How evidence is graded

Why funding is recorded

When a manufacturer pays for a study of its drug or device, favourable results are reported more often ¹.

In a Cochrane review, favourable efficacy results were more frequent with industry sponsorship: 638 against an assumed 502 per 1000 (risk ratio 1.27, 95% CI 1.17 to 1.37) ¹.

Why regulatory documents are read

Trials with disappointing results are published less often than those with favourable results ².

Manufacturers must report all phase II and III trial results to the US Food and Drug Administration (FDA) for approval ². Of trials judged negative by the FDA, 4 of 37 (11%) for older antidepressants and 7 of 15 (47%) for newer ones were published in agreement with it ².

The classes

Each cited publication is assigned one of six classes.

Class	Criterion
independent	Public or charitable funding, without industry money or an industry role. Donated products are permitted.
industry	Any industry funding, or an industry role in design, analysis or publication
regulatory	Assessments, labels and safety communications of regulators and statutory risk assessors: FDA, EMA, MHRA, PMDA, EFSA, FSANZ, JECFA
consensus	Guidelines, consensus statements, position papers, expert-panel classifications such as IARC
reference	Textbooks and curated databases
undetermined	Funding not determinable from the publication

The layers

On every page, independent findings are kept apart from other evidence.

Layer	Permitted classes
Main text	independent, regulatory, reference
Industry	industry, regulatory
Consensus	consensus, regulatory
Unresolved	all classes

The same evidence is required for a statement against consensus as for one in agreement with it. The absence of independent evidence is stated in the unresolved layer.

Absolute and relative effects

A relative effect is the fraction by which a treatment changes a risk. An absolute effect is the change in the number of people per hundred who have the outcome.

In SPRINT, 9361 adults were assigned a systolic target below 120 or below 140 mm Hg ⁷. The primary cardiovascular outcome occurred in 243 of 4678 participants (5.2%) and 319 of 4683 (6.8%) over a median 3.26 years (hazard ratio 0.75) ⁷. The relative reduction is 25%, the absolute reduction 1.62 percentage points, and the reported number needed to treat 61 ⁷. Serious adverse events possibly or definitely related to the intervention occurred in 220 (4.7%) and 118 (2.5%) participants ⁷. The number needed to harm, the reciprocal of the 2.18 percentage-point difference, is 46 ⁷.

Limits of this method

Funding is one source of error among several ¹. Trials without industry money can also be at risk of bias ¹.

Study design is not graded by class. Low serum 25-hydroxyvitamin D had been linked with cancer and cardiovascular risk in observational studies ⁸. In VITAL, an independent trial, colecalciferol reduced neither invasive cancer (6.1% of 12,927 against 6.4% of 12,944; hazard ratio 0.96) nor major cardiovascular events (3.1% against 3.2%; 0.97) over a median 5.3 years ⁸.

