1. Addressing Funding and Conflicts of Interest in Randomised Clinical Trials included in Cochrane Reviews Plans for the development of a ‘tool’ to assess reasons for concern
Asbjørn Hróbjartsson

2. Disposition
Background
Status
Challenges
Suggested approach

3. Jonathan Sterne vs. Lisa Bero debate
The Cochrane risk of bias tool should not include funding source as a standard item (Cochrane Library 2013):
“In summary, conflict of interest in reporting of medical research is a substantial and unresolved problem. …There are particular problems associated with pharmaceutical industry-funded research, but these should be dealt with by:
reporting and commenting on conflicts of interest as a standard component of Cochrane systematic reviews;
better procedures and an improved tool to assess reporting biases; and
more extensive use of mixed treatment comparisons.”

4. Jonathan Sterne vs. Lisa Bero debate
The Cochrane risk of bias tool should include funding source as a standard item because (Cochrane Library 2013):
“1. Funding source fits the definition of bias
2. There is empirically-based evidence of bias related to funding source
3. The observed bias related to funding source cannot be captured by the risk of bias criteria currently assessed with the risk of bias tool … .”

5. Broad consensus
Funding and Conflicts of interest may impact on risk of bias
Funding and conflicts of interest may impact on quality of the evidence
Funding and conflict of interest may be a basis for sensitivity analysis

6. Risk of bias tool Six bias domains Domains are mechanistically defined
Generation of allocation sequence
Concealment of allocation sequence
Blinding of patient and care provider
Blinding of outcome assessor
Incomplete outcome data
Selective reporting
Domains are mechanistically defined
Risk of bias tool builds on the idea of ‘core’ biases
Funding and conflict of interest is not regarded a core bias

7. Funding and CoI
Association between industry funding and trial conclusion (Lundh CDSR 2012)
Few direct comparisons
Strong incentives for positive trial results
Trial 329 (Le Noury BMJ 2015)
One failed stage III trial on idalopirdine for Alzheimer dropped the prize of Lundbeck’s stock by 15% (Bloomberg 2016)
Examples of selective reporting of industry initiated trials and outcomes
SSRI for adolescents and children (Whittington Lancet 2004)

8. Status in Cochrane reviews
Cochrane Handbook: Funding and conflict of interest should not be addressed as a risk of bias domain.
The Handbook provides no clear approach as to how funding and conflicts of interests should be addressed.
32% of Cochrane reviews published in 2014 incorporated funding into the “other bias” function (5% added funding as a separate domain).
Jørgensen et al. Syst Rev May 10;5:80.

9. Status risk of bias tool
Several changes
The ”other domain” option gone
No domain for ‘funding’
Plans for tool to address funding and conflicts of interest
Plans underway to develop a specific tool for reporting bias

10. Why is a tool important?
Considerable discrepancy between the Cochrane risk of bias tool approach vs. actual praxis
Concern of the risk that assessment of funding and conflict of interest is conducted inconsistently and crudely
Concern of the risk that assessment of funding and conflict of interest is not conducted, and thus does not impact on risk of bias assessments and on assessment of the quality of evidence

11. The aim of the tool
1) To systematically identify and summarise relevant information on funding and conflicts of interests 2) To coherently assesses reasons for concern for funding and conflicts of interests (assesses whether or not funding and conflict of interest may impact importantly on trial results and quality of evidence) 3) To incorporate a user-friendly interface between assessments of concern and: a) risk of bias, b) quality of evidence (e.g. risk of publication bias and indirectness) 4) To provide background for sensitivity/subgroup analyses

12. Framework Core group Wider group Set-up and time line
Isabelle Boutron, Lesley Stewart, Ian Shrier, Andreas Lundh, Asbjørn Hróbjartsson
Wider group
Lisa Bero, Peter Gøtzsche, Tom Jefferson, Kay Dickersin, Julian Higgins, Jonathan Sterne, Brett Thombs and An-Wen Chan.
Set-up and time line
Andreas Lund will work on the project on a day to day basis.
We hope to have a tool ready for pilot testing in 6 months.

13. Summary
Funding and conflict of interest are core aspects of a randomised trial
The present approach suggested for Cochrane Reviews is incomplete and challenging
We propose developing a ‘tool’ which systematically addresses funding and conflict of interest