Evidence-Based Approaches to Public Health: Epidemiology – Causality: Hill’s Criteria for Causation

In this tutorial, we will explore Hill’s criteria for causation, a set of principles that help researchers evaluate whether an observed association between an exposure and an outcome is likely to be causal. Hill’s criteria are widely used in epidemiology to assess causality, and understanding them is essential for the Certified in Public Health (CPH) exam.

By the end of this tutorial, you will understand each of Hill’s criteria, how they apply to evaluating causal relationships, and the limitations of these criteria. Practice questions will be provided to reinforce your understanding.

Table of Contents:

1. Introduction to Causality in Epidemiology
2. Hill’s Criteria for Causation
   • Strength of Association
   • Consistency
   • Specificity
   • Temporality
   • Biological Gradient (Dose-Response)
   • Plausibility
   • Coherence
   • Experiment
   • Analogy
3. Practice Questions
4. Conclusion

1. Introduction to Causality in Epidemiology

In epidemiology, causality refers to the relationship between an exposure (e.g., smoking) and an outcome (e.g., lung cancer) where the exposure directly influences the likelihood of the outcome. However, establishing a causal relationship is challenging because most epidemiological studies are observational, and multiple factors can influence the association between exposure and outcome.

To help determine whether an observed association is likely to be causal, Sir Austin Bradford Hill proposed nine criteria in 1965. These Hill’s criteria for causation are used to assess the evidence and help decide if a causal relationship exists between an exposure and an outcome.

2. Hill’s Criteria for Causation

Hill’s criteria for causation consist of nine principles that should be considered when evaluating an association between exposure and outcome. While not all criteria must be met to establish causality, they provide a framework for assessing the strength and validity of the evidence.

2.1 Strength of Association

Strength of association refers to the magnitude of the relationship between exposure and outcome. A strong association is more likely to be causal than a weak one. The strength of the association is often measured using relative risk (RR) or odds ratio (OR).

• Example: The relative risk of lung cancer among smokers compared to non-smokers is very high, providing strong evidence that smoking is causally related to lung cancer.

2.2 Consistency

Consistency refers to whether the association between exposure and outcome is observed consistently across different studies, populations, and settings. If multiple studies show the same result, this strengthens the case for causality.

• Example: Numerous studies in different countries have found a consistent association between smoking and lung cancer, further supporting a causal relationship.

2.3 Specificity

Specificity refers to the idea that a specific exposure should lead to a specific outcome. While this criterion is not always met (especially in complex diseases), it can strengthen the case for causality when observed.

• Example: Smoking causes lung cancer, but it is also associated with other diseases like heart disease. In some cases, one exposure may lead to multiple outcomes.

2.4 Temporality

Temporality refers to the time relationship between the exposure and the outcome. For an exposure to cause an outcome, it must occur before the outcome. Temporality is the only criterion that must always be satisfied to establish causality.

• Example: Smoking must occur before the development of lung cancer for smoking to be considered a cause of the disease.

2.5 Biological Gradient (Dose-Response)

The biological gradient (also known as the dose-response relationship) refers to the observation that an increase in exposure level should result in an increase in the risk of the outcome. This provides further evidence of a causal relationship.

• Example: Heavier smokers are at a greater risk of developing lung cancer compared to light smokers, indicating a dose-response relationship.

2.6 Plausibility

Plausibility refers to whether there is a biologically plausible mechanism that explains how the exposure leads to the outcome. This criterion depends on current scientific knowledge, which may evolve over time.

• Example: There is a plausible biological mechanism by which the carcinogens in cigarette smoke cause damage to lung cells, leading to cancer.

2.7 Coherence

Coherence means that the observed association should not conflict with known facts about the natural history and biology of the disease. The association should align with the existing body of evidence.

• Example: The association between smoking and lung cancer is coherent with what is known about the biological effects of smoking on the respiratory system.

2.8 Experiment

Experiment refers to evidence from experiments, such as randomized controlled trials or natural experiments, that support the causal relationship. While experimental evidence is not always available, it can provide the strongest evidence for causality.

• Example: In smoking cessation trials, people who stop smoking have lower rates of lung cancer than those who continue smoking, providing experimental support for a causal relationship.

2.9 Analogy

Analogy refers to the idea that if a similar exposure-outcome relationship has been observed for a different exposure, it strengthens the case for causality. While this criterion is the weakest, it can provide additional support.

• Example: Since certain chemicals are known to cause cancer, it is plausible that other similar chemicals may also be carcinogenic.

3. Practice Questions

Test your understanding of Hill’s criteria for causation with the following practice questions. Try answering them before checking the solutions.

Question 1:

A study shows that the risk of heart disease increases as the number of cigarettes smoked per day increases. Which of Hill’s criteria does this evidence support?

Answer 1:

Question 2:

A researcher finds that studies conducted in different countries all show a strong association between high cholesterol and heart disease. Which of Hill’s criteria does this support?

Answer 2:

Question 3:

For a causal relationship to exist, the exposure must occur before the outcome. Which of Hill’s criteria does this describe?

Answer 3:

4. Conclusion

Hill’s criteria for causation provide a valuable framework for evaluating whether an observed association between an exposure and an outcome is likely to be causal. While not all criteria need to be met for causality to be established, they help researchers assess the strength of the evidence and guide public health decision-making.

Remember:

• Strength of association and consistency strengthen the case for causality.
• Temporality is the only criterion that must always be met to establish a causal relationship.
• Plausibility and coherence provide biological and logical support for causality.
• Experiment and biological gradient offer direct and dose-response evidence for causal relationships.

Final Tip for the CPH Exam:

Make sure you understand how each of Hill’s criteria applies to evaluating causality and practice applying these criteria to real-world examples of exposure-outcome relationships. This knowledge will be crucial for answering questions related to causality on the Certified in Public Health (CPH) exam.

Humanities Moment

The featured image for this CPH Focus is Huffcap Pear (1811) by Thomas Andrew Knight (English, 1759–1838). Knight was a British horticulturalist and botanist who served as the second President of the Royal Horticultural Society (1811–1838). Born at Wormesley Grange in Herefordshire, he conducted pioneering research on plant physiology and fruit tree propagation, using his inherited estate to breed crops such as strawberries, cabbages, and peas. Knight’s influential works, including his Treatise on the Culture of the Apple and Pear (1797), established him as a leading figure in 18th and 19th-century horticulture, although his personal papers were lost after his death.