Investigation of bias related to differences between case and control interview dates in five INTERPHONE countries

Ann Epidemiol. 2016 Dec;26(12):827-832.e2. doi: 10.1016/j.annepidem.2016.09.013. Epub 2016 Oct 8.

Abstract

Purpose: Associations between cellular telephone use and glioma risk have been examined in several epidemiological studies including the 13-country INTERPHONE study. Although results showed no positive association between cellular telephone use and glioma risk overall, no increased risk for long-term users, and no exposure-response relationship, there was an elevated risk for those in the highest decile of cumulative call time. However, results may be biased as data were collected during a period of rapidly increasing cellular telephone use, and as controls were usually interviewed later in time than cases.

Methods: Further analyses were conducted in a subset of five INTERPHONE study countries (Australia, Canada, France, Israel, New Zealand) using a post hoc matching strategy to optimize proximity of case-to-control interview dates and age.

Results: Although results were generally similar to the original INTERPHONE study, there was some attenuation of the reduced odds ratios and stronger positive associations among long-term users and those in the highest categories for cumulative call time and number of calls (eighth-ninth and 10th decile).

Conclusions: Proximity and symmetry in timing of case-to-control interviews should be optimized when exposure patterns are changing rapidly with time.

Keywords: Case-control study; Cellular telephone; Glioma; Matching.

MeSH terms

  • Adult
  • Australia
  • Bias
  • Canada
  • Case-Control Studies
  • Cell Phone / statistics & numerical data*
  • Environmental Exposure / adverse effects*
  • Environmental Exposure / statistics & numerical data
  • Epidemiologic Research Design*
  • Female
  • France
  • Glioma / etiology*
  • Humans
  • Interviews as Topic*
  • Israel
  • Male
  • Matched-Pair Analysis
  • Middle Aged
  • New Zealand
  • Odds Ratio
  • Risk Factors
  • Time Factors

Grant support