Supporting France Universités’ reflections on the state of biomedical research in France

The report presents the findings of an independent study commissioned by France Universités to SIRIS Academic to analyse the current state of biomedical research in France, using a bibliometric in-depth analysis.

The full report is available here. What follows is an executive summary presenting the main analyses and conclusions.

he state of biomedical research in France has been debated repeatedly over the last decade. It has been examined, its difficulties exposed and recommendations for change advanced, most recently in the context of the preparation of the law for multiannual programmation of research (LPR) in 2019 .¹ 

In the wake of the COVID-19 pandemic that exposed further weaknesses in the system, the performance and organisation of the French biomedical research have been placed under even higher scrutiny, pushing discussions further. As a result, in 2021 the Academie of Medicine and Pharmacy published a two-volume report² summing up the state of the debate along two main lines: 1) there is indeed a decline of French biomedical research; 2) this decline is mainly attributable to i) the lower share of research funding to biomedical research in France with respect to international peers; ii) the fragmentation of structures, fundings and actors in the research system; iii) a lack of alignment between Universities and University Hospitals (CHUs). 

This fragmentation and multitude of structures and funding, as well as the misalignment between Universities and CHUs , is perceived as having a strong³ negative impact on the ability to do translational research and engage in multidisciplinary and interdisciplinary research, with consequences on advanced patient care and therapeutic innovation. Although this perception is strongly supported by experts from the sector, there was a lack of evidence that could substantiate it, namely due to the difficulty in addressing quantitatively and robustly issues such as interdisciplinarity, as well the separate contributions of the different actors. 

In light of this reality, France Universités commissioned an independent study aimed at bringing additional arguments and evidence to this debate, based on a bibliometric analysis of publications. The study, conducted during 2022, shed light on two main aspects: 

1. Through original bibliometric analyses, it refined the view of the production, scientific impact, thematic specialisation and interdisciplinarity of French biomedical research in an international comparison. 
2. It provided an understanding of the respective weight of the main types of institutional actors in the French ecosystem, with a special focus on the Universities and CHUs.

This executive summary report provides an overview of the most important conclusions. It is intended to further inform ongoing discussions among all relevant stakeholders over the current state of biomedical research in France and possible future actions to take at national level, namely on a reflection on the role and expected contribution of the universities and CHUs as well as more broadly on the setting of a biomedical research national strategy. The study assesses the validity of some of the existing assumptions and perceptions and provides additional evidence-based insights on where the issues may be arising.

Main conclusions 

The question of France’s overall performance and positioning in biomedical research has been discussed and analysed multiple times, most recently by Terra Nova and Alain Fischer , whose conclusions are in line with previous reports. The present study⁴ aims at refining and completing this discussion, by providing precise data and analysis to validate or inform key assumptions and hypotheses. This section provides an overview of the main conclusions and identifies discussion points that arise from them. 

Both the proportion and the growth of biomedical research in France are lower than in other comparable countries. If France does not prioritise biomedical research by increasing investment to the levels of competing countries, then it is crucial to rethink French biomedical research policy by better aligning decisions and priorities with those taken at a European level by other EU27 countries. 

• Compared with the UK, the Netherlands, Sweden, Germany, Italy, Spain and Canada, overall, measures of biomedical research production in France fare poorly. France has the lowest growth rate from 2010 to 2020, and is the only country below EU27+UK average of growth for the most part of the last decade. It also has the lowest overall production per capita.
• Although growth in overall production is not the only relevant measure of the health of research in a given country, this suggests that biomedical research in France has plateaued below its potential. 
• Most European countries, including France and most of the benchmarks analysed, have not increased their contribution to the world’s share of biomedical research in the last decade, which has remained largely stable. This is likely the result of the emergence and expansion of very large contributors to this field (and arguably not only to this field) such as China, India and Brazil. However, France’s share of the world’s publications is below countries such as Germany and the UK, not only in biomedical research but also in other types of research. 
• Compared to other countries, research in France is proportionally more concentrated in Physical Sciences, with lower dedication to Health and Life Sciences as well as very low dedication to Social Sciences. This is linked to long-term science policy choices and investment in other fields, specifically in the Physical Sciences and Mathematics.

Since research thematic specialisation is obviously a zero-sum game (France is less specialised in biomedical research with respect to the UK, and the UK is less specialised in physics than France), this brings up the question of whether this choice is a good long-term strategy and what the consequences are for the country. More widely, it underlines the fact that it probably makes sense to consider Europe as a whole and explore intra-European complementary areas of specialisation. 

French research is concentrated in classical fields rather than in cutting-edge areas. This increases the risk that French biomedical research will fall further behind competing countries, especially in emerging fields with strong input from the social sciences such as public health, and applied sciences such as bioengineering. Furthermore, French research appears to be more uneven than elsewhere in terms of its scientific impact. 

