5 Post-Doctoral Researchers Leading Innovative Projects
The research team Epilogy, Epidemiology and Evaluation of Therapeutics in Dermatology and Immune-Mediated Inflammatory Diseases has hosted and supported several post-doctoral researchers working on high-level projects in epidemiology, therapeutic evaluation, and advanced methodological research.
These post-doctoral projects focus on real-world data analysis, comparative effectiveness research, causal inference, drug safety, and methodological innovation in dermatological diseases and immune-mediated inflammatory diseases.
The research carried out by post-doctoral researchers within the team contributes to strengthening scientific expertise, developing innovative analytical methods, and producing robust evidence to improve the understanding of diseases, assess treatments, and support better patient care and public health decision-making.
Infectious Risk of New Immunomodulatory Therapies in Immune-Mediated Inflammatory Diseases: A National Cohort Using Data from The French National Health Data System
Post-doctoral researcher
PhD in Public Health, Nurse
Supervisor: Emilie Sbidian - Funding: ANSM
Research project
Immune-mediated inflammatory diseases (IMIDs) are a heterogeneous group of chronic conditions that share common pathways. They include psoriasis, psoriatic arthritis, spondyloarthritis, Crohn’s disease, ulcerative colitis and rheumatoid arthritis. Biologic and targeted synthetic therapies have improved disease control and quality of life for many patients. However, by modulating the immune system, they can also increase the risk of serious infections. Identifying which patients are most at risk is important for safer treatment decisions.
This thesis uses the French National Health Data System (SNDS) to study large real-world cohorts of IMID patients starting biologic or targeted synthetic DMARDs from 2010 to 2023.
The objectives are to :
- Construct a nationwide cohort of IMID patients starting biologic or targeted synthetic therapies, and to describe their demographics, comorbidities, and trends in therapy use over time.
- Develop and validate a model for predicting serious infection risk in rheumatoid arthritis.
- Assess the risk of herpes zoster in IMIDs, identify patients at greatest risk, and inform targeted prevention strategies.
Safety and Efficacy of Nintedanib in Systemic sclerosis in REAl Life (SENS-REAL) study
Post-doctoral researcher
Associate professor at the University of Aix-Marseille - Internist at La Timone hospital (AP-HM)
Supervisor: Emilie Sbidian - Funding: CHU La Timone
Research project
Systemic sclerosis (SSc) is a rare systemic autoimmune disease. Its complex pathophysiology is based on dysfunctions involving endothelial cells, fibroblasts, and the immune system. The primum movens appears to be an alteration of the microcirculation, preceding post-ischemic fibrosis. This condition can affect vital organs such as the heart, lungs, kidneys, and gastrointestinal tract. One of the most severe complications, and one that has the greatest impact on patients’ prognosis, is systemic sclerosis-associated interstitial lung disease (SSc-ILD).
Nintedanib, an antifibrotic agent initially developed for idiopathic pulmonary fibrosis, was approved in 2020 for SSc-ILD based on the results of two clinical trials: INBUILD and SENSCIS. However, the primary endpoints of these studies were mostly surrogate endpoints based on pulmonary function parameters (forced vital capacity, DLCO, etc.), with a relatively short follow-up (52 weeks) and no assessment of survival impact. Moreover, clinical trials typically include highly selected patient populations (median ages, fewer comorbidities, often less severe disease than in the general population), which limits the generalizability of the results. In addition, real-world evidence on the efficacy and safety of nintedanib in this indication remains limited.
To address this gap, this study aims to leverage real-world clinical practice data. We will use the extensive data available from the French Système National des Données de Santé (SNDS), which covers the entire French population and provides an exhaustive, representative cohort for the use of nintedanib in SSc-ILD.
The objectives are to :
- Assess the impact of nintedanib vs nintedanib + mycophenolate mofetil on overall survival.
- Evaluate the impact of nintedanib vs nintedanib + mycophenolate mofetil on the risk of hospitalization for respiratory exacerbation, hospitalization for any cause, and initiation of long-term home oxygen therapy.
- Compare nintedanib vs nintedanib + mycophenolate mofetil patients in order to identify factors associated with nintedanib prescription in real-world settings.
- Assess treatment safety by analyzing discontinuation rates and hospitalizations for serious adverse events (severe diarrhea, deterioration of general condition).
