Population Attributable Risk

Population attributable risk measures how much of the disease occurring in a whole population can be attributed to a particular exposure - that is, the share of cases that would, in principle, not occur if the exposure were removed. Unlike measures of individual risk such as the relative risk, it combines the strength of an exposure's effect with how common the exposure is, making it central to setting public-health priorities for chronic disease.

Definition

The population attributable risk is the excess disease rate in the total population that is associated with an exposure, equal to the population disease rate minus the rate that would be expected if no one were exposed; expressed as a proportion of the total population rate it becomes the population attributable fraction, the share of cases attributable to the exposure under a causal assumption.

Scope

The entry covers the population attributable risk and the population attributable fraction (or proportion), how they depend jointly on relative risk and exposure prevalence, their interpretation and key assumptions, and the common errors in their use. It is a methodological topic and does not provide clinical guidance.

Core questions

• How does population attributable risk differ from relative risk and from attributable risk in the exposed?
• Why does the population attributable fraction depend on both the strength of effect and the prevalence of exposure?
• What causal and no-confounding assumptions are required to interpret an attributable fraction as preventable disease?
• Why can attributable fractions for several risk factors sum to more than 100%?

Key concepts

• Population attributable risk
• Population attributable fraction (proportion)
• Attributable risk in the exposed
• Exposure prevalence
• Relative risk dependence
• Preventable fraction
• No-confounding and causal assumptions
• Overlapping (non-additive) fractions

Mechanisms

Population attributable risk asks how much of a population's disease experience an exposure accounts for. Because it compares the actual population rate with the rate expected under no exposure, its size grows both with the relative risk of the exposure and with how prevalent that exposure is: a modest risk factor that is very common can contribute more population disease than a strong factor that is rare. Expressed as the population attributable fraction, the measure is often read as the proportion of disease that elimination of the exposure would prevent, but this interpretation holds only if the association is genuinely causal, free of confounding, and the exposure could be removed without otherwise changing risk. Because chronic diseases are multicausal, the same cases can be attributed to more than one component cause, so attributable fractions for different factors can overlap and sum to well above one hundred per cent.

Clinical relevance

Population attributable risk guides public-health resource allocation by indicating which exposures account for the largest share of disease in a population, complementing individual-level risk measures used in clinical care. This entry explains the measure at a population level for reference purposes and is not a basis for individual diagnostic or treatment decisions.

Epidemiology

Attributable-fraction reasoning underpins global priority-setting: the Global Burden of Disease project estimates the share of deaths and disability attributable to dozens of behavioural, metabolic, and environmental risk factors across populations. The approach also connects to Rose's argument that shifting the exposure distribution of a common factor across a whole population can prevent more disease than focusing only on high-risk individuals.

History

Morton Levin introduced the attributable-fraction idea in 1953 in the context of lung cancer and smoking, providing a way to express the population impact of an exposure. The measure became a staple of chronic-disease epidemiology and public-health planning, and Rockhill, Newman and Weinberg's 1998 review clarified its assumptions and the common ways it is misapplied, particularly the over-literal reading of attributable fractions as straightforwardly preventable disease.

Debates

Can the population attributable fraction be read as the proportion of disease that is preventable?

Interpreting an attributable fraction as preventable disease assumes the association is causal, unconfounded, and that removing the exposure leaves other risks unchanged; reviewers caution that these assumptions are often unmet and that overlapping fractions for multiple causes make simple preventability claims misleading.

Key figures

• Morton Levin
• Beverly Rockhill
• Geoffrey Rose
• Sander Greenland

Seminal works

• levin-1953
• rockhill-1998
• rose-1985

Frequently asked questions

How is population attributable risk different from relative risk?

Relative risk compares the risk in exposed and unexposed individuals, while population attributable risk measures how much of the disease in the whole population is due to the exposure. It depends on both the relative risk and how common the exposure is, so a weak but widespread factor can have a large attributable risk.

Why can attributable fractions for different risk factors add up to more than 100%?

Because chronic diseases are multicausal, a single case can be attributed to several component causes at once; each factor's attributable fraction counts those shared cases, so the fractions overlap and their sum can exceed one hundred per cent.