Post-marketing trials are paramount for enhancing evidence about novel medicines once they are used at a massive scale

We systematically studied all post-marketing trials conduceted at a global scale for a consistent sample of novel medicines approved by the two leading regulators

We found extreme variations in the way medicines were studied after approval.

Development and evaluation of interventions to improve the peer-review process in biomedical research

This doctoral thesis focused on medication adherence among patients with one or several chronic conditions and taking multiple medications.
The first PhD project was the development of an instrument to assess medication adherence, with a “drug by drug” perspective. Validity and reliability of the tool were tested among 243 patients in France. The two other projects focused on the discordance between patients’ and physicians’ points of view about medication importance and on the estimation of the threshold for unacceptable risk of non-adherence according to 544 physicians for the most prescribed drugs in France. The results suggested a huge variation in physicians’ estimations according to the drug of interest.

In this thesis we mapped registered clinical trials at a macro-level, i.e. we analyzed what clinical research is undertaken worldwide, where it is conducted, which diseases are studied, and who is supporting it.
We highlighted major gaps in the effort of health research as compared to health needs, and the influence of different sponsors in the globalization of clinical research.

We found that SSA RCTs investigate locally high burdened diseases, however these trials are frequently led by high income countries.
Moreover, SSA RCTs at high risk of biais could be changed in trials at low risk of bias with simple methodological adjustments of minor cost, and the completeness of the intervention description could be improved with the adherence to CONSORT guidelines.

This thesis assessed the prevalence of spin in Google Health news, classified different strategies of spin and identified that factors related to publication process were associated with high dissemination through online media. This work highlighted the importance of spin in mass media and provided a tool (classification of spin) to improve the dissemination and communication of research results.

We analyzed all the FDA-approved drugs with a pharmacogenetic biomarker in their label and showed that (1) predictive biomarkers in the label of drugs is restricted in large part to oncology; (2) the level of evidence for clinical utility is often unknown because studies are restricted to the subgroup of biomarker-positive patients from the initial stages of the drug development; (3) there are very important variations between drugs in the use of enriched trials and in testing of the drug in several indications or with several predictive biomarker; and finally (4) that only one third of the anti-cancer drugs approved by the FDA between 2000 and 2015 have meaningful clinical benefit.

Phrase: In my PhD thesis, I modeled the macroscopic behavior of the scientific publication and peer-review systems, and I showed that the scientific community can collectively meet the annual demand for peer review, however, only 20% of researchers have been performing 69%-94% of reviews. Moreover, my results indicated that peer-review systems which allow sharing of past reviews of rejected manuscripts between journals showed important advantages, such as improved peer-review efficiency, less reviewer effort and better scientific dissemination.

In this project, I aimed to 1) compare treatment effect estimates from non-randomized studies with PS analysis and RCTs and 2) describe and assess the reporting and potential bias of PS analysis used in a sample of published observational studies assessing surgical procedures.

Systematic reviews are essential tools to synthesize available evidence for therapeutic evaluation. Multiple treatments are now frequently available for a given condition. Patients and physicians want to know which one is the best among all treatments. Thus we need to retrieve and synthesize all available evidence across all treatments and furthermore to maintain it updated when new evidence and new treatments become available.

Our objective was to evaluate the limits of the current ecosystem of evidence synthesis and to develop an alternative methodology.

The objective of this PhD thesis was to develop and validate a methodology with dynamic case-vignette to study discrepancies with good practices and to assess determinants related to the organization of care and / or caregivers in postpartum hemorhage management.

An accurate and complete presentation by authors of all the important information in report of randomized controlled trial (RCT) and systematic reviews (SR) is essential to allow a critical appraisal of the study by readers. Spin was defined as a specific way of reporting, deliberate or not, that can lead to a “beautification” of the data. We showed that 1) spin are disseminated in press releases and news and they are associated with an overestimation of the beneficial effect of treatment, 2) the addition of a limitation section in abstract of systematic reviews and meta-analysis does not impact the confidence in the study results by readers. Finally, we developed a classification scheme of spin in systematic reviews and meta-analyzes, and we ranked spin in abstract according to their perceived severity

As chronic diseases are becoming more prevalent, their management strategies are becoming more complicated. Patients are feeling increasingly burdened by their treatments. This burden of treatment represents the “work of being a patient”. It affects treatment adherence, quality of life and outcomes. In order to take into account this burden, clinicians and researchers needed tools to assess it.

Network meta-analyses are primarily meta-analyses; they are probably subject to the same types of bias and should therefore be performed according to the explicit and rigorous methods used in systematic reviews. We assessed whether network meta-analyses follow the key methodological recommendations for reporting and conduct of systematic reviews, and how the results of network meta-analyses are reported.