Table 2.
Relation between source of funding and outcome of research
| Author | Outcome question assessed by research | Results |
|---|---|---|
|
Funding source and publication status
|
|
|
| Dickersin et al17 | Factors associated with publication of findings from clinical trials | Publication rate at one of two centres was considerably higher for studies funded by National Institutes of Health than for studies funded by drug industry (90.5% v 65.0%), but no indication that there was any difference in the tendency to publish significant results |
| Easterbrook et al21 | Are clinical trials with statistically significant results more likely to be published and what is the magnitude of this bias | Drug company sponsored trials of any type are significantly less likely to be published or presented compared to unfunded studies (OR=0.36; 95% CI 0.16 to 0.81); drug company sponsored clinical trials are significantly less likely to be published or presented compared to unfunded studies (0.17; 0.05 to 0.53) |
| Ioannidis24 | Is time to completion and time to publication of clinical trials affected by statistical significance of results; description of natural history of trials | Trials where data managed by industry of shorter duration than those federally sponsored (P<0.01) but no difference in time to appearance in peer reviewed literature (P=0.33)† |
| Liebeskind et al31 | Evidence of publication bias in randomised controlled trials | Median time between enrolment and publication of industry funded studies with positive results 3.5 years v 4.7 years for negative studies; time difference for studies with any type of funding 3.5 v 4.4 years for industry funded studies |
| Massie et al33 | Whether source of funding of clinical trials on angina affected publication process | 48% of drug company funded trials published in symposium compared with 26% of studies with other types of funding or where source of funding not stated |
| Stern et al37 | Extent to which publication is influenced by study outcome | Pharmaceutical industry funding not a statistically significant predictor of time to publication |
|
Funding source and economic outcomes
|
|
|
| Azimi et al12 | How often do cost effectiveness analyses encourage a strategy requiring additional expenditures | Industry funded studies more likely to support a strategy requiring additional expenditures compared with those without such funding (9/10 v 15/34, P=0.01); among the 39 articles that supported additional expenditure the median cost effectiveness ratio of the 9 studies funded by industry was significantly higher than for the remainder (P=0.02) |
| Freidberg et al23 | Relation between drug company sponsorship and economic assessment of oncology drugs | Drug company sponsored studies more likely to report favourable qualitative conclusions (19/20 v 15/24, P=0.04); overstatement of quantitative results not significantly different (6/20 v 3/24, P=0.26) |
| Kamal-Bahl et al26 | Relation between drug company sponsorship and economic assessment of drugs for respiratory syncytial virus prophylaxis | Industry funded studies more likely to report possibility of cost effectiveness or cost savings of prophylaxis in entire high risk infant population either in point estimates or sensitivity analysis (4/4 v 0/8, P=0.002); when likelihood of reporting cost effectiveness or cost savings in either entire high risk populations or specific infant subgroups compared across studies, no difference between industry funded and non-industry funded studies (4/4 v 3/8, P=0.08) |
| Neumann et al34 | Relation between drug company sponsorship and cost effectiveness ratios and per cent cost saving reported | For incremental cost effectiveness ratio: industry funded studies $6000 v non-industry funded studies $13,000 (P=0.003); for per cent cost saving: industry funded studies 21% v non-industry funded studies 9% (P=0.002)‡ |
| Sacristan et al*36 | Relation between drug company sponsorship and results of cost effectiveness studies | General medical journals: 3/6 cost-effectiveness studies with industry funding had positive results v 31/63 with no funding or other source of funding; Pharmacoeconomics: 18/18 cost effectiveness studies with industry funding had positive results v 4/6 with no funding or other source of funding |
|
Funding source and outcomes of clinical trials and meta-analyses
|
|
|
| Chard et al13 | Results of clinical trials for osteoarthritis of the knee | Projects commercially funded more likely to support intervention than non-commercially or non-specified funding source studies (P=0.024)* |
| Cho et al14 | Relation between drug company sponsorship of clinical trials and study outcome | Proportion of trials with favourable outcome higher in drug company sponsored research than in trials without company sponsorship (39/40 v 89/112, P<0.01) |
