Attributable risk

In clinical and epidemiological studies, the attributable risk is the risk that can be ascribed to exposure to a risk factor . In contrast, the relative risk compares the disease risks of exposed and non-exposed people.

The attributable risk indicates the percentage by which a disease frequency can be reduced if the risk factor were switched off. For example, the risk of developing lung cancer is higher for smokers than for non-smokers - but non-smokers also have fewer cases of lung cancer.

In contrast to the relative risk, the attributable risk takes into account the rarity or frequency of an illness. To illustrate, a hypothetical example - two behaviors a person could choose between:

Behavior A doubles the risk of lung cancer and halves the risk of oral cancer .

Behavior B halves the risk of lung cancer and makes oral cancer twice as common.

A poorly informed person would choose behavior at random or even agree to behavior A, because the relative risk is the same for both diseases - namely, doubling or halving the risk of the disease. Oral cancer occurs much less often (around 10,000 diseases / year in Germany; lung cancer: 50,000). The probability of suffering one of the two diseases is lower with behavior B (i.e. oral cancer: 20,000 cases, lung cancer 25,000, a total of 45,000 cases per year).

The attributable risk of a population corresponds to the attributable risk multiplied by the population.

	Number of people with risk factor	Number of people without a risk factor
Number of sick people	${\ displaystyle a}$	${\ displaystyle b}$
Number of not sick people	${\ displaystyle c}$	${\ displaystyle d}$

Risk of people who were exposed to the risk factor (corresponds to the positive predictive value (PPW)):

{\ displaystyle RE = PPW = {\ frac {a} {a + c}}}

Risk of persons who were not exposed to the risk factor (corresponds to the negative predictive value (NPW)):

{\ displaystyle RNE = NPW = {\ frac {b} {b + d}}}

The attributable risk then corresponds to the difference:

{\ displaystyle AR = RE-RNE = PPW-NPW = {\ frac {a} {a + c}} - {\ frac {b} {b + d}}}

The relative risk (RR), on the other hand, is calculated as a quotient. The term attributable risk is, however, inaccurately defined in the literature, because there is also the term risk differences z ( RD ). Attributable risk and risk difference are sometimes equated. The attributable fraction and the population-related attributable risk , the percentage difference between an exposed and a non-exposed population, must also be distinguished from the risk difference .

An example with fictitious data

Suppose you want to investigate the relationship between the occurrence of heart attacks and smoking. You watch 10,000 patients and find out whether they smoke or not and whether they have ever had a heart attack. The result is the following crosstab:

	Number of people who smoke	Number of people who do not smoke
Number of people with a heart attack	130	70
Number of people without a heart attack	1870	7930

The following attributable risk results:

{\ displaystyle AR = {\ frac {130} {130 + 1870}} - {\ frac {70} {70 + 7930}} = 0 {,} 05625}

By multiplying by the factor 100, a percentage value can now be formed from the probability value. The risk of suffering a heart attack is therefore 5.625 percentage points lower for a non-smoker than for a smoker (less than 0.9% for a non-smoker versus 6.5% for a smoker).

Web links

Tutorial on the attributable risk in epidemiology (Engl.) ( Memento of 23 April 2009 at the Internet Archive )

Individual evidence

↑ ^a ^b Medical Psychology and Sociology . Georg Thieme Verlag, 2010, ISBN 978-3-13-136422-7 , p. 245 ( limited preview in Google Book Search).
↑ ^a ^b Arterial hypertension . Springer-Verlag, 2013, ISBN 978-3-642-18507-6 , pp. 52 ( limited preview in Google Book search).
↑ Preventive medicine, epidemiology and social medicine: for medicine and dentistry . facultas.wuv Universitäts, 2007, ISBN 978-3-7089-0094-0 , p. 44 ( limited preview in Google Book search).
^ Hermann Faller, Hermann Lang: Medical Psychology and Sociology . Springer-Verlag, 2011, ISBN 978-3-642-12584-3 , p. 88 ( limited preview in Google Book search).

[:0-1] Medical Psychology and Sociology . Georg Thieme Verlag, 2010, ISBN 978-3-13-136422-7 , p. 245 ( limited preview in Google Book Search).

[:1-2] Arterial hypertension . Springer-Verlag, 2013, ISBN 978-3-642-18507-6 , pp. 52 ( limited preview in Google Book search).

[3] Preventive medicine, epidemiology and social medicine: for medicine and dentistry . facultas.wuv Universitäts, 2007, ISBN 978-3-7089-0094-0 , p. 44 ( limited preview in Google Book search).

[4] Hermann Faller, Hermann Lang: Medical Psychology and Sociology . Springer-Verlag, 2011, ISBN 978-3-642-12584-3 , p. 88 ( limited preview in Google Book search).