug surveillance, applicable to future treatments for IBD as well as other chronic diseases. This is an opportunity to work on an ambitious project in collaboration with a multidisciplinary team of researchers, including experts in clinical medicine, pharmacoepidemiology, and artificial intelligence.

The selected candidate will be involved in all aspects of data science. Our main working languages are R and Python.

 

References

- https://cordis.europa.eu/project/id/101163425

- https://sante.sorbonne-universite.fr/actualites/erc-starting-grant-forecast

 

Main Responsibilities

• Extraction and visualization of structured or textual data.
• Conduct feasibility studies and research projects on IBD.
• Use of Natural Language Processing (NLP) algorithms to extract information from medical reports.
• Use machine learning models to study the efficacy and safety of treatments.
• Participate in drafting statistical analysis plans for pharmacoepidemiological studies.
• Conduct statistical analyses on the project’s databases, including the SNDS: population and variable selection, descriptive statistics, modeling, and graphical representations.
• Assist in disseminating results through scientific papers (co-author) and presentations at national and international conferences.

Desired Profile

• Ph.D. in Mathematics, Computer Science, Biostatistics, Epidemiology, and/or Data Science.
• Expertise in statistics and machine learning.
• Experience in data manipulation.
• Proficiency in R and/or Python, with knowledge of best coding development practices.
• Excellent ability to collaborate across teams and to explain complex concepts to non-experts.

Bonus

• Experience in public health or epidemiology.
• Experience with real-world health data, particularly SNDS.

Important Note: We are hiring people, not positions.
If, after reading this job description, you feel that you may not meet all the requirements but that the role aligns with where you would like to be in your next position, we encourage you to apply!