Causal Methods for Quantitative Synthesis of Mediation Analyses
Post-doctoral researcher
Supervisor: Thang Vo - Funding: Chaire Pr Junior
Research project
Mediation analysis is a common type of statistical analysis in psychology, sociology, epidemiology, and medicine. Such analysis aims at assessing the relative magnitude of different pathways and mechanisms by which a treatment or an exposure may affect an outcome. Systematic reviews and meta-analyses of mediation studies are increasingly being implemented in practice. Nonetheless, the methodology for conducting such review and analysis is still in a development phase, with much room for improvement.
In this project, we develop novel statistical methods to meta-analyze results of different mediation analyses, taking into account:
- The difference in the target population of these studies.
- Partial information about the treatment mechanism obtained from studies that assess the treatment-mediator or mediator-outcome relationship.
- Restriction against individual-level data access in one or multiple individual studies.
New methods will be evaluated by numerically simulated data, and illustrated by real-world clinical data.
Implementation of novel network meta-analysis methods to assess the safety of systemic treatments: an application on the evaluation of the safety of current systemic treatments for moderate-to-severe psoriasis in adults
Post-doctoral researcher
Supervisor: Laurence Le Cleach - Funding: CCA Bettencourt
Research project
Network meta-analysis (NMA) is an important statistical method in comparative effectiveness research that allows the comparison of several treatments in a single analysis, providing an estimation of the relative efficacy and safety, as well as a ranking of every included interventions. However, NMAs rely for the vast majority on evidence from aggregated data of randomized controlled trials (RCTs). Hence, NMAs inevitably inherit from some of their drawbacks and limitations, particularly in the case of safety assessments (highly selected population, short-term assessment, rare serious adverse events with a low number of exposed participants). Because NMAs’ results are important for clinical decision makers and guidelines developers, it is crucial that they provide precise and robust data relative to treatments efficacy and safety. Recently, new approaches have been developed to overcome these obstacles, such as the combination of both randomised and non-randomised studies, and the use of individual participants data (IPD) rather than aggregated data (AgD).
In this project, we will use the Cochrane review on psoriasis to apply these novel NMA approaches and meta-research methods to assess the safety of current systemic treatments, and the robustness of our findings.
Our objective is to provide data that accurately describe the relative safety of current systemic treatments for moderate-to-severe psoriasis in adults. To do that we will :
- Perform a network meta-analysis that includes both randomised and non-randomised studies.
- Compare the RCTs’ individual participants data to the published aggregated data, explore the reasons for discrepancy and assess the impact on the risk of bias judgment.
- Evaluate the robustness and the reproducibility of the RCTs’ findings and its impact on the NMA’s results by re-analysing the IPD under various methodological scenarios.
Causally Interpretable Network Meta-Analysis (CI-NMA): fit-for-purpose methods for decision-making
Post-doctoral researcher
Supervisor: Thang Vo - Funding: Novo Nordisk
Research project
Network meta-analysis (NMA) has become a cornerstone methodology for the assessment of innovative health technologies, in the absence of direct head-to-head comparisons between the interventions of interest. It provides valuable information to prescribing physicians, regulatory agencies and payers on the relative efficacy and safety of drugs and has a crucial impact on market and patient access.
This project will develop fit-for-purpose NMA methods that are flexible, bias-robust and produce causally interpretable results in specific target populations. Such methods are increasingly attractive for pharmaceutical companies, regulators and reimbursement agencies worldwide. Policy decisions are made for specific healthcare settings and require treatment effect estimates that are maximally relevant to the target population for decision-making.
Within the context of health technology assessment (HTA), the landscape is being disrupted by the new European Union HTA Regulation, which demands:
- A dramatic increase in the use of NMA due to unavailable head-to-head comparisons between all competitors.
- Considerable analytical complexity with respect to the type of NMAs being conducted.
- The generation of comparative effectiveness results in many different member state populations.
The aim of CI-NMA is to develop novel methods for case-mix standardization that are bias-robust and allow for causally interpretable network meta-analysis, in the context of both full access and restricted access to individual participant data (IPD). CI-NMA includes three work packages (WPs).
In WP1, we will develop robust and powerful approaches for case-mix standardization under limited access to IPD, which enable the use of machine learning methods in the estimation process, reducing the dependence on modeling assumptions and potential for bias while maintaining valid inference.
In WP2, we will develop novel methods for CI-NMA under full access to IPD aiming to: (1) compare and rank different treatment options for specific target populations; and (2) quantify the importance of case-mix heterogeneity in the trial network.
Finally, in WP3, we propose a new strategy to include aggregate data from trials without IPD into CI-NMA, integrating the methods developed in WP1 and WP2.