Does source of funding and conflict of interest influence the outcome and quality of spinal research?

Background/context: There has been longstanding controversy surrounding the influence of funding source on the conduct and outcome of medical research. In 2011, a systematic review of the use of recombinant bone morphogenetic protein-2 revealed underreporting of unfavorable outcomes in some industry-sponsored trials. We hypothesize that Industrial funding and the presence of potential conflict of interest will be associated with low levels of evidence (LOE) and greater proportions of favorable outcomes in spinal research.

Purpose: The aim of this study is to investigate the association between funding source and potential conflict of interest on the LOE and study outcome in the current spinal research.

Study design/setting: Systematic review of all the spinal publications in five leading spinal, orthopedics, neurosurgery, and general medical journals during 2010 (print and online). Supplements were included.

Outcome measure: Outcome and the LOE of research papers.

Methods: Two reviewers independently assessed all publications. Commentaries, editorials, letters, open operating theatres, case reports, narrative reviews, and study protocols were excluded. The self-reported potential conflict of interest and type of funding was extracted from each paper. Funding type was classified as foundation, industry, public, intramural, multiple (including industry), multiple (without industry), and unfunded. The outcome of each study was classified as favorable, unfavorable, equivocal, or not applicable. Clinical publications were ranked using the LOE guidelines produced by the Oxford Center for Evidence-Based Medicine.

Results: Overall, 1356 papers were analyzed, out of which 864 were suitable for LOE grading. There was good interobserver reliability for assignment of LOE grade, κ=0.897 (p<.01) and study outcome κ=0.804 (p<.01). A significant association was found between LOE and source of funding (p<.01). Industry-funded studies had the greatest proportion of level IV evidence (65%). There was a significant association between the funding source and study outcome (p=.01). The proportion of industry-funded studies with favorable outcomes (88%) was higher than that of publicly and foundation-funded studies (73% and 74%, respectively). The associated odds ratio for reporting favorable outcomes in industry-funded studies compared with studies with public and foundation funding was 2.7 (95% confidence interval [CI], 1.4-5.3), and 2.6 (95% CI, 1.3-5.2), respectively. A significant association between LOE and study outcome (p<.01) was also identified. Level I studies had the highest proportions of unfavorable (14%) and equivocal (23%) outcomes. Level IV studies had the highest proportion of favorable outcome (85%). There was no association between self-reported conflict of interest and LOE (p=.83) or study outcome (p=.25).

Conclusion: We demonstrated a significant association between source of funding, study outcome, and LOE in spinal research. A large proportion of industry funded research was shown to provide level IV evidence and report favorable outcome.