Conflicts of interest and how to find them
Where disclosures live in a paper, what funding and authorship arrangements do to results, and why disclosure is a starting point rather than a resolution.
Conflicts of interest are disclosed in a declaration section, usually near the end of a paper, alongside a funding statement. Both are easy to find and routinely skipped by readers.
Disclosure identifies a relationship, it does not neutralise it. Research on funding effects across medicine consistently finds that industry-funded studies report favourable results more often than independently funded ones. That is a reason to weight, not to dismiss, and the appropriate response is to look harder at the design rather than to discount the result automatically.
Where to look
In a journal article, look for four things. The funding statement, naming who paid. The conflict of interest declaration, naming relationships. The author affiliations, which may include the manufacturer. And the acknowledgements, which sometimes reveal that a company provided material, performed analysis or assisted with writing.
That last one is worth dwelling on. Assistance with writing means someone other than the named authors drafted text. It is disclosed and it is common in industry-supported publication. It is not misconduct when disclosed, and it does mean the framing of the paper may have been shaped by a party with an interest.
The forms it takes
- Direct funding of the study by a manufacturer.
- Employment or consultancy of authors by an interested party.
- Equity in a company whose product is studied.
- Patents held on the technology.
- Material supplied free, which is a contribution even when no money moves.
- Speaker fees and advisory roles, often disclosed as a list.
- Non-financial interests, including having publicly advocated a position that the study tests.
The last is real and rarely disclosed. A researcher who has built a career on a hypothesis has an interest in it, and no declaration form captures that.
A disclosed conflict of interest means a study's results can be disregarded.
- Proposed mechanism
- Financial interest produces biased results.
- What has been shown
- Research across medicine consistently finds an association between industry funding and favourable reported outcomes. The mechanisms proposed include study design choices, comparator selection, endpoint choice, analysis decisions and selective publication rather than fabrication. An association at the level of a literature does not establish that any specific study was distorted.
- Highest level reached
- Larger human studies
- Main confounders
- Industry funds studies of products it believes will succeed, so selection is part of the explanation. Independent studies are not free of interest either, including career and publication incentives.
GradeSUPPORTED, CONSISTENT
What would change thisLittle at the level of the association, which is well replicated. What would change the assessment of a specific study is examining its design directly: pre-registration, comparator, blinding, endpoint specification and whether the analysis matches the registered plan.
How funding influences without fabrication
The unhelpful version of this discussion assumes fraud. The mechanisms identified in the literature are subtler and mostly involve decisions that are defensible individually.
- Comparator choice. Comparing against no treatment rather than an active alternative makes a favourable result more likely.
- Dose and protocol. Selecting favourable conditions for the product and less favourable ones for the comparator.
- Endpoint selection. Choosing the measure most likely to move.
- Timepoint selection. Assessing when the effect is largest.
- Analysis choices. Which participants are included, how missing data are handled.
- Publication decisions. Studies with unfavourable results are less likely to be submitted or published, which distorts the literature without distorting any individual paper.
The last is the largest effect and the least visible, because it operates on the set of papers that exist rather than on their contents. Trial registration exists partly to make it detectable, by creating a public record of studies that were started.
| Check | Where | What you learn |
|---|---|---|
| Funding statement | End of paper | Who paid |
| Conflict declaration | End of paper | Author relationships |
| Affiliations | First page | Whether authors work for an interested party |
| Acknowledgements | End of paper | Material supplied, analysis performed, writing assistance |
| Registration number | Methods or abstract | Whether the study was registered and whether the plan matches |
Our own position
It would be incoherent to write this article without stating our own interests. This publication earns money from paid listings and from a labelled sponsor line in a newsletter. Both are disclosed on every appearance and neither can influence an evidence grade. We publish what we refuse at any price on the rate card, and the full model is on the about page.
We also have a non-financial interest worth naming: a publication built around a grading device has an interest in the device appearing useful, and therefore an interest in claims being gradeable. We mitigate that by using the grade NOT ASSESSABLE where a claim cannot be tested as worded, rather than forcing it onto the scale. Readers can check whether we do so consistently, because every panel is public.
Interests that are not financial
The declaration form is built for money, and money is the easiest interest to disclose because it leaves records. Several of the interests that most shape a literature leave none.
- Intellectual commitment. A researcher who has spent a decade on a hypothesis is not neutral about it, and no form captures that.
- Career incentive. Positive, novel findings advance careers more reliably than null ones. That pressure operates on every author regardless of funding.
- Field-level interest. Everyone working in an emerging area benefits if the area looks promising, including reviewers and editors of its specialist journals.
- Clinical investment. A practitioner who has bought equipment or trained in a technique has committed resources to it before assessing it.
None of these is improper and none is avoidable. They are worth naming because a reader who checks only the funding line will conclude that a study with independent funding is interest free, and no study is. The right response is the same as for financial interests: look at the design.
Practitioners and disclosure
The same logic applies in clinics. A practitioner who is a paid speaker for a manufacturer, or who holds distribution rights for a product, has an interest in recommending it. This is not improper and it is worth knowing, and it is a reasonable thing to ask directly. The question lands better than people expect, and a practitioner who answers it straightforwardly has told you something useful about how they work.
Questions readers ask
Where do I find conflicts of interest in a paper?
In the declarations near the end, alongside the funding statement, plus the author affiliations on the first page and the acknowledgements, which may reveal supplied material, analysis support or writing assistance.
Does industry funding mean a study is wrong?
No. Research across medicine finds an association between industry funding and favourable reported outcomes, which is a reason to examine the design carefully rather than to discard the result. Selection of which products to study is part of the explanation.
How does funding influence results without fraud?
Through defensible individual choices: comparator selection, dose and protocol, endpoint choice, timepoint selection, analysis decisions and whether unfavourable results are submitted for publication at all.
What is publication bias?
The tendency for studies with unfavourable results to be less likely to be published, which distorts the literature as a whole without distorting any individual paper. Trial registration exists partly to make it detectable.
Should I ask my practitioner about their interests?
It is a reasonable question. A practitioner who is a paid speaker for a manufacturer or who holds distribution rights has an interest in recommending that product. Asking is not an accusation, and a straightforward answer tells you something useful.