Passive smoking in the workplace: classical and Bayesian meta-analyses

  • B. J. Biggerstaff
  • R. L. Tweedie
  • K. L. Mengersen
Original Article


There are currently several classical and Bayesian methods of meta-analysis available for combining epidemiological results. We describe and compare these in a consistent framework, and apply them to published studies of the relative risk of lung cancer associated with exposure to environmental tobacco smoke in the workplace. We find that although all methods give reasonably similar combined estimates of relative risk of lung cancer associated with this exposure (none of which is significantly raised above unity, in either a frequentist or a Bayesian sense), the approximations arising from classical methods appear to be nonconservative and should be used with caution. The Bayesian methods, which account more explicitly for possible inhomogeneity in studies, give slightly lower estimates again of relative risk and wider posterior credible intervals, indicating that inference from the non-Bayesian approaches might be optimistic.

Key words

Passive smoking Meta-analysis Environmental tobacco smoke (ETS) Hierarchical Bayes Random effects 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Breslow NE, Day NE (1986) Statistical methods in cancer research, vol. II. IARC, Lyon. IARC Scientific Publication 82Google Scholar
  2. 2.
    Brownson RC, Reif JS, Keefe TJ, Ferguson SW, Pritzl JA (1987) Risk factors for adenocarcinoma of the lung. Am J Epidemiol 125:25–34Google Scholar
  3. 3.
    Kabat GC, Wynder EL (1984) Lung cancer in nonsmokers. Cancer 53:1214–1221Google Scholar
  4. 4.
    Carlin JB (1992) Meta-analysis for 2 × 2 tables: a Bayesian approach. Statist Med 11:141–158Google Scholar
  5. 5.
    DerSimonian R, Laird NM (1986) Meta-analysis in clinical trials. Controlled Clin Trials, 7:177–188Google Scholar
  6. 6.
    DuMouchel W (1990) Bayesian meta-analysis. In: Berry DA (ed) Statistical methodology in the pharmaceutical industry. Dekker, New York, pp 509–529Google Scholar
  7. 7.
    Fontham ETH, Correa P, Wu-Williams A, Reynolds P, Greenberg AS, Buffler PA, Chen VW, Boyd P, Alterman T, Austin DF, Liff J, Greenberg SD (1991) Lung cancer in non smoking women: a mutlicenter case-control study. Cancer Epidemiol Biomark Prevent 1:35–43Google Scholar
  8. 8.
    Garfinkel L, Auerbach O, Joubert L (1985) Involuntary smoking and lung cancer: a case study. JNCI 75:463–469Google Scholar
  9. 9.
    Hedges LV, Olkin I (1985) Statistical methods for meta-analysis. Academic, New YorkGoogle Scholar
  10. 10.
    Holcomb C (1988) Impact of environmental tobacco smoke on airline cabin air quality. Environ. Technol Lett 9:509–514Google Scholar
  11. 11.
    Oldaker GB III, Conrad JC, Jr (1987) Estimation of effect of environmental tobacco smoke on air quality within passenger cabins of commercial aircraft. Environ Sci Technol 21:994–999Google Scholar
  12. 12.
    Janerich DT, Thompson WD, Varela LR, Greenwald P, Chorost S, Tucci C, Zaman MB, Melamed MR, Kiely M, McKneally MF (1990) Lung cancer and exposure to tobacco smoke in the household. New Engl J Med 323:632–636Google Scholar
  13. 13.
    Kabat GC (1990) Epidemiologic studies of the relationship between passive smoking and lung cancer. In: Proc 1990 Toxicol Forum, p 187–199Google Scholar
  14. 14.
    Kalandidi A, Katsouyanni K, Voropoulou N, Bastas G, Saracci R, Trichopoulos D (1990) Passive smoking and diet in the aetiology of lung cancer among non-smokers. Cancer Causes Contr 1:15–21Google Scholar
  15. 15.
    Koo LC, Ho JH, Saw D, Ho C (1987) Measurements of passive smoking and estimates of lung cancer risk among non-smoking Chinese females. Int J Cancer 39:162–169Google Scholar
  16. 16.
    Lam TH, Cheng KK (1988) Passive smoking is a risk factor for lung cancer in never smoking women in Hong Kong. In: Aoki M, Hisamichi S, Tominaga S (eds) Smoking and Health 1987, proceedings of the 6th World Conference on Smoking and Health, TokyoGoogle Scholar
  17. 17.
