Internal and external validity

Internal and external validity

Validity in clinical trials refers to the degree to which the results of a study are true, credible, and applicable. Two essential types of validity are:

• Internal validity: Are the observed effects in the study due to the intervention and not other factors?
• External validity (generalizability): Can the study findings be applied to people, settings, or times beyond those studied?

Both are critical to the scientific credibility and real-world impact of a trial.

Internal Validity

Internal validity refers to the degree to which the study design and conduct minimize bias, allowing for a trustworthy causal interpretation between the intervention and the outcome.

Threats to Internal Validity:

1. Selection bias – Unequal baseline characteristics if randomization fails
2. Performance bias – Differences in care apart from the intervention
3. Detection bias – Outcome assessment influenced by knowledge of treatment
4. Attrition bias – Incomplete outcome data or differential loss to follow-up
5. Confounding – Other variables influencing the outcome

Enhancing Internal Validity:

• Proper randomization and allocation concealment
• Use of blinding (participants, personnel, outcome assessors)
• Predefined analysis plans
• Handling missing data appropriately
• Monitoring and quality assurance procedures

External Validity

External validity, or generalizability, is the extent to which the results of a trial can be applied to other populations, settings, and times.

Factors Affecting External Validity:

1. Eligibility criteria – Overly restrictive inclusion/exclusion limits generalizability
2. Study setting – Single-center or highly specialized centers may not reflect typical care
3. Intervention delivery – May differ in real-world conditions (e.g., adherence, resources)
4. Follow-up and outcome measures – Frequency and methods may be impractical outside trials
5. Participant characteristics – Age, sex, comorbidities, socioeconomic factors

Enhancing External Validity:

• Use of broad eligibility criteria
• Multicenter and diverse geographic locations
• Inclusion of underrepresented populations
• Consider pragmatic trial design
• Use of real-world outcomes and standard care comparators

Balancing Internal and External Validity

There is often a trade-off:

• Highly controlled explanatory trials → High internal validity, low generalizability
• Pragmatic trials → Closer to real-world settings, but may compromise internal control

The appropriate balance depends on the study objective:

• Mechanistic or proof-of-concept trials prioritize internal validity
• Policy or practice-oriented trials emphasize external validity

Conclusion

Strong clinical trials must strive for high internal validity to establish credible evidence and sufficient external validity to ensure relevance and impact. Clear reporting of trial methods, settings, and populations helps readers assess both.

---

Bibliography

1. Shadish WR, Cook TD, Campbell DT. Experimental and Quasi-Experimental Designs for Generalized Causal Inference. 2nd ed. Houghton Mifflin; 2002. Foundational text on internal and external validity.
2. Rothwell PM. External validity of randomised controlled trials: “to whom do the results of this trial apply?” The Lancet. 2005;365(9453):82–93.
3. Hernán MA, Robins JM. Causal Inference: What If. Boca Raton: Chapman & Hall/CRC; 2020. Chapters on threats to internal validity and generalizability.
4. Kennedy-Martin T, Curtis S, Faries D, Robinson S, Johnston J. A literature review on the representativeness of randomized controlled trial samples and implications for the external validity of trial results. Trials. 2015;16:495.
5. Zwarenstein M, Treweek S, Gagnier JJ, et al. Improving the reporting of pragmatic trials: an extension of the CONSORT statement. BMJ. 2008;337:a2390.

---

Adapted for educational use. Please cite relevant trial methodology sources when using this material in research or teaching.