In the fields of medicine and public health, we often ask the question “What part of Problem X is caused by Factor Y.” Specifically, we want to know the “fraction” of a problem that is caused by each of the possible (or theoretical) causal factors. Or simply, who or what is to blame, and by how much? For example, what proportion of lung cancer is attributable to smoking? What proportion of cancers are attributable to environmental exposure factors versus genetic or dietary factors? From a public health and medical perspective, these are important questions, as they help us figure out ways in which we might help prevent various diseases and conditions by identifying the true causes.
This concept is referred to as “attributable risk” or “attributable fraction.”[1] As a method, it is one of the most valuable tools we have in public health and health economics. It offers a way to disentangle the seemingly intractable problem of determining the magnitude of each factor. There are number of ways in which these tools can be used, including: (1) as part of a statistical model, such as a regression analysis, which tells us the role of one variable while holding constant (or “controlling for”) all the other potential variables; or (2) as a means of allocating “responsibility” using an accounting approach, perhaps based on sales data. These two approaches can also be combined; for example, we might use the regression approach to determine the role of asbestos exposure in lung cancer,[2] and then use the accounting approach to allocate abatement responsibility among asbestos manufacturers. These methods are commonly employed by epidemiologists, biostatisticians, health economists, and environmental economists in a wide variety of applications, including, for example, cancer attributable risk,[3] environmental exposure,[4] substance use disorder,[5] tobacco use,[6] and obesity.[7]
These tools have direct application in litigation, and we and other health economists have successfully applied them in a variety of litigation settings. For example, in product liability cases, we might be interested “quantifying” the role (if any) of a defendant in contributing to some sort of outcome, or some impact resulting in damages or requiring abatement. In a personal injury context, these tools can help determine how certain pre-existing conditions might result in reduced life expectancy. In a contract breach context, these tools can be used to determine the impact of an external factor (i.e., the hiring of a new vendor, or the exodus of key employees) on business performance. There are many other examples. In sum, these tools, if applied correctly, are powerful tools and are very “reliable” due to their common use in public health, medicine, and health economics.
(By John E. Schneider, PhD) (2023)
1) See generally O. Gefeller, “Definitions of attributable risk–revisited,” Public Health Rev 23, no. 4 (1995); S. G. Heeringa et al., “Attributable fraction estimation from complex sample survey data,” Ann Epidemiol 25, no. 3 (2015); C. Poole, “A history of the population attributable fraction and related measures,” ibid.
2) See, for example, E. K. Moon et al., “Variations of lung cancer risk from asbestos exposure: impact on estimation of population attributable fraction,” Ind Health 51, no. 1 (2013); M. P. Purdue et al., “The proportion of cancer attributable to occupational exposures,” Ann Epidemiol 25, no. 3 (2015); D. C. Whiteman and L. F. Wilson, “The fractions of cancer attributable to modifiable factors: A global review,” Cancer Epidemiol 44 (2016).
3) See, for example, “The fractions of cancer attributable to modifiable factors: A global review.”
4) See, for example, S. De Matteis, D. Consonni, and P. A. Bertazzi, “Exposure to occupational carcinogens and lung cancer risk. Evolution of epidemiological estimates of attributable fraction,” Acta Biomed 79 Suppl 1 (2008); Moon et al., “Variations of lung cancer risk from asbestos exposure: impact on estimation of population attributable fraction.”; Purdue et al., “The proportion of cancer attributable to occupational exposures.”; K. Torén and P. D. Blanc, “Asthma caused by occupational exposures is common – a systematic analysis of estimates of the population-attributable fraction,” BMC Pulm Med 9 (2009).
5) See, for example, R. N. Hansen et al., “Economic costs of nonmedical use of prescription opioids,” Clin J Pain 27, no. 3 (2011); J. Rehm et al., “The relationship between different dimensions of alcohol use and the burden of disease-an update,” Addiction 112, no. 6 (2017); J. Rehm and S. Imtiaz, “A narrative review of alcohol consumption as a risk factor for global burden of disease,” Subst Abuse Treat Prev Policy 11, no. 1 (2016); J. Rehm, K. D. Shield, and E. Weiderpass, “Alcohol consumption. A leading risk factor for cancer,” Chem Biol Interact 331 (2020); H. A. Whiteford et al., “Global burden of disease attributable to mental and substance use disorders: findings from the Global Burden of Disease Study 2010,” Lancet 382, no. 9904 (2013).
6) See, for example, M. Pérez-Ríos and A. Montes, “Methodologies used to estimate tobacco-attributable mortality: a review,” BMC Public Health 8 (2008).
7) See, for example, K. I. Avgerinos et al., “Obesity and cancer risk: Emerging biological mechanisms and perspectives,” Metabolism 92 (2019).