Comparing how interests are declared in Cochrane systematic reviews with the procedure employed by the European Food Safety Authority in its 2012 Opinion on Thresholds of Toxicological Concern.
Recommendations on declarations of interest in the conduct of Cochrane reviews in medicine are comprehensive, covering not only financial conflicts of interest but also publishing histories and any other factor which might have influenced somebody’s contribution to a review. The rationale is whether or not the declaration of this information might influence a reader’s interpretation of the findings of the review.
These declarations are very prominent in Cochrane reviews, having their own dedicated section. Interests are clearly connected to the genesis, conduct and reporting of reviews by making explicit the contributions each author has made to the review [see here for a full scheme of how this can be done; a summary of the full Cochrane guidance is given at the end of this post].
One declaration of interest which provides a simple example of how this conflicts of interest extend beyond financial influences is at the end of Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins et al. 2008), which sets out a framework for interpreting results of a review.
Declarations of interest: Holger Schünemann, Andrew Oxman, Gunn Vist, Paul Glasziou and Gordon Guyatt have, to varying degrees, taken leadership roles in the GRADE Working Group from which many of the ideas in this chapter have arisen.
Here, 4 of the 7 authors of the chapter have declared professional history and organisational affiliations. These are relevant because the GRADE scheme features heavily in the framework outlined by the chapter; understanding this derives from the authors’ involvement in GRADE helps explain this.
This shows Cochrane’s concern that a person’s experience can bias their interpretation of data, however well-intentioned they are – people will recommend what they know and trust, but what they know and trust may not be the best thing to recommend.
In comparison, declarations of interest and clarity of roles of involved parties in the genesis and conduct of the European Food Safety Authority’s Opinion on the use of Thresholds of Toxicological Concern (TTCs) in risk assessment appear limited (EFSA 2012).
Starting with the positives, it is at least made clear the TTC was developed by EFSA’s Scientific Committee on the basis of preparatory work by the TTC Working Group. All members of both the Committee and the Working Group are listed. Members’ declarations of interest are available on-line, separate to the document.
However, there is no dedicated section describing which specific persons or organisations were responsible for the various components of the Opinion; this information is stated in an acknowledgement footnote at the beginning of the document.
Although conception of the Opinion is attributed to EFSA as an organization (as a task it gave to its own Scientific Committee), it is not obvious whether the Working Group or another party had a role in focusing the Opinion on the specific topic of the use of TTCs.
Research by Pesticides Action Network (PAN) Europe (PAN 2011) indicates that the contribution of the Working Group to the genesis of the TTC Opinion was much more substantial than the report of the Opinion suggests. According to PAN the TTC Working Group organised 17 meetings between 2008 and 2011, presenting an “advanced draft” after each meeting. PAN’s research suggests the WG was in fact instrumental in developing the Opinion, with the Scientific Committee playing only a “more formal role”.
This is of particular potential concern because, according to PAN’s research, 8 of the 11 TTC Working Group members have a publishing history favouring the use of TTCs. Although this need not immediately disqualify the opinions of the Working Group members, within the framework of conducting Cochrane reviews declaration of this would at least be mandatory, particularly if the reviewers are likely to be appraising their own research (as must have happened in the TTC Working Group).
There are several reasons for this. For example, when people review their own data, they are more likely to view it in favourable light. They may also bring a preconceived approach to evaluating data which (though potentially valuable) can also skew analysis by down-playing of potentially important considerations which experts of different opinion, or non-specialists in the subject-matter, might bring to the table. Since active researchers are likely to see a problem in a certain way, the manner in which data is evaluated could be shaped in such a way as to lead to their preferred conclusion rather than the one ultimately best-supported by the overall body of existing research.
The employment of people with a publishing history in a subject under review does not invalidate review findings but it does increase the risk that a review will only present a partial view of treatment options.
Given the focus of the EFSA Opinion specifically on TTCs and their validity (little in the way critique is offered), the dominance of the Working Group by experts with a publishing history favouring the use of TTCs could explain why the Opinion is focused in this direction. If these interests had been declared rather than uncovered by PAN, it could have led people to interpret the findings of the Opinion very differently; hence the need for declaration not only of financial but also relevant research interests.
Cochrane Collaboration guidance on conflicts of interest
Authors should report any present or past affiliations or other involvement in any organization or entity with an interest in the review that might lead to a real or perceived conflict of interest. Situations that might be perceived by others as being capable of influencing a review author’s judgements include personal, political, academic and other possible conflicts, as well as financial conflicts.
Authors must state if they have been involved in a study included in the review. A summary of the Collaboration’s policy on conflicts of interest appears in Chapter 2 (Section 2.6).
Financial conflicts of interest cause the most concern, and should be avoided, but must be reported if there are any. Any secondary interest (such as personal conflicts) that might unduly influence judgements made in a review (concerning, for example, the inclusion or exclusion of studies, assessments of the validity of included studies or the interpretation of results) should be reported. (Higgins et al. 2008, Chapter 4)
Note that commercial sponsorship of a Cochrane review is prohibited.