Evaluation of Confounding and Selection Bias in Epidemiological Studies of Populations Exposed to Low-Dose, High-Energy Photon Radiation

Our plain-English summary of Evaluation of Confounding and Selection Bias in Epidemiological Studies of Populations Exposed to Low-Dose, High-Energy Photon Radiation, by M.K. Schubauer-Berigan, A. Berrington de Gonzalez, E. Cardis, et al (Journal of the National Cancer Institute. Monographs, 2020).

The study

• Authors: M.K. Schubauer-Berigan, A. Berrington de Gonzalez, E. Cardis, D. Laurier, J.H. Lubin, M. Hauptmann, D.B. Richardson
• Journal: Journal of the National Cancer Institute. Monographs
• Published: 2020;2020(56):133-153
• Type of article: Systematic review
• Cited by: 33 articles, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Abstract

Background: Low-dose, penetrating photon radiation exposure is ubiquitous, yet our understanding of cancer risk at low doses and dose rates derives mainly from high-dose studies. Although a large number of low-dose cancer studies have been recently published, concern exists about the potential for confounding to distort findings. The aim of this study was to describe and assess the likely impact of confounding and selection bias within the context of a systematic review.

Methods: We summarized confounding control methods for 26 studies published from 2006 to 2017 by exposure setting (environmental, medical, or occupational) and identified confounders of potential concern. We used information from these and related studies to assess evidence for confounding and selection bias. For factors in which direct or indirect evidence of confounding was lacking for certain studies, we used a theoretical adjustment to determine whether uncontrolled confounding was likely to have affected the results.

Results: For medical studies of childhood cancers, confounding by indication (CBI) was the main concern. Lifestyle-related factors were of primary concern for environmental and medical studies of adult cancers and for occupational studies. For occupational studies, other workplace exposures and healthy worker survivor bias were additionally of interest. For most of these factors, however, review of the direct and indirect evidence suggested that confounding was minimal. One study showed evidence of selection bias, and three occupational studies did not adjust for lifestyle or healthy worker survivor bias correlates. Theoretical adjustment for three factors (smoking and asbestos in occupational studies and CBI in childhood cancer studies) demonstrated that these were unlikely to explain positive study findings due to the rarity of exposure (eg, CBI) or the relatively weak association with the outcome (eg, smoking or asbestos and all cancers).

Conclusion: Confounding and selection bias are unlikely to explain the findings from most low-dose radiation epidemiology studies.

© The Author(s) 2020. Published by Oxford University Press. All rights reserved. For permissions, please email: [email protected].

In plain English

This article was a review of confounding and selection bias in 26 studies, published from 2006 through 2017, of cancer after low-dose radiation (mean absorbed dose under 100 mGy) in environmental, medical and occupational settings. The authors tabulated which factors each study controlled for and gathered direct and indirect evidence of confounding. They also ran theoretical adjustments for smoking and asbestos in nuclear workers and for predisposing conditions in children given CT scans. For asbestos, they assumed heavy exposure raised the risk of death from all cancers except leukemia 1.3-fold, based on two cohorts of chrysotile textile workers. On that assumption, asbestos exposure would have had to exceed 100% in the highest dose category of the INWORKS nuclear worker study to explain its results; at a 1.6-fold risk, it would have had to rise from 8% to 61%. In the US and UK worker cohorts, pleural cancer and other mesothelioma were strongly positively associated with dose. The authors concluded that confounding and selection bias are unlikely to explain the findings of most low-dose radiation studies.

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: Systematic literature review assessing confounding and selection bias in published studies, with theoretical (indirect) adjustment calculations
• Who was studied: 26 eligible published studies: 8 of environmental, 4 of medical and 14 of occupational radiation exposure (including nuclear workers, Chornobyl recovery workers, atomic test participants and US radiologic technologists)
• Exposure: Low-dose (mean absorbed dose under 100 mGy), predominantly low-linear energy transfer radiation. For the asbestos calculation the authors assumed about 10% of nuclear workers might have been highly exposed, with the risk estimate taken from cohorts of textile workers exposed to chrysotile asbestos.
• Main results:
  • In the INWORKS nuclear worker study, assuming heavy asbestos exposure raised all-cancer (excluding leukemia) risk 1.3-fold, asbestos prevalence would have had to rise from 6% in the lowest dose category to greater than 100% in the highest to explain the observed dose response. The authors described this as impossible.
  • If the asbestos-related risk were 1.6, asbestos prevalence in INWORKS would have had to rise from 8% to 61% in the highest dose category, which the authors called implausible. They noted a low correlation between asbestos and radiation dose at the Portsmouth Naval Shipyard (r = 0.09).
  • In the US and UK nuclear worker cohorts, pleural cancer and other mesothelioma were strongly positively associated with radiation dose. The authors took this to suggest that asbestos exposure may be a positive confounder for lung cancer in these studies and in INWORKS.
  • For INWORKS, at overall smoking prevalences of 50 to 70% and a smoking-related all-cancer relative risk of 1.6, the calculated smoking prevalences in the highest dose categories exceeded 1.0. The authors concluded that smoking could not explain the full range of observed radiation risks.
  • For pediatric CT and leukemia, Down syndrome prevalence would have had to rise from 0.2% to 4% across dose categories to explain the dose response. The UK study found a prevalence of just 0.03%.
  • The most consistent finding was likely downward bias (toward the null) in nuclear worker studies from healthy worker survivor bias of unknown magnitude.
• The authors' conclusion: The authors concluded that confounding and selection bias are unlikely to explain the findings from most low-dose radiation epidemiology studies.
• Limitations the authors state:
  • Confounding from workplace coexposures, such as asbestos, was not directly evaluated in most occupational cohorts.
  • The indirect adjustment did not address the precision of the adjusted estimates. The authors considered this beyond the scope of the article because the required information does not exist for most studies.
  • Healthy worker survivor bias cannot be directly adjusted for with conventional statistical methods.
  • The reviewed studies controlled for relatively few factors.

Topics

Medical terms in this study, explained in our Lexicon: Confounding, Selection Bias, Systematic Review, Confounding by Indication, Epidemiology.

Related studies

• Evaluation of selection bias in a cross-sectional survey — I.J. Selikoff, H. Seidman (1991)
• Re: Bias in the proportionate mortality ratio analysis of small study populations: A case on analyses of radiation and mesothelioma — M.J. Hein, M.K. Schubauer-Berigan (2015)
• Commentary: Picking the winners-Leaning into selection bias — E. Wakeam, B. Kidane (2020)
• Bias in the proportionate mortality ratio analysis of small study populations: a case on analyses of radiation and mesothelioma — J.Y. Zhou (2014)
• Population pharmacokinetics and exposure-response relationship of amatuximab, an anti-mesothelin monoclonal antibody, in patients with malignant pleural mesothelioma and its application in dose selection — A. Gupta, Z. Hussein, et al. (2016)
• Effects of Photon Radiation on DNA Damage, Cell Proliferation, Cell Survival, and Apoptosis of Murine and Human Mesothelioma Cell Lines — S. Keam, K.M. MacKinnon, et al. (2022)

Browse return to medical research.