Risk Factors and Causation
Risk factors and causation is the area of chronic-disease epidemiology that asks why non-communicable diseases occur: it identifies the exposures, behaviours, and host characteristics associated with raised disease risk and develops the conceptual and quantitative tools for judging when such associations reflect cause. Because chronic diseases typically arise from many interacting causes acting over long periods, this area emphasises multicausal models, graded exposure-response, and population-level attribution rather than single necessary causes.
Definition
Risk factors and causation comprises the identification of exposures statistically associated with chronic-disease occurrence and the framework of criteria, models, and measures used to evaluate whether and to what extent those exposures are causal at the individual and population level.
Scope
The area orients the reader across the topics that make up causal reasoning for chronic disease: how exposures and risk factors are defined and classified, how diseases unfold over time (natural history and progression), how multiple causes combine (multifactorial etiology), how risk changes with the level of exposure (dose-response), and how much disease in a population is attributable to a given factor (population attributable risk). It treats causation as a methodological subject within epidemiology and not as clinical guidance.
Sub-topics
Core questions
- What distinguishes a risk factor that is merely associated with disease from one that is causal?
- How do multiple causes combine to produce chronic disease, and what does it mean for a cause to be a component of a sufficient cause?
- How does disease risk change as the level or duration of an exposure increases?
- How much of the disease burden in a population could in principle be removed by eliminating a given exposure?
Key concepts
- Risk factor and exposure
- Association versus causation
- Necessary and sufficient causes
- Component causes and interaction
- Biological gradient (dose-response)
- Latency and induction period
- Population attributable fraction
- Modifiable versus non-modifiable risk factors
Key theories
- Sufficient-component cause model (causal pies)
- Rothman's model represents each sufficient cause as a set (a 'pie') of component causes that together produce disease; a single component is rarely sufficient or necessary alone, which formalises the multicausal nature of chronic disease and explains interaction between factors.
- Bradford Hill viewpoints on causation
- Hill set out nine considerations - including strength, consistency, biological gradient (dose-response), temporality, and plausibility - as aids to judgement when deciding whether an observed association is causal; they are heuristics for inference, not a checklist or a statistical test.
- Population versus individual causes of disease
- Rose distinguished the causes of cases within a population from the causes of the population's overall incidence, showing that the determinants of who becomes ill can differ from the determinants of how common a disease is, with implications for prevention strategy.
Mechanisms
Causal reasoning in this area proceeds from observed associations to causal judgement using explicit models and criteria. Associations are first checked against alternative explanations - chance, bias, and confounding - and then weighed using considerations such as temporality, strength, consistency, and biological gradient. The sufficient-component cause model clarifies that most chronic-disease risk factors are component causes that act only in combination with other factors, so the same outcome can arise through several distinct causal constellations and the apparent effect of one factor depends on the prevalence of its complementary causes. Graded exposure-response relationships strengthen causal inference and, together with attributable-fraction measures, translate individual-level risk into the share of population disease that a factor explains.
Clinical relevance
The risk factors established through this area underpin much of preventive medicine and clinical risk assessment, since recognising modifiable determinants is what makes chronic disease preventable in principle. The material here describes how causal knowledge is generated and quantified at the population level; it is reference and educational content and is not a basis for individual diagnostic or treatment decisions.
Epidemiology
Large prospective cohorts such as the Framingham studies provided the empirical engine for chronic-disease risk-factor epidemiology, identifying blood pressure, smoking, and lipids as determinants of cardiovascular disease. At the global scale, the Global Burden of Disease project quantifies how much death and disability is attributable to dozens of behavioural, metabolic, and environmental risk factors across populations and over time.
History
Twentieth-century chronic-disease epidemiology shifted attention from single infectious agents to webs of interacting determinants. Hill's 1965 address codified considerations for causal judgement, Rothman's 1976 paper supplied a formal multicausal model, and Rose's 1985 essay reframed causation at the population level. Long-running cohorts such as Framingham, begun in 1948, demonstrated empirically that chronic diseases have measurable, modifiable risk factors, and the Global Burden of Disease enterprise later systematised population attribution.
Debates
- Are the Bradford Hill considerations criteria or heuristics?
- Hill himself framed his nine viewpoints as aids to judgement rather than mandatory rules, and there is continuing discussion about whether they should be applied as a formal checklist, weighted, or replaced by explicit causal-model reasoning.
- Should prevention target high-risk individuals or whole populations?
- Rose argued that shifting the population distribution of a risk factor can prevent more disease than targeting only high-risk individuals, a strategic tension that remains central to chronic-disease policy.
Key figures
- Austin Bradford Hill
- Kenneth Rothman
- Geoffrey Rose
- William Kannel
- Sander Greenland
Related topics
Seminal works
- hill-1965
- rothman-1976
- rose-1985
- kannel-1979
Frequently asked questions
- What is the difference between a risk factor and a cause?
- A risk factor is an exposure or characteristic statistically associated with higher disease risk; calling it a cause additionally requires judging that the association is not due to chance, bias, or confounding and that altering the factor would alter risk. Many risk factors are component causes that act only together with others.
- Why is chronic disease described as multifactorial?
- Most non-communicable diseases result from the interaction of several genetic, behavioural, metabolic, and environmental factors over long periods, so no single exposure is usually necessary or sufficient on its own to produce the disease.