It is useful to explore the topics and areas of specialisation and multidisciplinarity within a given country’s portfolio , since it shows where it may have a competitive edge⁵ at a European or global level and where it stands to contribute and impact the most. In this regard:

• One of the main issues for biomedical research in France appears to be the specialisation in classical fields rather than in cutting edge or emergent fields. This is true of French research in the Physical Sciences, Social Sciences and Health and Life Sciences, but, arguably, is felt most strongly in the Health and Life Sciences where new diagnostic tools and treatments, such as vaccines, require more advanced, innovative approaches.
• Biomedical research in France is dedicated to traditional areas such as Cancer, Cardiovascular Diseases, Infectious Diseases, Neuroscience, Immunology, Surgery and Genetics, in a pattern mostly shared amongst the benchmark countries. 
• The areas of Cancer, Infectious diseases and Immunology are those where France is more specialised in comparison with benchmarks; this is not surprising considering the existence and success of specialised and/or research intensive centres on these topics (e.g. Institut Gustave Roussy, Institut Curie or Institut Pasteur). In line with this, France also shows significant dedication and high specialisation in haematology and virology. 
• However, France is clearly not specialised in other fields, such as those of Public Health and Policy, and Psychiatry and Mental Health. These are fields that have garnered increased attention in the last years and in which countries such as the UK and the Netherlands (and Spain in the case of Public Health and Policy) have higher contributions. 
• France’s biomedical research shows a degree of multidisciplinarity similar to⁶ he benchmark countries. 
• Overall, the level of interdisciplinarity is similar between France and⁷ benchmarks, albeit with different patterns. Interdisciplinarity in Biomedical research in France has contributions mainly from the Physical Sciences, and, in contrast to the Netherlands or the UK, interdisciplinary research with the Social Sciences is low. 
• Interestingly, and most likely linked to the known strengths in Physical Sciences, interdisciplinary biomedical research in France is more specialised in Environmental Sciences and Physics, rather than in biomedical applied fields such as Biomedical Engineering, Bioengineering or Biomaterials.
• It has been argued that weaknesses of the system, such as a lack of funding and evaluation structures specific for translational research, have a negative impact on the ability to do this type of research. The analysis shows that the share of basic and translational research in France is similar to other countries such as Spain and Germany, and its performance (at least measured through the ability to publish in Nature Index journals) is on par to countries such as the Netherlands, and only slightly lower than the UK .⁸ 
• Despite lower production, France’s biomedical research has a similar scientific impact to countries such as Spain and the UK, with a similar percentage of its publications in top tier journals (all around 50%)⁹,Top 1-10% Scimago journals ).¹⁰
• However, France also has a proportionally higher number of publications in lowest tier journals (30-100% Scimago journals), compared to benchmarks.

The organisational complexity of the French biomedical research system appears to have a negative impact on the capacity at steering research policy and addressing major national priorities. This seems, at least in part, linked to a difficulty in aligning CHU research priorities with that of universities and a broader issue in terms of primary affiliation identity of joint CHU-University academic staff, as well as the scientific impact of the research performed at the CHUs.

The dichotomy in the impact of research just mentioned could be the result of the¹¹ diversity of practices and types of research being developed in Universities and CHUs, and across the different sites. Several factors can influence this: the different practices normally seen in fundamental vs clinical research, and specifically within clinical research, factors such as the type of clinical research taking place (e.g clinical case studies) or the prioritisation of excellent research by institutions. The higher proportion of publications in lower tiers could be indicative of a low or/and uneven focus on (excellent) research. 

Biomedical research that originates from both clinical and fundamental research functions in a continuum, requires coordination and exchange between all the institutions/groups that perform these types of research, which is naturally complex. In France, this relies on a triad of actors: Universities - CHUs - Research organisations (such as Inserm). 

One of the main intuitions of the actors is that France’s under-performance is largely due to organisational complexity, which dilutes efforts and diminishes efficiency. There is also the idea that the CHU model, once quite successful, is in need of revision to allow higher quality research, namely translational clinical research. 

By analysing the institutions behind French biomedical research publications, with a special focus on the Universities-CHUs relation, this study attempted to see whether evidence supports these perceptions: 

