Article type
Year
Abstract
Background: Over recent years, research has shown that pharmaceutical industry funding may be associated with published outcomes that are favourable to the funder. [1,2] Device companies have similar financial interests to pharmaceutical companies, however it has yet to be demonstrated that a similar bias exists in device trials.
Objectives: This study aimed to determine whether a relationship exists between the funding source in endometrial ablation surgical device trials and the selection and statistical significance of the results.
Methods: A systematic search of online databases was conducted to identify randomised controlled trials published in English from 1994 comparing the efficacy of different endometrial ablation techniques for women with heavy menstrual bleeding. Trials were classified as industry funded, not-for-profit funded, mixed (a combination of industry and not-for-profit funding) or not reported. The primary outcome of the trial was defined as the first reported outcome if there was no stated primary outcome or power calculation.
Results: Thirty-four publications representing 22 studies were identified which met all eligibility criteria (8 industry funded, 7 not-for-profit, 7 funding source not reported). Among these, 21 stated a primary outcome or reported an outcome in the power calculation. Nine studies reported a statistically significant primary outcome, with four favouring the comparator. There was no detectable association between type of funding and the statistical significance of the primary outcome (p=0.64) or between trial funding and the reporting of a positive outcome (p=0.60). The trials reported nine different primary outcomes, of which two were considered patient-centred outcomes (patient satisfaction and re-intervention rates). There was a significant difference in funding source between trials which used patient-centred primary outcomes and those which did not, with industry funded trials less likely to report patient-centred outcomes as the primary outcome (p=0.02).
Conclusions: This small study found no evidence that device company sponsorship is associated with the reporting of positive outcome. Our finding that funding source is associated with the selection of patient-centred outcomes deserves further consideration. Further research to address these questions should be conducted with a larger number of studies.
References:
1 Bekelman JE et al. 2003,JAMA,289(4),454-465.
2 Lexchin J et al. 2003,BMJ,326,1167.
Objectives: This study aimed to determine whether a relationship exists between the funding source in endometrial ablation surgical device trials and the selection and statistical significance of the results.
Methods: A systematic search of online databases was conducted to identify randomised controlled trials published in English from 1994 comparing the efficacy of different endometrial ablation techniques for women with heavy menstrual bleeding. Trials were classified as industry funded, not-for-profit funded, mixed (a combination of industry and not-for-profit funding) or not reported. The primary outcome of the trial was defined as the first reported outcome if there was no stated primary outcome or power calculation.
Results: Thirty-four publications representing 22 studies were identified which met all eligibility criteria (8 industry funded, 7 not-for-profit, 7 funding source not reported). Among these, 21 stated a primary outcome or reported an outcome in the power calculation. Nine studies reported a statistically significant primary outcome, with four favouring the comparator. There was no detectable association between type of funding and the statistical significance of the primary outcome (p=0.64) or between trial funding and the reporting of a positive outcome (p=0.60). The trials reported nine different primary outcomes, of which two were considered patient-centred outcomes (patient satisfaction and re-intervention rates). There was a significant difference in funding source between trials which used patient-centred primary outcomes and those which did not, with industry funded trials less likely to report patient-centred outcomes as the primary outcome (p=0.02).
Conclusions: This small study found no evidence that device company sponsorship is associated with the reporting of positive outcome. Our finding that funding source is associated with the selection of patient-centred outcomes deserves further consideration. Further research to address these questions should be conducted with a larger number of studies.
References:
1 Bekelman JE et al. 2003,JAMA,289(4),454-465.
2 Lexchin J et al. 2003,BMJ,326,1167.