How to read a study abstract before you trust the headline

This is not clinical evidence and not medical advice. It does not replace your care team. Discuss anything here with your clinician before you act on it. Someone forwards a PubMed link. Or a headline links a study, and the study turns out to be one paragraph on a screen. That paragraph is the abstract: a short summary that sits on top of a much longer paper. NCCIH describes it as a brief description of the key points you will find in the paper. Many abstracts now come with labeled sections. NLM calls this a structured abstract, with headings such as Introduction, Methods, Results and Discussion so readers can take it in quickly. On PubMed, those labels show in capital letters followed by a colon. The labels are your map. Read them in order and you will know roughly what kind of study this was before you decide what it means.

This is where the authors say what question they set out to answer. NCCIH lists objectives as the questions the researchers hope to answer. Read it slowly. A study built to ask whether a drug is safe at a given dose is asking something different from a study built to ask whether it works better than current care. Also check the date on the paper. NCCIH suggests looking at whether the study is one year, five years or fifteen years old, because standards and practice change and newer research can support or contradict older work.

Methods is the short part most people skip, and it carries the most weight. NCCIH describes methods as the type of research, the study design, the participants, the outcomes measured and the steps taken to avoid bias. MedlinePlus suggests the same handful of questions for health news: did the research involve people or animals? If people, how many, and who were they? How long did the study run? What type of study was it? Who paid for it? Write those five answers next to the link. Also look for a comparison group. In a clinical trial, NCI's dictionary describes the control group as the group that does not receive the new treatment, so the two groups can be compared to see if the new treatment works. No comparison group means much less to compare against.

In a randomized clinical trial, NCI says, people are divided by chance into groups that compare different treatments or other interventions. Using chance makes the groups similar, so the effects can be compared more fairly. At the time of the trial, it is not known which treatment is best. An observational study is different. NCI defines it as a study in which people are observed or outcomes are measured, with no attempt to affect the outcome, for example no treatment is given. Both kinds have their place. When you see an observational design, though, consider asking your care team how much the groups might have differed in ways the authors could not control.

It counts as early science, not as evidence a treatment helps people. NCI defines a preclinical study as research using animals to find out whether a drug, procedure or treatment is likely to be useful, done before any testing in humans. In vitro, in NCI's dictionary, means in the laboratory, outside the body. NCCIH places basic research as the first step in any research, with clinical trials in people giving the clearest information on whether a treatment is effective and safe in humans. Those trials usually come only after smaller preliminary studies. So look for the words mice, rats, cell lines or in vitro in the methods line. If they are there, 'promising' describes a lab result. NCCIH also describes systematic reviews and meta-analyses, which look across many studies on one topic; when many studies reach the same conclusion, that helps show the results are reliable.

Find the primary outcome first. NCI's dictionary defines the primary endpoint as the main result measured at the end of a study to see if a treatment worked, such as the number of deaths or a difference in survival between groups. It is decided before the study begins. Anything else reported is secondary to that main question. An abstract that leads with a secondary result, or a subgroup, while the main outcome sits further down, deserves a second read. Two kinds of numbers often show up here. If the result is a percent change in risk, our blog piece "Absolute risk vs relative risk in cancer headlines" walks through how to read it. If it is a survival time, "What median survival means in a cancer paper" covers that number.

Treat it as the authors' reading of their own results. NCCIH describes the discussion or conclusion as what the results mean and the authors' interpretation of them. That is useful, but it is still interpretation. Hold the conclusion up against the methods: who was studied, how many, for how long, and whether there was a comparison group. MedlinePlus warns to be wary of dramatic writing, promises of cures and claims that sound too good to be true. NCI says the same in its guide to trustworthy cancer information: if a source makes claims that are too good to be true, they usually are.

A lot. The abstract is a summary of key points; NCCIH notes that the article itself holds the detailed methods, results, discussion and references. PubMed records contain the citation and the abstract, not the full article. The abstract page links to the full text when it is available, for example from the publisher or from PubMed Central, NIH's free archive. Some publisher sites charge a fee. PubMed suggests the Free full text filter, your hospital or university library, or a local librarian who may be able to get a copy through interlibrary loan.

MedlinePlus puts it simply: ask questions before you trust what you read, and discuss it with your health care provider, who can tell you whether good information is good for you. Consider bringing the link plus your notes on these points: was this in people, animals or cells? How many people, and how similar were they to me? Was it randomized, and what was the comparison group? What was the primary outcome, and did the study meet it? How long were people followed? Who paid for it? Then consider asking the bigger one: does this change anything for my situation, or is it too early to say? NCI's advice on reading cancer articles fits here too: speak with experts, ask lots of questions and decide whether the information applies to you.

It is a place to explore public cancer research and find the source behind a headline, so you can open the abstract yourself. It is not a second opinion. Triangle has not clinically reviewed the science for your case. Use it to find the study, use the questions above to read it, then take what you found to your clinician.