Considering intellectual, in addition to financial, conflicts of interest proved important in a clinical practice guideline

Article type
Authors
Akl EA1, El-Hachem P2, Abou-Haidar H3, Neumann I4, Schunemann HJ4, Guyatt GH4
1American University of Beirut, Lebanon
2Mount Sinai School of Medicine, USA
3McGill University, Canada
4McMaster University, Canada
Abstract
Background: The conflict of interest (COI) policy of the American College of Chest Physicians 9th iteration of the Antithrombotic Guidelines (AT9) restricted panelists from discussing recommendations on which they disclosed a primary COI. To what extent, beyond assessing financial COI, assessing intellectual COI affected management of COI is uncertain.

Objectives: To describe financial and intellectual COI among AT9 panelists and assess how frequently intellectual COI would have, in the absence of financial COI, resulted in restrictions on participation in decision-making.

Methods: We classified financial and intellectual COI into primary (leads to voting restriction) and secondary (no restrictions). We analyzed COI disclosures of panelists and in recommendations as units of analysis.

Results: Of 104 panelists, 102 made 4030 disclosures for 431 recommendations. The median number (and range) of recommendations for which the panelists disclosed COI was: 0 (0–33) for secondary financial COI, 0 (0–21) for primary financial COI, 1 (0–63) for secondary intellectual COI, and 0 (0–32) for primary intellectual COI. Of the 102 panelists, 37 (36%) disclosed a primary intellectual but no primary financial COI for at least one recommendation. Among 431 recommendations, the median number (and range) of panelists per recommendation who disclosed COI was: 0 (0–4) for secondary financial COI, 0 (0–5) for primary financial COI, 1 (0–6) for secondary intellectual COI, and 0 (0–7) for primary intellectual COI. Of the 431 recommendations, 63 (14.6%) had at least one panelist with a primary intellectual COI but no primary financial COI.

Conclusions: There was relatively low prevalence of COI in AT9. The distribution of COI was skewed (many with none, some with many). In the absence of financial COI, a substantial number of disclosures would have resulted in restrictions based on intellectual COI. The Cochrane Collaboration should ask systematic review authors to disclose both their financial and intellectual COI.