Reading research
Reading research
This is not a review of specific studies but a guide to reading them. Here you will find the main types of study and what each can show, a step-by-step way to check a story about a "breakthrough", and the headline words that should make you cautious. You will not find summaries of individual papers or conclusions about specific treatments.
A single study rarely changes medicine. Usually it adds one more piece to a big picture that is put together over years.
Types of study
| Type | What it shows | Limitations |
|---|---|---|
| Randomised controlled trial (RCT) | Whether an intervention works compared with placebo or another treatment | Participants may differ from typical patients; with a small sample, results are unstable |
| Cohort study | How habits or living conditions relate to later events in a large group of people | Shows an association but does not prove cause: other factors affect the result |
| Case-control study | How people with a disease differed in the past from people without it | Depends on participants' memory and how groups were selected; does not prove causation |
| Systematic review | What all relevant studies on a question say when taken together | The conclusion is no better than the quality of the studies included |
| Meta-analysis | A pooled statistical result from several studies | Only comparable studies can be combined; errors in the original data remain |
| Cell and animal studies | Possible mechanisms and ideas for further testing | Results do not translate directly to humans |
How to read a news story about a study
Find the original source
Look for the name of the journal or organisation in the story. If there is no source, treat the conclusions with caution.
Check whether the work was peer-reviewed
A preprint is a paper that has not yet been peer-reviewed. Its conclusions may still change.
The phrase "published on a preprint server" means independent experts have not yet checked the paper.
Find out who was studied
Cells, animals or people. Conclusions about people can only be drawn from studies in people.
Identify the type of study
Observational studies show an association, while an RCT allows conclusions about cause. Check the table above.
Look at the sample size and comparison group
The fewer the participants, the higher the chance of a random result. Without a control group it is hard to tell the effect of a method from the placebo effect or the natural course of the illness.
Compare relative and absolute risk
Relative risk shows by what factor the probability changed; absolute risk shows by how much. Both matter for a decision.
If a story only says "twice as likely", look for how many people out of how many fell ill in each group.
Red flags in headlines
- "Scientists have proved" when it is a single study
- "Cure for cancer found" when the work was done on cells or mice
- "Food X lowers risk" when the study is observational
- Percentages without baseline figures: "risk reduced by 50%"
- A reference to "a study" without naming the journal or organisation
- A call to buy the product in question urgently
Related sections
- Methods and approaches
What counts as evidence-based, the main types of intervention and how to check a claim about a method.
- Medical glossary
Plain-language definitions of terms from research, prevention, lab tests and medicine leaflets.
- Sources and limitations
Which sources the reference relies on and what it cannot do.
Sources
- Cochrane Library
- Peer-reviewed medical journals
- General principles of evidence-based medicine
Information is current as of the preparation date. The Russian check-up schedule, recommendations from health organisations and medicine registration rules may change — check current arrangements with your clinic and official sources.