Science
Preclinical models: what they can and cannot show
A guide to reading cellular and animal data without overinterpretation.
A preclinical model isolates a question in a controlled system. Cell lines can examine signalling, while animal models can explore interactions…
02Internal and external validityGood randomisation, appropriate controls and blinded assessment strengthen internal validity. External validity concerns how far a result travels beyond the model.…
03Translation is not automaticA positive preclinical finding is not a human recommendation, and a negative finding does not exclude every other condition. Responsible synthesis…
The value of a model
A preclinical model isolates a question in a controlled system. Cell lines can examine signalling, while animal models can explore interactions between organs and responses over time. Their value lies in experimental clarity and hypothesis generation.
Internal and external validity
Good randomisation, appropriate controls and blinded assessment strengthen internal validity. External validity concerns how far a result travels beyond the model. Differences in species, biological background and experimental conditions limit generalisation.
Translation is not automatic
A positive preclinical finding is not a human recommendation, and a negative finding does not exclude every other condition. Responsible synthesis states the model, endpoint and uncertainty, and looks for convergence from independent methods before proposing a broader mechanism.
LABORATORY APPLICATION
Turning the principle into an auditable record
Before using a preclinical source, record species or cell line, biological background, sample size, comparator, primary endpoint and exclusions. Then write one sentence stating what the model supports and a separate sentence stating what it cannot support. This limits over-generalisation.
PRIMARY & OFFICIAL SOURCES
Sources and further reading
COMMON QUESTIONS
Questions specific to this topic
Why use simple models?
They allow control of variables and clear testing of a mechanism, but do not represent the whole biology.
What is external validity?
It is the extent to which a result can reasonably travel beyond the specific model and its conditions.
Is one positive preclinical result enough?
No. Replication, suitable controls, transparency and convergence from complementary methods are needed.