[Durgical](https://durgical.com/en/) › [DurgiPedia](https://durgical.com/en/durgipedia/) › [Epidemiology](https://durgical.com/en/epidemiology/) › [Bias, confounding and causation](https://durgical.com/en/epidemiology/areas/bias/) › Internal and external validity

# Internal and external validity

- Concept

Internal: whether the result is true for those who took part. External: whether it applies to other people, such as your patient.

Updated Nov 13, 2026

## Key points

- Threats to internal validity: selection bias, information bias, confounding, losses and chance.
- Threats to external validity: strict selection criteria, highly specialised centres, an intervention hard to reproduce, a different population.
- Without internal validity there is no external validity; the reverse can happen.
- The relative effect (RR) is usually more stable across populations than the absolute one; the NNT changes with the patient's baseline risk.
- Explanatory trials favour internal validity; pragmatic trials favour external validity.

On this page

- The basics Key points
- What it is
- Key facts
- Worked example
- Internal versus external validity
- Markers
- In practice
- Common pitfalls
- Memory aid

- Study Test yourself

- Connections Connections

## What it is

Internal validity is the degree to which a study's result reflects the truth in its own participants, free of bias, confounding and chance. External validity (generalisability) is the degree to which that result applies to other populations, places and times. Internal validity comes first: a biased result cannot be generalised. But a study that is flawless inside may still apply poorly to a particular patient.

## Key facts

- Applying to a patient: risk with treatment ≈ patient's baseline risk × the study RR; ARR = baseline risk − that risk; NNT = 1 / ARR.
- Applicability questions: would my patient have entered the study? Can the intervention be done here? Were the outcomes that matter measured?

## Worked example

- Made-up trial in a university hospital: concealed allocation, blinding and 3 % losses. Good internal validity.
- Result: event risk 20 % with control and 14 % with treatment. RR = 14 / 20 = 0.70; ARR = 6 points; NNT = 1 / 0.06 ≈ 17.
- External validity: only 8 % of participants were over 75; in the reader's practice they are 40 %.
- A 78-year-old patient with an estimated baseline risk of 40 %. If the RR of 0.70 holds: 40 × 0.70 = 28 %.
- ARR = 40 − 28 = 12 points; NNT = 1 / 0.12 ≈ 8.3, rounded up to 9.
- But with so few older people in the trial, it is not known whether the RR holds or whether harms are the same at that age.
- Reading: the study is sound inside, but applying it to this patient is an extrapolation; each kind of validity is judged on its own.

## Internal versus external validity

| Aspect | Internal validity | External validity |
| --- | --- | --- |
| Question | Is the result true in these participants? | Does it hold for other people and places? |
| Threats | Bias, confounding, losses, chance | Strict selection, specialised centre, different population |
| How it is protected | Randomisation, concealment, blinding, complete follow-up | Broad criteria, several sites, a realistic intervention |
| Who judges it | Critical appraisal of the methods | The clinician, against the patient and the setting |

## Markers

- IV
- EV
- RR
- NNT

## In practice

- Before applying a trial, compare its baseline table with the patient in front of you.
- In Colombia, results from high-complexity centres may not be reproduced in rural areas with other resources.

## Common pitfalls

- Believing that a large sample guarantees external validity.
- Rejecting every study whose population is not identical to one's own: what matters is whether there are reasons for the effect to differ.
- Applying the study NNT without adjusting for the patient's baseline risk.

## Memory aid

‘True first, useful second’: internal validity decides whether to believe the study; external validity, whether it helps your patient.

## Test yourself

Think of the answer, then tap the question to see it.

## What is internal validity?

The degree to which the result reflects the truth in the participants themselves, without important bias, confounding or random error.

## Can a study have external validity without internal validity?

No: a biased result cannot be generalised.

## A trial shows RR 0.80; your patient's baseline risk is 25 %. What is the approximate NNT?

25 × 0.80 = 20 %; ARR = 5 points; NNT = 1 / 0.05 = 20.

## Which type of trial puts external validity first?

The pragmatic trial, with broad criteria and conditions close to everyday practice.

[Add to my cards (4)](https://durgical.app/estudiar/?add=/en/epidemiology/internal-and-external-validity/)

## Connections

## Related 6

- [Selection bias](https://durgical.com/en/epidemiology/selection-bias/)
- [Confounding](https://durgical.com/en/epidemiology/confounding/)
- [Information (measurement) bias](https://durgical.com/en/epidemiology/information-bias/)
- [Critical appraisal of a paper](https://durgical.com/en/epidemiology/critical-appraisal-of-a-paper/)
- [Pragmatic and explanatory trials](https://durgical.com/en/epidemiology/pragmatic-and-explanatory-trials/)
- [Number needed to treat](https://durgical.com/en/epidemiology/number-needed-to-treat/)

[← All of epidemiology](https://durgical.com/en/epidemiology/)

Educational content on epidemiological and statistical methods; the examples use made-up numbers to teach the calculation. Not a clinical or legal guide: current Colombian rules are in the official sources.

Sources: [Chapter 8: Assessing risk of bias in a randomized trial](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-08) (Cochrane, 2024); [Chapter 14: Completing ‘Summary of findings’ tables and grading the certainty of the evidence](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-14) (Cochrane, 2026).

## **Spotted a mistake?** We are always improving for the community. Tell us what is wrong and our team will review it. Report a mistake

Representative model, for educational purposes only.

---
Source: https://durgical.com/en/epidemiology/internal-and-external-validity/
License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Cite as: DurgiPedia (durgical.com), https://durgical.com/en/epidemiology/internal-and-external-validity/