Conflict of interest guidelines for clinical guidelines

To the Editor: We welcome the article by Williams and colleagues on conflicts of interest in clinical guidelines,1 and agree that such conflicts can undermine trust in guidelines, compromise patient care and skew health care spending. One challenge in managing conflicts is the lack of evidence that disclosure minimises their impact. For example, research at the United States Food and Drug Administration found that, despite disclosure, for every committee member with a conflict, there was a 10% greater likelihood of the meeting favouring the drug reviewed.2 Recusal (self-disqualification) is also problematic, especially in Australia where there may be a relatively small pool of experts available. Given these problems, we are concerned that some of the recommendations made by Williams and colleagues lack supporting evidence and fail to adequately address conflicts of interest.

One of their recommendations, a tiered system within guideline panels limiting those with conflicts to contributions regarding the nature of the disease under discussion, for example, ignores what we know about the way that conflicts of interest operate. Research has shown that this apparently innocuous element of guideline development is not immune to influence, as conflicts of interest can influence how disease is defined, leading to broader disease definitions and lower treatment thresholds.3

A second recommendation was that guideline developers declare that “any benefits received have not influenced their professional opinion”. But surely this misses the real danger of such conflicts, which is not that of making deliberately biased decisions. Rather, the issue is that professional opinion may be unconsciously swayed by conflicts.4

Finally, they suggest an exclusion period of 3 years between receipt of industry support and involvement in guideline development. But no evidence is provided to ensure that this is “sufficient time for the conflict to have lost its potential for ‘undue influence’”. Without supporting evidence, the choice of 3 years, rather than 4 or 2 and a half, appears arbitrary.

We need more research into the way conflicts of interest operate, and perhaps we need to seek a fresh pool of experts who, while less experienced, may be significantly less conflicted.