Bias from industry trial funding? A framework, a suggested approach, and a negative result

Abstract

Bias from funding sources of trials would threaten their validity. Meta-analyses of high quality acute pain and migraine trials were used to explore the hypothesis that industry funding of clinical trials produced more favourable results than non-profit sponsorship. Analyses were planned to evaluate whether industry-sponsored trials had different results from trials funded by academic or other non-profit sources, but of 176 trials, only two were supported by non-profit sources, while 31 provided no statement of support. An alternative method is proposed within industry-sponsored trials, looking at conflicting industry interests for the same drug, used either as test or comparator treatment. Fifty-three trials used an analgesic as test and 90 as comparator, allowing comparisons to be made for aspirin 600/650 mg, ibuprofen 400 mg, paracetamol (acetaminophen) 1000 mg, rofecoxib 50 mg and sumatriptan 50 and 100 mg. Only for sumatriptan 50 and 100 mg, with the outcome of headache response at 2 h, was there any significant difference between the drug used as a test or as a comparator. The direction was for higher (worse) NNTs with sumatriptan as comparator. Investigating potential industry bias through the funding source of trials is unlikely to be adequate because of a dearth of trials funded by non-profit organisations. We propose a method based on potential conflict of interest within industry-sponsored trials. Using this method, established clinical trial results in acute pain and migraine appear to be unbiased.