Evaluation of selection bias in a cross-sectional survey

Our plain-English summary of Evaluation of selection bias in a cross-sectional survey, by I.J. Selikoff, H. Seidman (American Journal of Industrial Medicine, 1991).

The study

• Authors: I.J. Selikoff, H. Seidman
• Journal: American Journal of Industrial Medicine
• Published: 1991;20(5):615-27
• Cited by: 15 articles, as counted by NIH iCite
• Read it: PubMed · Publisher
• In J-M's library: Johns-Manville's company medical library held a copy (J-M no. 004914).

Abstract

Selection bias is inherent in all occupational cohorts. Selection bias at entry has long been known and is commonly referred to as a "healthy worker effect." Less well appreciated is selection during the life of a cohort resulting from life-style factors (e.g., cigarette smoking); aging with accompanying chronic diseases, economic and demographic circumstances; and diseases that might result from exposures suffered by the cohort being studied, that influence whether individuals remain in a trade. These factors weigh differently at different times. Thus, at any point in time, "surviving" members of a cohort reflect an amalgam of selection factors. When such groups are studied in cross-sectional surveys there can be uncertainty whether clinical, radiological and physiological findings are necessarily representative for the trade or occupation as a whole. We analyzed the results of a large clinical field survey of long-term asbestos insulation workers to investigate whether the non-participants differed substantially from those who were examined. Five thousand three hundred and fifty-five (5,355) men, of an initial cohort of 17,800 established January 1, 1967, had reached 30 or more years from onset of their work by July 1, 1981. All were invited to come for examination. Two thousand and seventy-seven (2,077) came, and 3,278 did not. We questioned a sample of 1,393 non-responders to see why they failed to appear. The answers did not give evidence of significant health-related selection influence. Sickness only infrequently kept them away. We then followed both groups--those examined and those not examined--to the end of 1987 for their mortality experience. There was no great difference. The non-responders had somewhat fewer deaths overall and proportionately fewer of asbestos-associated cancers, such as mesothelioma and lung cancer. The results indicated that, in this cohort, there did not seem to be health-related selection bias that determined whether or not cohort members responded to invitations for examinations.

In plain English

This study asked whether asbestos insulation workers who took part in a medical survey differed from those who did not. In 1967 the authors enrolled 17,800 members of the insulation workers' union in the United States and Canada. Of these, 5,355 men alive and at least 30 years past the start of their work by July 1, 1981 were invited for examination. 2,077 came, between 1981 and 1983, and 3,278 did not. The authors questioned 1,393 of the men who did not come. Personal reasons, including disinterest, were the most common (559, or 40%). Almost 20% fell under failure to receive invitations, which included lost contact with the union and late or missing invitations; illness accounted for 17% and pending litigation almost 10%. Both groups were then followed for deaths to the end of 1987. The authors compared observed deaths with those expected from U.S. white male death rates. For all causes the ratio was 1.09 among those examined and 0.93 among those not examined. For lung cancer it was 3.92 and 2.36. The authors concluded that health-related selection bias was not great and did not necessarily invalidate the survey results.

