Attributable Burden and Population Impact

It is one thing to know how many people fall ill or die from a disease, and another to know how much of that burden a particular cause is responsible for. Attributable burden answers the second question: it estimates the share of disease, death, or DALYs that would be avoided if a given infection, pathogen, or exposure were removed. The population attributable fraction extends this from individual risk to the whole population, accounting for how common the cause is.

Definition

Attributable burden is the portion of disease, death, or health loss in a population that can be ascribed to a particular cause; the population attributable fraction is the proportion of total cases that would not have occurred if the causal exposure had been absent, given a causal relationship.

Scope

The topic covers attributable and population-attributable measures - the attributable fraction, the population attributable fraction, and attributable burden expressed in deaths or DALYs - their interpretation, and the strong causal assumptions they require. It is a reference to how the impact of a cause is quantified and apportioned, not a basis for clinical action.

Core questions

• What share of an outcome would be avoided if a given infection or exposure were eliminated?
• How does the population attributable fraction differ from individual relative risk?
• Why do attributable fractions depend on both the strength and the prevalence of a cause?
• What causal assumptions must hold for attributable burden to be interpretable?

Key concepts

• Attributable fraction (among the exposed)
• Population attributable fraction (PAF)
• Exposure prevalence
• Counterfactual / theoretical-minimum exposure
• Comparative risk assessment
• Causal assumption
• Attributable deaths and DALYs

Mechanisms

Attributable measures combine the strength of an association with how widespread the cause is. The attributable fraction among the exposed reflects how much of their risk is due to the exposure, derived from the relative risk. The population attributable fraction scales this by the prevalence of the exposure in the population, so a modest risk factor that is very common can account for more burden than a strong but rare one. Multiplying the fraction by the total deaths or DALYs from an outcome gives the attributable burden. These quantities are counterfactual: they describe what would happen under a comparison scenario in which the cause is removed, and they are only valid if the association is genuinely causal and confounding is controlled - assumptions that are easily violated and frequently misinterpreted (Rockhill, Newman & Weinberg, 1998; Rothman, Greenland & Lash, 2008).

Clinical relevance

Attributable burden quantifies the population-level contribution of an infection or exposure and informs where prevention could yield the largest aggregate gain; it is an apportioning tool for populations and carries no implication for individual diagnosis or treatment.

Epidemiology

Comparative risk assessment within global burden studies estimates attributable deaths and DALYs for many risk factors and causes by comparing observed exposure with a theoretical-minimum-risk scenario, allowing the population impact of competing causes to be ranked (Murray & Lopez, 2013; Vos et al., 2020).

Evidence & guidelines

Methodological literature cautions that attributable fractions are routinely miscomputed and misread, especially when summed across non-independent causes (Rockhill, Newman & Weinberg, 1998), and core texts set out their derivation and causal preconditions (Rothman, Greenland & Lash, 2008). The Global Burden of Disease comparative risk framework operationalises attribution at scale (Murray & Lopez, 2013; Vos et al., 2020).

History

The attributable fraction emerged from mid-twentieth-century risk-factor epidemiology as a way to express the public-health relevance of an exposure, and the population attributable fraction generalised it to whole populations. Comparative risk assessment later embedded attribution into global burden estimation, while methodological critiques highlighted persistent misuse (Rockhill, Newman & Weinberg, 1998; Murray & Lopez, 2013).

Debates

Can attributable fractions for multiple causes be added together?

Because causes can act jointly and overlap, population attributable fractions for different exposures need not sum to one and can exceed it; treating them as an additive partition of total burden is a common and consequential error.

Key figures

• Beverly Rockhill
• Kenneth J. Rothman
• Sander Greenland
• Christopher J. L. Murray

Seminal works

• rockhill-1998
• murray-2013

Frequently asked questions

How does the population attributable fraction differ from relative risk?

Relative risk measures how much more likely an outcome is among the exposed. The population attributable fraction also accounts for how common the exposure is, so it expresses the share of total disease in the whole population that the cause is responsible for.

Why can a weak risk factor have a large attributable burden?

Because attributable burden depends on prevalence as well as strength. A factor that raises risk only modestly but affects a large fraction of the population can account for more cases than a strong factor that is rare.