Funding concentration leads to knowledge gaps in Danish biomedical Research

Presentation at the "Funded and Unfunded Science: Academic Inequalities and Epistemic Gaps" conference, Oct. 4-5, 2021, Prague. Presented by Jens Peter Andersen, authored by Jens Peter Andersen & Emil Bargmann Madsen.

Transcript

1. 5 OCTOBER 2021 SENIOR RESEARCHER FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY FUNDING CONCENTRATION LEADS TO KNOWLEDGE GAPS IN DANISH BIOMEDICAL RESEARCH Jens Peter Andersen & Emil Bargmann Madsen
2. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY BACKGROUND • Ràfols & Yegros-Yegros (2017): comparing research outputs (focus on Spain) to health needs (disease burden). Calls for better planning. • Yegros-Yegros et al. (2020): Using complete Medline • Show research output concentrated in handful of diseases most prevalent in high- income countries • Publicly funded research efforts focused in similar diseases as industrial R&D and often on diseases attracting more citations • Own report on the funding landscape in Denmark gives first impression of distribution of funding in a Danish context.
3. PURPOSE To evaluate the concentration of funding, specifically on diseases, contrasted to the societal health impact of diseases.
4. DATA Combining publications, scientists, funding and disease burden.
5. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY OVERVIEW We use a data set linking funding given to individuals, from Danish private and public funders, to their publications in Web of Science and PubMed Medline, and the Global Burden of Disease (GBD) data on Disability Adjusted Life-Years (DALY) for Denmark.
6. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY FUNDING Funding organization Number of grants Public research funders Independent Research Fund Denmark 8,507 Danish Council for Strategic Research 582 Danish Council for Technology and Innovation 94 Danish Advanced Technology Foundation 211 Danish Ministry of Higher Education and Research 46 Danish National Research Foundation 154 Private research funders Carlsberg Foundation 2,432 Danish Cancer Society 1,557 Lundbeck Foundation 2,090 Nordea Foundation 7 Novo Nordisk Foundation 2,516 TrygFonden 517 Villum Foundation 308 Velux Foundation 264 International research funders European Research Council 137 • 15 reseach funding organizations • 19,422 research grants awarded between 2004 and 2016 • 7,539 PIs • ~ 52.9 bn DKK or €7.1 bn
7. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY AUTHORS • Authors are identified by their names and publications. From the funders we have the names (manually cleaned), and we manually attach 1-2 publications from Web of Science. • Using author name and publications, we find the most exact possible author profile (exhaustive publication list), using the disambiguation algorithm of Caron & Van Eck (2014). Only authors with a profile match are included. • Profile matches were manually checked for correctness.
8. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY PUBLICATIONS • Web of Science contains PubMed ID’s (PMID) on all records in PubMed. We use this to match between databases. • From PubMed we extract the MeSH subject headings for each publication. • Using the Unified Medical Language System (UMLS) thesaurus, we link MeSH to ICD-10 codes (International Classification of Diseases). • Using the translation table of the Institute for Health Metrics and Evaluation (IHME), we map ICD-10 codes to causes in the IHME Global Burden of Disease (GBD) data.
9. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY DISABILITY-ADJUSTED LIFE YEARS (DALY) • From the IHME GBD, we extract DALY’s for all causes in Denmark, for the period 2004- 2016. • DALY consists of two elements: • Life years lost to death (life expectancy – age at death) • Years lived with disability, weighted (e.g. a fractured femur is weighted .208, Alzheimer’s .666)
10. People CWTS Web of Science IHME Global Burden of Disease UMLS PubMed Funders Grant receives Person has Author has Web of Science record PMID PubMed record translates to ICD-10 Disease translates to Causes aggregates Publication represented by MeSH Funding amount person_id* author_id* DALY Rate
11. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY LIMITATIONS OF DATA Not possible to directly link grants to publications in the Danish data. We therefore have to use a proxy, linking all publications of an author to all their grants. This is OK, because: 1. Funding is given to individuals, not to articles. 2. Nearly all authors have no or few focus changes over time. Several authors deal with more than one cause, but do so consistently.
12. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY DEFINITIONS For a set of authors, 𝐴, author 𝐴𝑖 has a set of grants 𝐺𝑖 and publications 𝑃𝑖. The funding for 𝐴𝑖 is 𝐹 𝐴𝑖 = σ 𝑗 𝐺𝑖 𝐹(𝐺𝑖𝑗 ), where 𝐹(𝐺𝑖𝑗 ) is the funding for grant 𝑗 of author 𝐴𝑖’s grants. To calculate the funding a given publiaction receives, we first calculate an author weight, 𝑤𝑖𝑗, which is 𝑤𝑖𝑗 = ൗ 1 𝐴∋𝑃𝑖𝑗 , with 𝑤𝑖 = σ 𝑗 𝑃𝑖 𝑤𝑖𝑗. The publication funding, 𝐹(𝑃𝑖𝑗 ), is then: 𝐹 𝑃𝑖𝑗 = 𝑤𝑖𝑗𝐹(𝐴𝑖) 𝑤𝑖 As a publication can address multiple causes, 𝐶, the funding a particular publication provides to a cause is: 𝐹 𝐶𝑖𝑗 = 𝐹(𝑃𝑖𝑗) 𝐶𝑖𝑗 , and the total funding to a given cause in a given year is the sum of 𝐹(𝐶𝑖𝑗 ) of the publications published in that year.
13. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY DEFINITION CONSEQUENCES • The definitions for funding per publication and cause allow us to have a proxy for how well-funded a given disease area is. The causal link is not exact. • There is an additional limitation introduced from the time element: there is no clear logic in when a grant is given, a publication is published and when there are changes in health. Here we use a four-year delay between publication year and DALY. • We can only work at the aggregate level, and our conclusions must be careful!
14. RESULTS
15. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY DESCRIPTIVE OVERVIEW
16. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY VARIABLE RELATIONSHIPS
17. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY DALY RATE CHANGE BY FUNDING -1
18. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY DALY RATE CHANGE BY FUNDING -2
19. DISCUSSION
20. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY OVERALL FINDINGS • No straightforward relationship between burden or change in burden and funding intensity • Slight trend towards more funding for more pressing issues, but: • Global disease burden also plays a role, e.g. tropical medicine (e.g. malaria) • Some causes receive disproportional funding, e.g.: • Cancer: public opinion, government policy (24 hour treatment initiation) • Diabetes/metabolism: largest, private funder • Underfunded topics, e.g.: • Alcohol misuse • Liver diseases
21. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY CALL FOR TRANSPARENCY Strategic investment and responsible research for society requires transparent information on what research is funded. Trasparent data would require: • Who is funded, by how much, for what (incl. Keywords (intention) and publications (outcomes))? • What is funding spent on? (Salary, equipment, other) • Involved scientists Allowing further research on concetrations and dispersion, societal focus, training, societal impact
22. FUNDED AND UNFUNDED SCIENCE JENS PETER ANDERSEN 5 OCTOBER 2021 SENIOR RESEARCHER DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY THE END… Slides can be found here: All artwork produced by Danielle Navarro, licensed CC-0 https://art.djnavarro.net/ Jens Peter Andersen on Twitter: @ipoga Emil Bargmann Madsen on Twitter: @EmilBargmann
23. DEPARTMENT OF POLITICAL SCIENCE AARHUS UNIVERSITY DANISH CENTRE FOR STUDIES IN RESEARCH AND RESEARCH POLICY