    Lee PN (1992) Environmental tobacco smoke and mortality. Karger, BaselGoogle Scholar
  18. 18.
    Lee PN (1988) Misclassification of smoking habits and passive smoking: a Review of the evidence. Springer, Berlin Heidelberg New YorkGoogle Scholar
  19. 19.
    Lee PN (1993) An assessment of the epidemiological evidence relating lung cancer risk in never smokers to environmental tobacco smoke exposure. In: Kasuga H (ed) Environmental tobacco smoke. Springer, New York, pp 28–70Google Scholar
  20. 20.
    Light RJ, Pillemer DB (1984) Summing up: the science of reviewing research. Harvard University Press, CambridgeGoogle Scholar
  21. 21.
    Mengersen KL, Tweedie RL Biggerstaff BJ (1994) The impact of method choice in metaanalysis. Austral J Statist (in press)Google Scholar
  22. 22.
    Olkin I (1992) Meta-analysis: methods for combining independent studies. Statist Sci 7:226Google Scholar
  23. 23.
    NRC (1992) Committee on Applied and Theoretical Statistics. Combining information: statistical issues and opportunities for research. National Academy Press, WashingtonGoogle Scholar
  24. 24.
    EPA Review (1992) Health effects of passive smoking: assessment of lung cancer in adults and respiratory disorders in children. National Academy Press, United States EPA, WashingtonGoogle Scholar
  25. 25.
    EPA Draft Review (1990) Health effects of passive smoking: assessment of lung cancer in adults and respiratory disorders in children. National Academy Press, United States EPA, WashingtonGoogle Scholar
  26. 26.
    Robins J, Breslow N, Greenland S (1986) Estimators of the Mantel-Haenszel variance consistent in both sparse data and large-strata limiting models. Biometrics 42:311–323Google Scholar
  27. 27.
    Sato T (1990) Confidence limits for the common odds ratio based on the asymptotic distribution of the Mantel-Haenszel estimator. Biometrics 46:71–80Google Scholar
  28. 28.
    Shimizu H, Morishita M, Mizuno K, Masuda T, Ogura Y, Santo M, Nishimura M, Kunishima K, Karasawa K, Nishiwaki K, Yamamoto M, Hismichi S, Tominaga S (1988) A case-control study of lung cancer in nonsmoking women, Tohuku J Exp Med 154B:389–397Google Scholar
  29. 29.
    Siegel M (1993) Involuntary smoking in the restaurant workplace. JAMA 270:490–493Google Scholar
  30. 30.
    Svensson C, Pershagen G, Klominek J (1989) Smoking and passive smoking in relation to lung cancer in women. Acta Oncol 28:623–629Google Scholar
  31. 31.
    Tweedie RL, Mengersen KL (1992) Lung cancer and passive smoking: reconciling the biochemical and epidemiological approaches. Br J Cancer 66:700–705Google Scholar
  32. 32.
    Tweedie RL, Mengersen KL, Eccleston JA (1994) Garbage in, garbage out: can statisticians quantify the effects of poor data? Chance 7:20–27Google Scholar
  33. 33.
    Wald NJ, Nanchahal K, Thompson SG, Cuckle HS (1986) Does breathing other people's tobacco smoke cause lung cancer? Br Med J 293:1217–1222Google Scholar
  34. 34.
    Waternaux C, DuMouchel W (1993) Combining information across sites with hierarchical Bayesian linear models. In: Proceedings of the Section on Bayesian Statistics. ASA, 1993. Joint Statistical Meetings, San FranciscoGoogle Scholar
  35. 35.
    Wu AH, Henderson BE, Pike MC, Yu MC (1985) Smoking and other risk factors for lung cancer in women. JNCI 74:747–751Google Scholar
  36. 36.
    Wu-Williams AH, Dai XD, Blot W, Xu ZY, Sun XW, Xiao HP, Stone BJ, Yu SF, Feng YP, Ershow AG, Sun J, Fraumenti JF Jr, Henderson BE (1990) Lung cancer among women in north-east China. Br J Cancer 62:982–987Google Scholar
  37. 37.
    Yusuf S, Peto R, Lewis J, Collins R, Sleight P (1985) Beta-blockade during and after myocardial infarction: an overview of the randomized trials. Progr Cardiovasc Dis 27:335–371Google Scholar

Copyright information

© Springer-Verlag 1994

Authors and Affiliations

  • B. J. Biggerstaff
    • 1
  • R. L. Tweedie
    • 1
  • K. L. Mengersen
    • 2
  1. 1.Department of StatisticsColorado State UniversityFort CollinsUSA
  2. 2.School of MathematicsQueensland University of TechnologyBrisbaneAustralia

Personalised recommendations