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: Analysis of non-participation in a cross-sectional clinical field survey. It combined interviews of non-responders about their reasons with follow-up of deaths among participants and non-participants.
• Who was studied: Men from a cohort of 17,800 members of the insulation workers union (IAHFIAW) in the United States and Canada, enrolled January 1, 1967. 5,355 who had reached 30 or more years from onset of work by July 1, 1981 were invited. 2,077 were examined from November 15, 1981 to November 24, 1983 in 19 U.S. cities, and 3,278 did not attend. The first 1,393 consecutive non-responders in the initial cities (Chicago, Columbus, Boston, Atlanta, Tampa, Baltimore) were questioned.
• Exposure: Long-term work as asbestos insulation workers, at least 30 years from onset of work. The article does not name fiber types or give exposure doses.
• Main results:
  • Among the 1,393 questioned non-responders, 257 (18.45%) fell under 'failure to receive invitations'. This category included men no longer in contact with their local union (160), unpublished telephones (61), and invitations not received or received too late (36).
  • Personal factors were the largest reason among questioned non-responders: 559 (40.13%). This included 304 citing disinterest.
  • Illness was the reason for 240 (17.23%) of questioned non-responders. The authors described it as mostly unrelated to asbestos-associated disease. The reported diagnoses were not verified by the authors.
  • Pending litigation (advice of an attorney) kept 136 (9.76%) of questioned non-responders from coming.
  • The authors stated that four-fifths of the non-responses among the 1,393 questioned were for reasons other than health.
  • 9% of questioned non-responders reported that chronic illness kept them away, equal to 2.4% of those eligible.
  • From July 1981 to the end of 1987, the ratio of observed to expected deaths from all causes was 1.09 for the 2,077 examined men and 0.93 for the 3,278 non-participants.
  • The observed to expected ratio for cancers at all sites was 2.54 for examined men and 1.72 for non-participants. For lung cancer it was 3.92 and 2.36.
  • Pleural mesothelioma deaths were 117.88 per 100,000 person-years among examined men and 109.73 among non-participants. Peritoneal mesothelioma deaths were 287.33 and 198.56 per 100,000 person-years.
  • There were proportionately more asbestosis deaths among men who did not come for examination.
  • The authors checked the 30-year cutoff against 2,820 men at 25 to 29 years from onset. Of these, 216 died by December 31, 1987. Their observed to expected ratio was 1.16 for all causes and 2.33 for lung cancer. The authors described this as not substantially different from the 30-plus-year group.
  • When deaths in 1981 to 1984 were compared with 1985 to 1987, there was no initial surge of cancer deaths, among either examined or unexamined men.
• The authors' conclusion: The authors concluded that some health-related selection bias operated, but it was not great and did not necessarily invalidate the survey results. Non-participants' mortality over 6 years was, if anything, somewhat better than that of participants. In their view, properly constituted and defined cohorts can be studied and valid clinical, radiological and physiological data obtained.
• Limitations the authors state:
  • The illness diagnoses reported by non-responders were not verified by the authors.
  • The observed to expected ratios were calculated for information only. Observed deaths used 'Best Evidence' diagnoses while expected deaths came from death certificate rates, which are not strictly comparable.
  • The original 1967 cohort was itself a survivor group. Men who had earlier left the union or died were not known to the authors.
  • Every investigation and every cohort must be evaluated individually for potential bias. The authors stated that more data are needed on selection bias.

Topics

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

Medical terms in this study, explained in our Lexicon: Selection Bias, Healthy Worker Effect, Insulator, Mesothelioma, Lung Cancer.

Related studies

• Evaluation of Confounding and Selection Bias in Epidemiological Studies of Populations Exposed to Low-Dose, High-Energy Photon Radiation — M.K. Schubauer-Berigan, A. Berrington de Gonzalez, et al. (2020)
• Asbestos-related disease in railroad workers. A cross-sectional study — L.C. Oliver, E.A. Eisen, et al. (1985)
• Commentary: Picking the winners-Leaning into selection bias — E. Wakeam, B. Kidane (2020)
• Are the knowledge of non-malignant asbestos-related diseases and lung function impairment differentially associated with psychological well-being? A cross-sectional study in formerly asbestos-exposed workers in Germany — J. Lang, M.K. Felten, et al. (2019)
• Nonmalignant pleural lesions due to environmental exposure to asbestos: a field-based, cross-sectional study — M. Metintas, S. Metintas, et al. (2005)
• Modelling prevalence and incidence of fibrosis and pleural plaques in asbestos-exposed populations for screening and follow-up: a cross-sectional study — C. Paris, A. Martin, et al. (2008)

Browse more research in American Journal of Industrial Medicine, or return to medical research.