Estimating the contribution of individual risk factors to disease in a person with more than one risk factor

Our plain-English summary of Estimating the contribution of individual risk factors to disease in a person with more than one risk factor, by P. McElduff, J. Attia, B. Ewald, et al (Journal of Clinical Epidemiology, 2002).

The study

• Authors: P. McElduff, J. Attia, B. Ewald, J. Cockburn, R. Heller
• Journal: Journal of Clinical Epidemiology
• Published: 2002;55(6):588-92
• Cited by: 21 articles, as counted by NIH iCite
• Read it: PubMed · Publisher

Abstract

Various measures have been proposed to express the excess risk of an outcome attributable to one particular risk factor, such as attributable risk and risk fraction. However, there is sometimes a need, both in epidemiological studies and in awarding compensation in legal cases, to simultaneously consider the contribution of several risk factors to a disease outcome, when a biological model is not available. We propose a method that allocates the proportional contribution of several risk factors to a disease outcome, based on the weighted contribution of the risk fraction for each risk factor. We demonstrate the use of this method using figures for renal cell carcinoma, and discuss the caveats in using this method for epidemiologic studies, and in awarding compensation in legal cases.

In plain English

This was a methods paper, not a study of patients. The authors proposed a way to split a disease's excess risk among several risk factors in one person. They illustrated it with an Australian appeal case about whether an employee's workplace asbestos exposure had contributed to renal-cell carcinoma. The employee was also male, obese and a smoker. The authors used relative risks from published studies and assumed the factors acted independently. On that basis the method gave asbestos, sex and smoking 9.5% each (each with a relative risk of 1.4) and obesity 61.5% (relative risk 3.6). In a hypothetical normal-weight woman, two factors got 24.5% each. The text calls them smoking and asbestos, but the table labels them sex and asbestos. The authors presented the method as a simple tool to help attribute cause in individuals. They cautioned that odds ratios can replace relative risks only under certain conditions. They also said unavoidable factors such as gender should not count legally, and that the risk fraction represents only a minimum bound for the probability of causation.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Methods paper proposing a statistical way to apportion excess risk among several risk factors in one individual. It was illustrated with a real legal case and hypothetical examples, and no new data were collected
• Who was studied: One employee in an Australian appeal court case who developed renal-cell carcinoma and had four known risk factors (male sex, obesity, cigarette smoking, workplace asbestos exposure). A hypothetical normal-weight woman and hypothetical risk-factor groups were also used
• Exposure: Workplace asbestos exposure (fiber type not stated), considered alongside cigarette smoking, obesity and male sex. The relative risk used for asbestos was 1.4, taken from cited published studies
• Main results:
  • In a hypothetical example (risk factor A relative risk 2, risk factor B relative risk 4.5, both together 9), the method allocated 19.8% of the contribution to A and 69.1% to B. The remaining 11.1% went to baseline risk and unknown factors
  • For the man in the court case, assuming independent risk factors, the proportional contribution of asbestos exposure to renal-cell carcinoma was 9.5%. Sex and smoking were also allocated 9.5% each
  • In the same man, obesity (relative risk 3.6, rate fraction 72.2%) was allocated 61.5% of the contribution
  • In a hypothetical normal-weight woman with two risk factors, each factor was allocated 24.5%. The text names the factors as cigarette smoking and asbestos, but Table 4 labels them sex and asbestos
  • When the hypothetical example included synergy (combined relative risk 15), risk factor A was allocated 20.7% and risk factor B 72.6%. The remaining 6.7% went to unknown factors
• The authors' conclusion: The authors proposed the method, with its caveats, as a simple tool to help attribute cause at the individual level. They disagreed with the Chief Justice's view that epidemiology has no role in individual compensation decisions
• Limitations the authors state:
  • Without a biologic model, the method assumes each risk factor is individually sufficient but not necessary to cause the disease
  • When case-control studies supply odds ratios instead of relative risks, conditions must be checked before using them: incidence in controls must be low, and the odds ratio must not be far from 1
  • Unavoidable factors such as age and gender should not be included when the method is used for legal apportionment. Whether factors like obesity count is left to the court
  • The risk fraction ignores time considerations and cannot simply be equated with the probability of causation. The authors said the risk fraction represents only a minimum bound for the probability of causation

Topics

Medical terms in this study, explained in our Lexicon: Attributable Risk, Carcinoma.

Related studies

• Are children more vulnerable to mesothelioma than adults? A comparison of mesothelioma risk among children and adults exposed non-occupationally to blue asbestos at Wittenoom — A. Reid, P. Franklin, et al. (2018)
• Risk Factors of Mortality from All Asbestos-Related Diseases: A Competing Risk Analysis — R. Abós-Herràndiz, T. Rodriguez-Blanco, et al. (2017)
• Occupational risk factors for renal cell carcinoma: agent-specific results from a case-control study in Germany. MURC Study Group. Multicenter urothelial and renal cancer study — B. Pesch, J. Haerting, et al. (2000)
• Endemic malignant mesothelioma: exposure to erionite is more important than genetic factors — M. Metintas, G. Hillerdal, et al. (2010)
• Response to: 'Are children more vulnerable to mesothelioma than adults? A comparison of mesothelioma risk among children and adults exposed non-occupationally to blue asbestos at Wittenoom' by Reid et al — A. Farioli, P. Boffetta, et al. (2019)
• The contribution of environmental monitoring in the epidemiological assessment of exogenous risk. The experience of ARPA in the Emilia-Romagna Region of Italy — A. Zavatti, P. Lauriola (1999)

Browse return to medical research.