Authored By:
Turner MC, Sadetzki S, Langer CE, Villegas PhD R, Figuerola J, Armstrong BK, Chetrit A, Giles GG, Krewski D, Hours M, McBride ML, Parent ME, Richardson L, Siemiatycki J, Woodward A, Cardis EThis paper reported the results of further analysis of the INTERPHONE study, which is an international collaborative case-control study investigating mobile phone use and tumours of the head and neck. The analysis was done on a subset of five countries out of the thirteen countries involved in the original study. The analysis found that the elimination of a possible bias may result in a stronger positive association between long-term mobile phone use (more than 10 years) and glioma, although it is not statistically significant (odds ratio, OR = 1.26; 95% confidence intervals, 95% CI = 0.90-1.78). The authors concluded that the bias due to differences in the interview dates between cases and controls have resulted in the underestimation of the risk originally reported in the INTERPHONE study.
Published In:
Ann Epidemiol 2016; 26 (12): 827-832.e2The biggest criticism of the INTERPHONE study was its methodology, which led to many biases. The impact of selection and recall biases in a case-control study of mobile phone use has previously been investigated in separate studies by Vrijheid et al. (2009a and 2009b). This study by Turner et al. provided evidence on another possible bias.
Glioma is a highly fatal and fast-progressing disease and a rapid case ascertainment (identifying the number of cases in the period and region that is studied) was an important issue in the INTERPHONE study. This will often mean that cases are interviewed much earlier in the study comparing to controls, which results in a time lag. In the INTERPHONE Study, this resulted in an apparent increase of duration of mobile phone use as well as over-estimation of mobile phone use in controls.
