Finding and reading published research
Behind most health headlines is a published paper. Reading the original is easier than it sounds, and it is one of the best ways to check what a story really said.
Where research is published
Researchers usually report their work in scientific journals. Each paper follows a familiar layout: a title, an abstract (a short summary), then the background, methods, results, discussion, and a list of references. Most papers also state who funded the work and whether the authors have any competing interests.
If you only have time to read one part, read the abstract, then skim the methods. The methods tell you who was studied and how. A careful abstract often says less than a news story does, which is a good sign. Be wary of a news story that is bolder than the paper behind it.
PubMed basics
PubMed is a free search site for biomedical research. It lists the title, authors and abstract of a very large number of papers, and links to the journal for the full text where it is available. Some full texts are free to read. Others sit behind a paywall.
- Search with specific words. A named condition plus a study type gives a shorter, more useful list than a single broad term.
- Use the filters. You can narrow results by article type, such as clinical trial or systematic review, and by date.
- Read the record carefully. A paper being listed on PubMed means it is indexed. It does not mean anyone has endorsed it.
The Cochrane Library is another place to look. It collects systematic reviews of health evidence, written to a strict method.
Types of publication
- Journal article (original research)
- Reports a new study, with methods and results.
- Systematic review
- Searches for and combines all suitable studies on one question, using set rules. A meta-analysis adds statistics to pool the numbers.
- Editorial
- An opinion piece by a journal’s editors, often about a paper in the same issue.
- Commentary
- An expert’s view on a topic or study. Useful context, but opinion rather than new evidence.
- Report
- A document from an organisation, agency or committee. These vary a lot, so check who wrote it and how.
- Preprint
- A paper shared before peer review. It may change, or never be published.
Peer review, explained
Before most journals publish a paper, other experts in the field read it and comment. They look for weak methods, gaps in logic and claims that go beyond the data. Authors may need to revise, or may be turned down.
Peer review catches many problems, but not all. Reviewers usually cannot re-run the study or see every record, and they are often volunteers working fast. Think of it as a useful filter, not a guarantee.
Spotting retractions and corrections
Sometimes a journal withdraws a paper. This is called a retraction. The reasons range from honest mistakes to serious misconduct. Smaller problems are fixed with a correction notice.
PubMed and journal pages usually show a clear notice on retracted papers. If a study matters to you, search its title together with the word “retraction”, and look for notices on the journal page before you rely on it.
Conflicts of interest
A conflict of interest is a tie that could, even without meaning to, shape how research is done or reported. Examples include funding from a company with a stake in the result, or working for such a company. Having a tie does not make a study wrong. It is a reason to read more carefully.
Look near the end of the paper for statements headed funding, competing interests or disclosures. If you cannot find any, that is worth noting too. Honest papers say who paid for the work and who did what.