Calculation of fractions of lung cancer incidence attributable to occupational exposure to asbestos and combustion products in Stockholm, Sweden

Our plain-English summary of Calculation of fractions of lung cancer incidence attributable to occupational exposure to asbestos and combustion products in Stockholm, Sweden, by P. Gustavsson, A. Ahlbom, T. Andersson, et al (European Journal of Epidemiology, 2003).

The study

• Authors: P. Gustavsson, A. Ahlbom, T. Andersson, P. Schéele
• Journal: European Journal of Epidemiology
• Published: 2003;18(10):937-40
• Cited by: 6 articles, as counted by NIH iCite
• Read it: PubMed · Publisher

Abstract

In a population-based case-referent study of lung cancer we wanted to estimate the over-all influence on the lung cancer incidence from several occupational exposures. Standard methods to do this are based on addition of separately estimated attributable fractions (AFs) by rather complex formulas. Although a simple and valid method for direct estimation of summary effects was published in 1990, it is not well known and has rarely been used. We here describe the method and apply it to the data from the case-referent study. The AF for withdrawal of occupational exposure to both asbestos and combustion products were nearly identical regardless of if it was calculated by an algorithm for summation of AF for the exposure factors separately (6.90%), by a bootstrap method (6.89%, 95% confidence interval, CI: 3.69, 10.04), or by the simple 'dichotomization'-method (6.88%, 95%CI: 3.81, 9.84). The method is very easy to apply to population-based case-referent studies analyzed by logistic regression.

In plain English

This was a statistical methods paper. The authors described a simple 'dichotomization' method, published by Hartge and colleagues in 1990, for estimating the combined attributable fraction of several exposures. This is the proportion of cases that could be prevented by withdrawing the exposures. They applied it to their earlier population-based case-referent study in Stockholm County, Sweden, of 1042 lung cancer cases and 2364 randomly selected referents. The analysis adjusted for other occupational exposures, historical environmental air pollution, residential radon and lifetime smoking. Compared with people unexposed to both factors, the relative risk of lung cancer was 1.6153 for occupational asbestos exposure without combustion products, 1.6704 for combustion products without asbestos, and 2.2467 for both. The attributable fraction for occupational exposure to asbestos and/or combustion products was 6.90% by a standard formula combining three non-overlapping exposure groups, 6.89% (95% CI 3.69 to 10.04) by bootstrapping, and 6.88% (95% CI 3.81 to 9.84) by the simple method. The authors concluded the method was simple and valid. They noted that results depend on the regression model, exposure indicators and data quality.

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: Statistical methods paper, illustrated with a hypothetical example and applied to data from a population-based case-referent (case-control) study of lung cancer
• Who was studied: 1042 lung cancer cases and 2364 referents selected randomly from the population of Stockholm County, Sweden; study years not given in this article
• Exposure: Occupational exposure to asbestos and to combustion products, grouped into three non-overlapping categories: asbestos only, combustion products only, or both; fiber types not specified
• Main results:
  • Compared with those unexposed to both factors, people exposed to asbestos but not combustion products had an adjusted relative risk of 1.6153. They made up 0.0934 of cases, giving an attributable fraction of 3.55%.
  • Those exposed to combustion products but not asbestos had a relative risk of 1.6704 and made up 0.0366 of cases, giving an attributable fraction of 1.46%.
  • Those exposed to both asbestos and combustion products had a relative risk of 2.2467 and made up 0.0337 of cases, giving an attributable fraction of 1.87%.
  • The summary attributable fraction for occupational exposure to asbestos and/or combustion products was 6.90%. It was calculated with a formula (Equation 1) that combines the three non-overlapping exposure groups, not by simply adding fractions for asbestos and for combustion products.
  • Bootstrapping with 10,000 samples gave a mean attributable fraction of 6.89% (95% CI 3.69 to 10.04).
  • The simple dichotomization method, using a single indicator for exposure to asbestos and/or combustion products, gave an attributable fraction of 6.88% (95% CI 3.81 to 9.84). With this method, the relative risk for the exposed group was 1.7242, and the exposed group made up 0.1638 of cases.
• The authors' conclusion: The authors considered the Hartge et al. method a simple and valid way of calculating summary attributable fractions for groups of exposures.
• Limitations the authors state:
  • The authors stated that the validity of the attributable fraction depends on the validity of the underlying regression model and exposure indicators, and on the quality of the data.
  • The bootstrap confidence interval was slightly wider than the one from the simple method. The authors suggested this may be because the two regression models differed in the number of exposure variables.

Topics

The National Library of Medicine indexes this study under subjects our site covers: Lung cancer.

Medical terms in this study, explained in our Lexicon: Lung Cancer, Occupational Exposure, confidence interval.

Related studies

• Low-dose exposure to asbestos and lung cancer: dose-response relations and interaction with smoking in a population-based case-referent study in Stockholm, Sweden — P. Gustavsson, F. Nyberg, et al. (2002)
• Apportioning causes, targeting populations and predicting risks: population attributable fractions — N. Chatterjee, P. Hartge (2003)
• Past exposure to asbestos and combustion products and incidence of cancer among Finnish locomotive drivers — P. Nokso-Koivisto, E. Pukkala (1994)
• Multicentre, population-based, case-control study of particulates, combustion products and amyotrophic lateral sclerosis risk — A.E. Visser, F. D'Ovidio, et al. (2019)
• INTEROCC case-control study: lack of association between glioma tumors and occupational exposure to selected combustion products, dusts and other chemical agents — A. Lacourt, E. Cardis, et al. (2013)
• The INTEROCC case-control study: risk of meningioma and occupational exposure to selected combustion products, dusts and other chemical agents — D.M. McElvenny, M. van Tongeren, et al. (2018)

Browse more research in European Journal of Epidemiology, or return to medical research.