| Clifford et al15 | Relation between funding source and trial outcome | Studies favouring new product: industry only funding 30/44; mixed industry and non-industry funding 16/22; not for profit funding 15/28 (P=0.461) |
| Davidson16 | Relation between drug company sponsorship of clinical trials and study outcome | Studies supported by companies significantly more likely to support new therapies than trials with other sources of funding or where funding not stated (33/37 v 43/70, P=0.002) |
| Dieppe et al18 | Relation between drug company sponsorship and outcome of meta-analyses of treatment of osteoarthritis of the knee | Trend for meta-analyses sponsored by pharmaceutical industry to show a significant beneficial effect compared with meta-analyses with other source of funding or unknown sources of funding (16/18 v 81/117, P=0.084) |
| Djulbegovic et al19 | Relation between drug company sponsorship of clinical trials of erythropoietin and study outcome | Drug company funded trials showed positive results in 21/23 studies (P<0.0001) compared with 5/7 studies funded partially or completely with public resources (P=0.096) |
| Djulbegovic et al20 | Relation between drug company sponsorship of clinical trials of the treatment of multiple myeloma and study outcome | Drug company funded trials showed positive results in 26/35 studies (P=0.004) compared with 54/101 studies funded by non-profit making organisations (P=0.608) |
| Freemantle et al22 | Investigate potentially confounding factors which may affect clinical trials that assessed the relative efficacy of antidepressants | Most important structural predictor of outcome was trial sponsorship—trend towards increased efficacy of sponsor's drug (coefficient 0.097; 95% CI -0.03 to 0.23) |
| Ioannidis24 | Is time to completion and time to publication of clinical trials affected by statistical significance of results; description of course of trials | No significant correlation between presence of statistical significance, study accrual, and whether data managed by industry or not (all correlation coefficients P<0.2)§ |
| Kemmeren et al27 | Relation between drug company sponsorship and outcome of studies comparing effects of second and third generation oral contraceptives on risk of venous thromboembolism | Odds ratio 1.3 (95% CI 1.0 to 1.7) for studies directly funded by pharmaceutical industry v 2.3 (1.7 to 3.2) in other studies |
| Koepp et al30 | Relation between drug company sponsorship of clinical trials on tacrine and study outcome | 0/5 studies without corporate support found clinical benefit; 6/7 studies with corporate support found benefit (1 study could not be located) |
| Mandelkern32 | Relation between drug company sponsorship of clinical trials in psychiatry and study outcome | 16/16 drug company funded studies reported favourable outcome; 10/16 non-industry funded studies reported favourable outcome |
| Thomas et al38 | Relation between drug company funding studies comparing topical glucocorticosteroids and whether study favours product made by sponsoring company ¶ | Product tested favoured in trial: 37/48 in company funded trials v 12/22 in trials with other sources of funding (P<0.001) |
| Vandenbroucke et al39 | Relation between drug company sponsorship and outcome of studies comparing effects of second and third generation oral contraceptives on risk of venous thromboembolism | 1/9 studies without industry funding found no higher risk of venous thromboembolism (relative risk 1.5-4.0, summary relative risk 2.4); 4/4 industry funded studies found no higher risk (relative risk 0.8-1.5, summary relative risk 1.1) |
| Wahlbeck et al40 | Relation between drug company sponsorship of clinical trials on clozapine and study outcome | Odds of relapsing significantly in favour of clozapine in drug company sponsored trials (odds ratio 0.5; 95% Cl 0.3 to 0.7) compared with non-sponsored trials (0.4; 0.1 to 1.4); both drug company sponsored and non-sponsored trials suggested clozapine mediates clinically important improvement compared with older drugs but industry sponsored trials were more positive (0.4; 0.2 to 0.7 v 0.3; 0.1 to 0.7); industry sponsored trials reported significantly fewer early dropouts compared with older drugs (0.5; 0.4 to 0.7) but non-sponsored trials did not (0.6; 0.3 to 1.2) |
| Yaphe et al41 | Relation between sources of support of research and published outcomes of randomised controlled drug trials | Positive outcomes: 181/209 trials with industry funding v 62/96 with non-industry funding (odds ratio 3.45, 95% Cl 1.90 to 6.62) |
*
Results for all 128 commercially sponsored trials (114 funded by pharmaceutical industry).
†
56 completed trials (of 98 where data was not managed by pharmaceutical industry) included in analysis.
‡
70 trials funded by pharmaceutical companies and 7 funded by device companies.
§
90 trials where full patient accrual completed (of 98 where data was not managed by pharmaceutical industry) included in analysis
¶
Outcome of studies assessed by independent evaluators.