References

Independent research
Lundh A, Lexchin J, Mintzes B, Schroll JB, Bero L. Industry sponsorship and research outcome Cochrane Database of Systematic Reviews, 2017.
doi:10.1002/14651858.MR000033.pub3 PubMed 28207928
Funding and finding
Funding
Internal sources: Center for Evidence-Based Medicine, Odense; The Nordic Cochrane Centre, Copenhagen (no financial support); University of Sydney; York University (author salaries). External sources: no sources of support supplied.
Finding
75 papers. Industry-sponsored drug and device studies more often had favourable efficacy results (RR 1.27, 95% CI 1.17 to 1.37; 25 papers, 2923 studies; 638 versus 502 per 1000) and favourable conclusions (RR 1.34, 1.19 to 1.51; 29 papers, 4583 studies; 863 versus 644 per 1000). Risk of bias did not differ for sequence generation, allocation concealment, follow-up or selective outcome reporting; industry studies more often had low risk of bias from blinding (RR 1.25, 1.05 to 1.50) and showed less agreement between results and conclusions (RR 0.83, 0.70 to 0.98).
Independent research
Turner EH, Cipriani A, Furukawa TA, Salanti G, de Vries YA. Selective publication of antidepressant trials and its influence on apparent efficacy: Updated comparisons and meta-analyses of newer versus older trials PLoS Medicine, 2022.
doi:10.1371/journal.pmed.1003886 PubMed 35045113
Funding and finding
Funding
The authors received no specific funding for this work.
Finding
30 trials of newer antidepressants with 13,747 patients. Transparent publication of FDA-negative trials rose from 11% for older antidepressants to 47% (7 of 15) for newer ones; all 15 FDA-positive newer trials were published transparently. Effect size inflation (journal minus FDA Hedges g) was 0.05 (0.29 versus 0.24) for newer and 0.10 (0.41 versus 0.31) for older antidepressants.
Funding unknown
Turner EH, Matthews AM, Linardatos E, Tell RA, Rosenthal R. Selective publication of antidepressant trials and its influence on apparent efficacy New England Journal of Medicine, 2008.
doi:10.1056/NEJMsa065779 PubMed 18199864
Funding and finding
Participants
12,564
Funding
not stated
Finding
Of 74 FDA-registered trials, 23 (31%), with 3449 participants, were not published. 48 of 51 published trials (94%) were reported as positive, against 38 of 74 (51%) in the FDA analysis. Journal-based effect sizes exceeded FDA-based effect sizes by 32% overall and by 11% to 69% for individual drugs.
Industry-funded
Marso SP, Daniels GH, Brown-Frandsen K, Kristensen P, Mann JF, Nauck MA, et al. Liraglutide and cardiovascular outcomes in type 2 diabetes New England Journal of Medicine, 2016.
doi:10.1056/NEJMoa1603827 PubMed 27295427
Funding and finding
Registration
NCT01179048
Participants
9,340
Follow-up
median 3.8 years
Funding
Supported by Novo Nordisk and by grants from the National Institutes of Health. A steering committee of 11 academic investigators and 4 sponsor employees designed the protocol in collaboration with the sponsor; the sponsor performed site monitoring and data collection; the data were analysed by Statogen Consulting and the sponsor; editorial support was funded by the sponsor.
Finding
Cardiovascular death, non-fatal myocardial infarction or non-fatal stroke in 608 of 4668 (13.0%) with liraglutide versus 694 of 4672 (14.9%) with placebo (HR 0.87, 95% CI 0.78 to 0.97); median follow-up 3.8 years; number needed to treat 66 over 3 years.
Expert consensus
Ross AC, Manson JE, Abrams SA, Aloia JF, Brannon PM, Clinton SK, et al. The 2011 report on dietary reference intakes for calcium and vitamin D from the Institute of Medicine: what clinicians need to know Journal of Clinical Endocrinology and Metabolism, 2011.
doi:10.1210/jc.2010-2704 PubMed 21118827
Funding and finding
Funding
Institute of Medicine committee assembled at the request of agencies of the US and Canadian governments.
Finding
Recommended Dietary Allowance of vitamin D 600 IU per day for ages 1 to 70 years and 800 IU per day from 71 years, corresponding to serum 25-hydroxyvitamin D of at least 20 ng/ml (50 nmol/l); tolerable upper intake level 4000 IU per day from 9 years. Evidence on cancer, cardiovascular disease, diabetes and autoimmune disorders was judged inconsistent, inconclusive as to causality and insufficient to set requirements.
Funding unknown
Ehrhardt S, Appel LJ, Meinert CL. Trends in National Institutes of Health funding for clinical trials registered in ClinicalTrials.gov JAMA, 2015.
doi:10.1001/jama.2015.12206 PubMed 26670975
Funding and finding
Funding
not stated
Finding
Newly registered interventional trials by year first received, 2006 to 2014: NIH-funded 1376 to 1048 (328 fewer, 24%); industry-funded 4585 to 6550 (1965 more, 43%); all trials 9321 to 18,400.
Independent research
SPRINT Research Group, Wright JT Jr, Williamson JD, Whelton PK, Snyder JK, Sink KM, et al. A randomized trial of intensive versus standard blood-pressure control New England Journal of Medicine, 2015.
doi:10.1056/NEJMoa1511939 PubMed 26551272
Funding and finding
Registration
NCT01206062
Participants
9,361
Follow-up
median 3.26 years; intervention stopped early for benefit
Funding
NIH contracts and an inter-agency agreement (National Heart, Lung, and Blood Institute; National Institute of Diabetes and Digestive and Kidney Diseases; National Institute on Aging; National Institute of Neurological Disorders and Stroke); Clinical and Translational Science Awards at several sites; resources and facilities of the Department of Veterans Affairs. NIH scientists participated in the design and held one vote on the steering committee; the coordinating centre analysed the data.
In-kind support
Azilsartan and azilsartan combined with chlorthalidone were donated by Takeda Pharmaceuticals International and Arbor Pharmaceuticals; neither company had any other role in the study.
Finding
Primary outcome in 243 of 4678 (5.2%; 1.65% per year) with a systolic target below 120 mm Hg versus 319 of 4683 (6.8%; 2.19% per year) with a target below 140 mm Hg (HR 0.75, 95% CI 0.64 to 0.89); number needed to treat 61 during the median 3.26 years. Serious adverse events possibly or definitely related to the intervention in 220 (4.7%) versus 118 (2.5%) (HR 1.88).
Independent research
Manson JE, Cook NR, Lee IM, Christen W, Bassuk SS, Mora S, et al. Vitamin D supplements and prevention of cancer and cardiovascular disease New England Journal of Medicine, 2019.
doi:10.1056/NEJMoa1809944 PubMed 30415629
Funding and finding
Registration
NCT01169259
Participants
25,871
Follow-up
median 5.3 years
Funding
NIH grants U01 CA138962 and R01 CA138962, with support from the National Cancer Institute, National Heart, Lung and Blood Institute, Office of Dietary Supplements, National Institute of Neurological Disorders and Stroke and National Center for Complementary and Integrative Health. The NIH sponsors had a collaborative role in design, conduct and interpretation; final decisions, data collection, management, analysis, review and approval of the manuscript and the decision to submit rested with the investigators.
In-kind support
Pharmavite LLC donated vitamin D3, matching placebo and calendar packaging; Pronova BioPharma and BASF donated Omacor fish oil; Quest Diagnostics measured serum 25-hydroxyvitamin D at no cost.
Finding
Vitamin D3 2000 IU per day versus placebo: invasive cancer in 793 of 12,927 versus 824 of 12,944 (HR 0.96, 95% CI 0.88 to 1.06); major cardiovascular events in 396 versus 409 (HR 0.97, 0.85 to 1.12); median 5.3 years.

Updated 2026-10-21