Statistical Issues in the Analysis of Data from Occupational Cohort Studies

  • N. Breslow
Part of the Recent Results in Cancer Research book series (RECENTCANCER, volume 120)

Abstract

The overwhelming majority of statistical analyses of occupational cohort data are conducted in terms of the standardized mortality ratio or SMR. Saracci and Johnson (1987) reviewed 55 cancer epidemiology papers published during 1982 in the American Journal of Epidemiology, the British Journal of Industrial Medicine and the Journal of Occupational Medicine. All but one of 20 occupational cohort studies that they identified employed the SMR as the principal method of analysis. In spite of its evident popularity, the SMR is particularly susceptible to the selection and confounding biases that affect observational studies generally (Hill 1953; Cochran 1983) and its uncritical use as a summary measure has received frequent criticism (e.g., Gaffey 1976). The continuing debate regarding the relative advantages of local vs national rates as a basis for comparison (Gardner 1986) is a reminder that many other factors besides membership in a particular occupational cohort affect disease rates. “Background” variation in SMRs computed for cohorts in different geographic areas can be expected to swamp the Poisson sampling variability that is reflected in the usual measures of statistical uncertainty.

Keywords

Nickel Arsenic Mastitis Asbestos 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Akaike H (1973) Information theory and an extension of the maximum likelihood principle. In: Petrov BN, Czaki F (eds) Second international symposium on information theory. Akademiac Kiado, Budapest, pp 267–281Google Scholar
  2. Andersen PK, Borch-Johnsen K, Deckert T, Green A, Hougoard P, Keiding N, Kreiner S (1985) A Cox regression model for the relative mortality and its application to diabetes mellitus survival data. Biometrics 41: 921–932PubMedCrossRefGoogle Scholar
  3. Armitage P, Doll R (1961) Stochastic models for carcinogenesis. In: Proceedings of the 4th Berkeley Symposium on mathematical statistics and probability: biology and problems of health. University of California Press, Berkeley, pp 19–38Google Scholar
  4. Baker RJ, Nelder JA (1978) The GLIM system: release 3. Numerical Algorithms Group, OxfordGoogle Scholar
  5. Berry G (1983) The analysis of mortality by the subject-years method. Biometrics 39: 173–184PubMedCrossRefGoogle Scholar
  6. Breslow NE, Cain KC (1988) Logistic regression for two-stage case-control data. Biometrika 75: 11–20CrossRefGoogle Scholar
  7. Breslow NE, Day NE (1985) The standardized mortality ratio. In: Sen PK, (ed) Biostatistics: statistics in biomedical, public health and environmental sciences. Elsevier, New York, pp 55–74Google Scholar
  8. Breslow NE, Day NE (1987) Statistical methods in cancer research II: design and analysis of cohort studies. International Agency for Research on Cancer, LyonGoogle Scholar
  9. Breslow NE, Langholz B (1987) Nonparametric estimation of relative mortality functions. J Chronic Dis 40 [Suppl 2]: 895–995Google Scholar
  10. Breslow NE, Lubin JH, Marek P, Langholz B (1983) Multiplicative models and cohort analysis. J Am Stat Assoc 78: 1–12CrossRefGoogle Scholar
  11. Cochran WG (1983) Planning and analysis of observational studies. Wiley, New YorkCrossRefGoogle Scholar
  12. Coleman M, Douglas A, Hermon C, Peto J (1986) Cohort study analysis with a FORTRAN computer program. Int J Epidemiol 15: 134–137PubMedCrossRefGoogle Scholar
  13. Cook RD, Weisberg S (1982) Residuals and influence in regression. Chapman and Hall, LondonGoogle Scholar
  14. Cox DR (1972) Regression models and life tables (with discussion). J R Stat Soc B 34: 187–220Google Scholar
  15. Day NE, Brown CC (1980) Multistage models and the primary prevention of cancer. J NCI 64: 977 –989Google Scholar
  16. Doll R, Peto R (1978) Cigarette smoking and bronchial carcinoma: dose and time relationships among regular smokers and life-long non-smokers. J Epidemiol Community Health 32: 303–313PubMedCrossRefGoogle Scholar
  17. Doll R, Morgan LG, Speizer F (1970) Cancers of the lung and nasal sinuses in nickel workers. Br J Cancer 24: 623–632PubMedCrossRefGoogle Scholar
  18. Efron B (1986) How biased is the apparent error rate of a prediction rule? J Am Stat Assoc 81: 461–470CrossRefGoogle Scholar
  19. Fox AJ, Collier PF (1976) Low mortality rates in industrial cohort studies due toselection for work and survival in the industry. Br J Prey Soc Med 30: 225–230Google Scholar
  20. Frome EL (1983) The analysis of rates using Poisson regression models. Biometrics 39:665–674PubMedCrossRefGoogle Scholar
