9.3.9 Cause and Effect Relationship

Describes Approaches for Arguing a Cause and Effect Relationship (Koch, Hill, Rothman, Susser)

The establishment of a causal relationship between exposure and an outcome is a fundamental objective in research. Several approaches have been proposed to determine causality, including Koch’s postulates, Hill’s criteria, Rothman’s causal pie model, Susser’s causal web and the triad of essential causal properties.

Koch’s postulates:

Robert Koch, a German physician, developed a set of postulates to establish a causal relationship between a microorganism and a disease. The postulates include:

• the organism must be present in every case of the disease
• the organism must be isolated and grown in pure culture
• the disease must be reproduced when the cultured organism is inoculated into a healthy host
• the same organism must be isolated again from the experimentally infected host.

Although Koch’s postulates have limitations, they provide a useful framework for establishing a causal relationship.

Hill’s criteria:

Sir Austin Bradford Hill, a British epidemiologist, proposed nine criteria to assess a causal relationship between an exposure and an outcome. These criteria provide a systematic approach to evaluating the evidence for causality. These criteria are not definitive rules but serve as a guide to evaluate the evidence:

• Strength of association: The larger the association, the more likely it is causal.
• Consistency: The association is observed in different studies and populations.
• Specificity: The exposure is associated with a specific outcome rather than multiple outcomes.
• Temporality: The exposure must precede the outcome.
• Biological gradient: There is a dose-response relationship between the exposure and the outcome.
• Plausibility: There is a biologically plausible explanation for the association.
• Coherence: The association is consistent with existing knowledge and theories.
• Experiment: Intervention studies support the causal relationship.
• Analogy: Similar associations have been observed for similar exposures and outcomes.

Rothman’s causal pie model:

Kenneth Rothman, an American epidemiologist, proposed the sufficient-component cause model, which considers causality as a complex interaction of multiple factors. In this model, a cause is seen as a combination of component causes that are individually insufficient but, when combined, are sufficient to produce the outcome. This approach emphasizes the importance of considering multiple risk factors and their interactions when assessing causality.

Susser’s causal web:

Mervyn Susser, a South African epidemiologist, proposed four criteria for causality, which focus on the research design and analysis:

• Temporality: The exposure must precede the outcome.
• Strength and consistency of association: The association should be strong and consistent across studies.
• Dose-response relationship: There should be a gradient in the risk of the outcome with increasing exposure levels.
• Alternative explanations: Other possible explanations for the observed association must be considered and ruled out.

Susser’s triad of essential causal properties:

Susser’s triad of essential causal properties is a fundamental concept in epidemiology that helps determine whether an observed association between an exposure and an outcome is likely to be causal. The triad includes three essential properties:

• association
• time order
• directionality.

Association refers to the presence of an observed relationship between the exposure and the outcome, while time order ensures that the exposure precedes the outcome. Directionality requires that there be a plausible mechanism linking the exposure to the outcome. This means that there should be a logical and biologically plausible explanation for how the exposure causes the outcome. By applying Susser’s triad, epidemiologists can evaluate the likelihood that an observed association between an exposure and an outcome is causal, which is critical for identifying potential causes of disease and developing effective interventions to prevent or mitigate their impact.

Approach	Description
Koch’s postulates	A set of criteria is used to determine if a specific microorganism is the cause of a particular disease. The criteria include that the microorganism should be present in all cases of the disease, should be isolated from the host and grown in pure culture, should cause the same disease when inoculated into a susceptible host, and should be re-isolated from the host.
Hill’s criteria	A set of guidelines is used to assess the causal relationship between a risk factor and a disease. The criteria include the strength of association, consistency, specificity, temporality, biological gradient, plausibility, coherence, experiment, and analogy.
Rothman’s causal pie model	A conceptual model that defines causation as a combination of component causes that act together to cause a disease or outcome. The model considers three components: the necessary cause, the sufficient cause, and the component causes.
Susser’s causal web model	A model that considers the causal relationships between multiple risk factors and disease. The model takes into account the multiple pathways through which risk factors can contribute to disease and the complex interactions between risk factors. The model also considers the importance of time in determining causal relationships.
Susser’s triad of essential causal properties	The triad includes three essential properties: association, time order, and directionality.

Each of these approaches contributes to the understanding of cause-and-effect relationships in public health and epidemiology. They emphasize different aspects of the causal inference process:

• Koch’s Postulates are more suited to infectious diseases.
• Hill’s Criteria offer a broader, more flexible guideline applicable to a variety of health outcomes.
• Rothman’s Causal Pies highlight the multifactorial and often complex nature of causation in diseases.
• Susser’s Framework underscores the importance of understanding both surface-level associations and deeper underlying mechanisms.

In modern epidemiology, these concepts are often used in combination or adapted to suit specific research questions and contexts, acknowledging that different diseases and health outcomes may require different approaches to establish causation.

References:

1. Koch, R. (1891). Über bakteriologische Forschung. Verhandlungen des X. Internationalen Medizinischen Congresses, Berlin, 35-45.
2. Hill, A. B. (1965). The environment and disease: association or causation? Proceedings of the Royal Society of Medicine, 58(5), 295-300.
3. Rothman, K. J. (1976). Causes. American Journal of Epidemiology, 104(6), 587-592.
4. Susser, M. (1973). Causal thinking in the health sciences: concepts and strategies of epidemiology. Oxford University Press.