• By identifying the research produced by the different types of institutions through the affiliations that are indicated by each last author on a publication (research group or leader of the research) we confirmed a clear problem of feeling of belonging to the universities which appear through the affiliation practices of CHU authors. 
• Nearly all authors mentioning a CHU affiliation should also indicate a university affiliation, since most are employed by a university. However, this is far from the case, as we found that a considerable percentage of all publications of CHU´s last authors did not mention the university affiliation. To be able to have an approximation of reality, we have “reconstructed” the Higher Education Institutions (HEI) affiliations to include the CHUs, and considered this as well as research conducted exclusively at universities (hereinafter called Universities) and research conducted at CHUs regardless of University relation being made explicit or not, (hereinafter called CHUs) as our predominant categories in the analyses. 
• Higher education institutions (and associated CHUs) contribute to more than 60% of all biomedical research in France (33% and 29%, in Universities and CHUs respectively) 
• The contribution of most institutions by research type is as expected, such as with CNRS publishing substantially more in basic and translational research; CHUs and Hospitals in public health, epidemiology and clinical research and university groups without CHU affiliation publishing more in basic and translational research. Cancer centres dedicate most (around ⅔) of their research to public health, epidemiology and clinical research. 
• Citation behaviours are distinct for basic & translational research vs the clinical research domain, in which the latter generally has a dichotomy of highly cited studies (e.g. big clinical collaborative studies on major diseases) and very low cited research (e.g. clinical case studies or regional public health studies); while research categorised as basic research tends to have a more uniform type of studies. This pattern is reflected in the citation metrics we see for Fundamental and Clinical research in Universities and CHUs. 
• CHUs do not fare well in the scientific impact of their publications, and especially so in basic and translational research, with citation metrics below Universities and Inserm in both research types. The low citation metrics in basic and translational research suggests that there is not a focus on excellent translational research in the CHUs, even if the dedication to this research type is expected to be low 
• Despite an overall lower scientific impact in both fundamental and clinical research, in research on major disease groups, CHUs actually have a higher share of their publications in the Top 1-10%, when compared to Universities. 
• CHUs and Inserm display the highest proportion of publications in Top 2-10% Scimago in Cardiovascular Diseases and Infections; while Inserm publications have the highest scientific impact in the fields of Neoplasms and Nervous System Diseases. 
• In fields of higher research dedication (larger fields in volume of publications), nearly all research in Life Sciences is published by Universities and nearly all research in Health Sciences by CHUs. There is however an interesting exception: Public Health is in fact a specialisation of Universities. 
• In fields of higher research dedication, Universities are more prone to interdisciplinarity. 
• There are no substantial differences between Universities, Inserm and CHUs in the dedication pattern of research to the main areas of high production and specialisation in France: Neoplasms, Infections, Infectious Diseases, Cardiovascular Diseases and Hemic and Lymphatic diseases. Although, and as expected, CHUs have a proportionally higher volume of research per disease group than other organisations. It is interesting to note that although producing more in basic and translational research, university groups have a reasonable percentage of research in diseases; likely in fields such as molecular basis of disease or disease models. 
• The national specialisation in infectious diseases and cancer is distributed through all institution types (CHUs, Universities, Inserm and others), with all having a reasonable contribution in publications in Top 1-10%. 
• CHUs and Hospitals display the lowest share of publications with international partners (around 45%, while all other actors show >50% share). 
• Of all types of institution analysed, CHUs, together with HE Institutions, have the highest share of publications led as first or last author. An increased capacity to lead, although in itself good, is also often a trade-off of reduced collaborative work, which may decrease international visibility, peer recognition and research excellence .¹² 
• Inserm, CNRS and IHUs show a fairly balanced distribution of leadership and internationalisation, showcasing the capacity to collaborate internationally, without losing the capacity to lead. 
• The patterns of internationalisation and leadership do not change substantially when just basic & translational or public health, epidemiology & clinical research are considered.

Overall, this study tested one main assumption and raised one additional main question (that can be subdivided in two more specific questions): 

1. Is there a decline of biomedical research in France? 
2. Where is research produced and what is the weight of the different actors? 3) And, is the current CHU model allowing high quality research, namely translational clinical research?

Three broad conclusions can be reached: 

1. Biomedical research is not declining per se (although it is declining relative to other countries here analysed), but its growth is plateauing below its potential and has been doing so in the last decade, in contrast to other European countries. Considering the intense competition from large emerging countries, this is a major cause for concern. Moreover, and beyond specialisation in large areas such as cancer and infectious diseases, France appears to be specialised in classical fields rather than in cutting edge or emergent fields. 
   
   It would be important to ask what are the biomedical research priorities at the national level, and is, overall, biomedical research a priority? And if not, what would be the effect for the country? ‍
2. More than 60% of the research produced in France originates from the combined Universities and CHUs. However the pattern of research that is being performed is distinct; with Universities carrying nearly all specialisations in Life Sciences and CHUs nearly all the research in Health Sciences, with little overlap. Universities are also more interdisciplinary.
3. CHUs are also not faring well in terms of scientific impact of publications; with the lowest rate of citations when compared to Universities and Inserm in both basic & translational but also, importantly, in public health, epidemiology and clinical research. Although this could indicate that different types of studies carrying different citation behaviours are occurring in CHUs vs Inserm/University groups, it nonetheless indicates a lower impact and/or lower interest from the global community. It also suggests that there may not exist a focus on excellent translational research in the CHUs. 
   
   Although CHUs have a lower research impact than what would/could be expected or desirable, further comparative studies with international institutions with a similar model to the CHUs would allow a deeper understanding of how the CHUs, and integration with the Universities to which they are associated, are truly faring.

‍

‍