Human, Animal and Laboratory Research: What Each Can Show

The short answer

Research in people, in animals and in laboratory dishes answers different questions. Only research in people can show what happens in people. Animal and cell studies are earlier steps: they help researchers understand how a substance behaves and decide whether it deserves testing in humans. A positive animal or cell result is not human evidence.

Three kinds of research

Why preclinical research comes first

Before a drug is tested in people, researchers must find out whether it has the potential to cause serious harm, also called toxicity. [4] The two types of preclinical research are in vitro and in vivo. [4] After preclinical testing, researchers review their findings and decide whether the drug should be tested in people. [4] Preclinical testing is defined as testing in experimental animals or in vitro for biological and toxic effects and potential clinical applications. [5]

Early human studies often lean on animal data: researchers adjust dosing schemes based on animal data to find out how much of a drug the body can tolerate. [6] That is a stepping stone, not a verdict.

Why results do not always carry over

Translational research is the effort to turn an observation in the laboratory, clinic or community into an intervention that will improve the health of individuals. [7] The step is not guaranteed.

A systematic review compared animal experiments with human trials for six treatments and found mixed agreement. [2] In head injury, corticosteroids did not show any benefit in clinical trials but did in animal models. [2] For the stroke drug tirilazad, patients had a worse outcome, even though in animal models it reduced infarct volume. [2] The authors concluded that disagreement may be due to bias or to animal models failing to mimic clinical disease adequately. [2] Those are six chosen examples, not a rule that animal work never predicts human results.

Common misunderstandings

• "It works in mice, so it works." An animal result describes animals.
• "It affects cells in a dish." That describes a dish, not a whole organism.
• "No human studies means it does not work." It means it has not been tested in humans. Absence of evidence is a gap, not a result.
• "Mechanism equals benefit." A proposed mechanism explains how something might work. It does not show that it does.

What this means in the Research Library

The library's publication records (6,705 with PubMed IDs and abstracts) carry a research model label that comes from the source data. About 2,717 are labeled animal experiments, 2,394 cell experiments, 1,360 animal and cell experiments and 116 clinical experiments, with a few mixed or other. In plain terms, most publication records in the library are not human studies. Search can filter by research model (human, animal, cell), so you can look at one kind at a time.

Labels come from source data and can be wrong or incomplete. The same is true elsewhere: PubMed's own species filters rely on indexing and can exclude relevant results. [8] When a label matters, open the abstract.

Continue with evidence types and study designs, why a registration is not a result, and relationship types in the knowledge graph.