Validation of self-reported cellular phone use

Background: In recent years, concern has been raised over possible adverse health effects of cellular telephone use. In epidemiological studies of cancer risk associated with the use of cellular telephones, the validity of self-reported cellular phone use has been problematic. Up to now there is very little information published on this subject. Methods: We conducted a study to validate the questionnaire used in an ongoing international case-control study on cellular phone use, the "Interphone study". Self-reported cellular phone use from 68 of 104 participants who took part in our study was compared with information derived from the network providers over a period of 3 months (taken as the gold standard). Results: Using Spearman's rank correlation, the correlation between self-reported phone use and information from the network providers for cellular phone use in terms of the number of calls per day was good (r = 0.62, 95% CI: 0.45-0.75), while that of the average duration of each call was rather moderate (r = 0.34, 95% CI: 0.11-0.54). Similar results were found when Kappa coefficients were estimated. A value of r = 0.56 (Spearman's correlation, CI: 0.38-0.70) was found for cumulative cellular phone use. Conclusion: Our study suggests that cellular phone use is easier to recall in terms of number of calls made than in terms of cumulative phone use and should thus be used as the basis for the dose-response analysis.