What counts
Financial interests are the clearest: employment or consultancy with an organisation with a stake in the findings, stock holdings, patents, paid speaking or advisory roles, and research funding from a commercial body whose product is being evaluated.
Non-financial interests count too, and are more often missed. Personal relationships with people affected by the outcome, strongly held positions publicly advocated, unpaid roles in organisations with an interest, and close professional rivalries can all qualify.
The test is not whether you were actually influenced — you are not well placed to judge that about yourself. It is whether a reasonable reader, knowing the facts, might wonder. That framing resolves most borderline cases quickly.
The test is not whether you were influenced. It is whether a reasonable reader, knowing the facts, might wonder.
Disclosure protects you
Authors hesitate because disclosure feels like an admission of bias. It is the opposite: a declared interest is context a reader can weigh, and it demonstrates you understand the standards of your field.
The damage comes from non-disclosure. An undisclosed interest that surfaces later — and they do surface, through funding records, patent filings and institutional disclosures — becomes a story about concealment rather than about the interest itself, and can lead to a correction or a retraction notice.
Journals very rarely reject papers over declared interests. They act decisively over undeclared ones.
Writing it
Be specific and factual. Name the organisation, the nature of the relationship, and the period. "A.B. has received consultancy fees from [company] during the conduct of this study" is clear and complete.
Cover all authors — the statement is collective, and each author is responsible for declaring their own interests. In practice the corresponding author should ask every co-author explicitly rather than assume.
Where there is genuinely nothing, say so: "The authors declare no competing interests." Journals expect a positive statement rather than an omission, and most disclosure periods look back a few years, so check the journal's window rather than only considering the present.