Evidence Types and Study Designs: From Systematic Reviews to Laboratory Studies
A plain guide to eleven kinds of health research, from systematic reviews to laboratory and animal studies, and what each one can and cannot tell you about a remedy.
Content last updated . This page has not been medically reviewed.
The short answer
Not all studies answer the same question. A study that follows thousands of people over time, a study that tests a substance on cells in a dish, and a doctor's write-up of one patient are all "research," but they show very different things. Before asking whether a result is strong, ask what kind of study it came from and what that kind of study is built to show.
A guide, not a rule: a well-run lower-level study can be more useful than a poor higher-level one, and results from animals or cells do not show an effect in people. See Evidence types and study designs.
This page walks through eleven common types. It is a reading guide, not health advice, and it does not rank remedies.
Eleven study types at a glance
| Study type | Plain meaning | What it can show | What it cannot show on its own |
|---|---|---|---|
| Systematic review | A search for all research on one question that meets set rules, then a combined appraisal [1] | A fair summary of what many studies found, with the methods written down in advance [2] | More than the studies inside it can support |
| Meta-analysis | A statistical method that combines the results of independent studies [3] | A pooled estimate across studies | Whether the studies combined were good ones |
| Randomized controlled trial | At least one test treatment and one control treatment, with people assigned by a random process [3] | A fair comparison between groups in the people studied | What happens outside that group, setting or time period |
| Cohort study | A defined group of people followed over time, with or without an exposure [4] | Patterns between exposures and later outcomes | That the exposure caused the outcome |
| Case-control study | Starts with people who have an outcome and a comparison group without it [5] | Differences between the two groups | How often an outcome happens in everyone |
| Cross-sectional study | Measures people at one particular time [6] | A snapshot of what is common | Which came first, or how things change |
| Case report | A clinical presentation that may be followed by further studies [3] | A signal worth looking into, including unexpected reactions | How common anything is, or whether it is a pattern |
| Pharmacokinetic study | How a substance is absorbed, changed, distributed and eliminated at a given dose [7] | What happens to a substance in the body | Whether that has any effect on health |
| Animal study | Research on non-human animals chosen for specific characteristics [8] | How a substance behaves in a living animal body | What will happen in people [9] |
| Laboratory (in vitro) study | Study of processes in an artificial environment outside the living organism [10] | A proposed mechanism, or a first look at activity | That the same thing happens in a whole person |
| Practice guideline | Directions or principles to assist health care decisions, written by agencies, organizations or expert panels [3] | How a group of experts interprets the evidence at a point in time | New evidence; it is a reading of evidence, not a study |
Reading the table
Reviews sit on top of other studies. A systematic review and a meta-analysis do not produce new data from new participants. They gather and weigh existing studies. Systematic reviews try to minimize bias with a question and methods written down in a protocol before the work begins. [2] A meta-analysis is described as a quantitative way of combining independent studies, and it is usually named as such by its authors. [3]Interventional and observational are different things. In a trial, the investigators assign the treatment. In an observational study, the investigator does not assign participants to specific interventions. [3] That difference is why observational findings are usually described as associations.
Early-stage research asks different questions. Pharmacokinetic, animal and laboratory work tell researchers how a substance behaves and whether it is worth testing in people. They are not tests of whether it works in people. A systematic review that compared animal experiments with human trials found that they sometimes disagreed, and its authors suggested the disagreement may be due to bias or to animal models not mimicking human disease well enough. [9]
Guidelines are interpretations. A practice guideline is a set of directions or principles to assist health care decisions. [3] It can be useful, but it reflects the evidence available when it was written and the judgment of the group behind it.
Common misunderstandings
- "Studied" does not mean "shown to work." A substance can be studied many times in cells and animals without a single human trial.
- A review is not automatically strong. What a review can support depends on the studies it includes.
- One kind of study does not replace another. A trial does not make laboratory work meaningless, and a mechanism does not stand in for a trial.
- Null and mixed results count. A study that finds no difference, or finds harm, is as much part of the evidence as one that finds a benefit.
What this means in the Research Library
The library holds imported records, not written profiles, and it shows connections without rating effectiveness. Its study records include registered trials, systematic-review registrations and publication records. Publication records carry a research model label (animal, cell, clinical and so on) taken from the source data. A record with many connections is not better studied or more effective.To go deeper, read how registered trials differ from results, what human, animal and laboratory research can show, and how to read Research Library records. For general claims checking, see evaluating safety claims.
Frequently asked questions
What is the difference between a systematic review and a meta-analysis?
A systematic review attempts to identify, appraise and synthesize all the empirical evidence that meets pre-specified eligibility criteria to answer a specific question. [1] A meta-analysis is the statistical combining of results from independent studies. [3] A review may include a meta-analysis, but it does not have to.
Is a randomized controlled trial always better than an observational study?
They answer different questions. In a randomized trial the treatment is chosen by a random process, which allows a fair comparison between groups. [3] In an observational study the investigator does not assign interventions. [3] Either can be done well or badly, so the design alone does not settle quality.
Why can a case report not tell me whether a remedy works?
A case report describes a clinical presentation that may be followed by evaluative studies. [3] It describes what happened in a case, not how often it happens or why.
Does a laboratory result mean a remedy works in people?
No. In vitro methods study reactions outside the living organism. [10] Preclinical work, which includes in vitro and in vivo studies, is done before testing in people. [11]
Where does a pharmacokinetic study fit?
It describes absorption, transport and elimination of a substance as a function of dose. [7] That helps explain how a substance moves through the body. It is not a test of whether it changes a health outcome.
References
- Cochrane. About Cochrane reviews: what is a systematic review?. Read 2026-10-22.
- Lasserson TJ, Thomas J, Higgins JPT. Cochrane Handbook for Systematic Reviews of Interventions, Chapter 1: Starting a review. Read 2026-10-22.
- National Library of Medicine. Publication Types (MeSH). Read 2026-10-22.
- National Library of Medicine. Cohort Studies (MeSH). Read 2026-10-22.
- National Library of Medicine. Case-Control Studies (MeSH). Read 2026-10-22.
- National Library of Medicine. Cross-Sectional Studies (MeSH). Read 2026-10-22.
- National Library of Medicine. Pharmacokinetics subheading (MeSH). Read 2026-10-22.
- National Library of Medicine. Models, Animal (MeSH). Read 2026-10-22.
- Perel P, Roberts I, Sena E, et al. Comparison of treatment effects between animal experiments and clinical trials: systematic review. BMJ. 2007;334:197. Read 2026-10-22.
- National Library of Medicine. In Vitro Techniques (MeSH). Read 2026-10-22.
- U.S. Food and Drug Administration. The Drug Development Process, Step 2: Preclinical Research. Read 2026-10-22.