  21. Gaffey WR (1976) A critique of the standardized mortality ratio. J Occup Med 18: 157–160PubMedCrossRefGoogle Scholar
  22. Gardner MR (1986) Considerations in the choice of expected numbers of appropriate comparisons in occupational cohort studies. Med Law 77: 23–47Google Scholar
  23. Gilbert ES (1982) Some confounding factors in the study of mortality and occupational exposures. Am J Epidemiol 116: 177–188PubMedGoogle Scholar
  24. Gilbert ES (1983) An evaluation of several methods for assessing the effects of occupational exposure to radiation. Biometrics 39: 161–171PubMedCrossRefGoogle Scholar
  25. Gilbert ES, Buchanan JA (1984) An alternative approach to analyzing occupational mortality data. J Occup Med 11: 822–828CrossRefGoogle Scholar
  26. Hakulinen T (1981) A Mantel-Haenszel statistic for testing the association between a poiychotomous exposure and a rare outcome.Am J Epidemiol 113:192–197Google Scholar
  27. Holford TR (1980) The analysis of rates and of survivorship using log-linear models.Biometrics 36: 229–306Google Scholar
  28. Kannel WB, Gordon T (1974) The Framingham study. US Government Printing Office, Washington (DHEW publ no (NIH) 74–559 )Google Scholar
  29. Lee AM, Fraumeni JF Jr (1969) Arsenic and respiratory cancer in man: an occupational study. J NCI 42: 1045–1052Google Scholar
  30. Marsh GM, Preininger ME (1980) OCCMAP: a user-oriented occupational cohort mortality analysis program. Am Statist 34: 245–246CrossRefGoogle Scholar
  31. Mauritsen RH (1988) EGRET manual. Statistics and Epidemiology Research Corporation, SeattleGoogle Scholar
  32. Miettinen OS (1972a) Standardization of risk ratios. Am J Epidemiol 96: 383–388PubMedGoogle Scholar
  33. Miettinen OS (1972b) Design options in epidemiologic research: an update. Scand J Work Environ Health 8 [Suppl 1]: 7–14Google Scholar
  34. Muirhead CR, Darby SC (1987) Modelling the relative and absolute risks of radiation-induced cancers (with discussion). J R Stat Soc A 150: 83–118CrossRefGoogle Scholar
  35. Nelson W (1969) Hazard plotting for incomplete failure data. J Quality Technol 1: 27–52Google Scholar
  36. Peto J (1985) Some problems in dose-response estimation in cancer epidemiology. In: Voug VB, Butler GB, Hoel DG, Peakall DB (eds) Methods for estimating risk of chemical injury: human and non-human biota and ecosystems. Wiley, New York, pp 361–380Google Scholar
  37. Peto J, Seidman H, Selikoff IJ (1982) Mesothelioma mortality in asbestos workers:implications for models of carcinogenesis and risk. Br J Cancer 45: 124–135Google Scholar
  38. Prentice RL (1986) A case-cohort design for epidemiologic cohort studies and disease prevention trials. Biometrika 73: 1–12Google Scholar
  39. Preston DL, Pierce DA (1986) Program documentation for AMFIT and PYTABS. Radiation Effects Research Foundation, HiroshimaGoogle Scholar
  40. Robins J (1987) A graphical approach to the identification and estimation of causal parameters in mortality studies with sustained exposure periods. J Chronic Dis 40 [Suppl 2]: 1395–1615Google Scholar
  41. Rothman KJ, Boice JD Jr (1979) Epidemiologic analysis with a programmable calculator. US Government Printing Office, Washington (NIH publ no 79–1649)Google Scholar
  42. Saracci R, Johnson E (1987) A note on the treatment of time in published cancer epidemiology studies. J Chronic Dis 40 [Suppl 2]: 77S–78SGoogle Scholar
  43. Shore RE, Hempelmann LH, Kowaluk E, Mansur PS, Pasternack BS, Albert RE, Haughie GE (1977) Breast neoplasms in women treated with x-rays for acute postpartum mastitis. J NCI 59: 813–822Google Scholar
  44. Thomas DC (1977) Addendum to a paper by Liddell FDK, McDonald JC, Thomas DC. J R Stat Soc A 140: 483–485Google Scholar
  45. Thomas DC (1987) Pitfalls in the analysis of exposure-time-response relationships. J Chronic Dis 40 [Suppl 2]: 71S–78SPubMedCrossRefGoogle Scholar
  46. Thomas DC (1988) Models for exposure-time-response relationships with applications to cancer epidemiology. Annu Rev Public Health 9: 451–482PubMedCrossRefGoogle Scholar
  47. Waxweiler RJ et al. (1983) A modified life-table analysis system for cohort studies. J Occup Med 25 (2): 115–124PubMedGoogle Scholar
  48. White JE (1982) A two stage design for the study of the relationship between a rare exposure and a rare disease. Am J Epidemiol 115: 119–128PubMedGoogle Scholar
  49. Yule GU (1934) On some points relating to vital statistics, more especially statistics of occupational mortality. J R Stat Soc 97: 1–84CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin·Heidelberg 1990

Authors and Affiliations

  • N. Breslow
    • 1
  1. 1.Department of Biostatistics, SC-32University of WashingtonSeattleUSA

Personalised recommendations