ID Number		1138
Study Type		Epidemiology
Model		Mobile phone use in adults and children and analysis of exposure and recall bias in support of epidemiologic studies
Details		Analysis of data (for evaluation of exposure assessment and recall bias) of three case-control studies conducted from 1992 to 1997 that investigated childhood cancer risk in Germany and exposure of the parents to microwaves as well as solvents, paints, insecticides, and other occupational and environmental agents during the year before conception, during pregnancy and after the child's birth. Each study used similar exposure assessments, including self-administered questionnaires and follow-up telephone interviews with both parents. Analysis showed both case and control Fathers reported more exposures during the prenatal period than during the postnatal period, even though the postnatal period was usually longer. Also, recall bias was different between case and control parents, with case Mothers and Fathers reporting more exposures than controls. Although for the studies examined, the authors feel it is unlikely recall bias entirely explains the odds ratios of 1.5 - 2.0, they did suggest recall bias may have influenced these and possibly other case control studies of childhood diseases. In a preliminary mobile phone-specific study to investigate use patterns, the group analyzed German school-age children compared to adults and what other socioeconomic, age, gender, and education factors might be correlated with mobile phone use. For 4th graders, ~53 % said they used a mobile. Among 12- and 13-year-olds, mobile phone use increased from 3% in 1998 to 69 % in 2002. Phone use was higher in school classes which had a higher proportion of socially disadvantaged children (the authors speculated that having a phone might serve as a status symbol). More than 80 percent of older teen-agers in Japan, Germany, the U.K., France, Australia, and the Scandinavian countries report using a mobile phone, with the lowest teen use rate in North America. The authors suggest that children may use mobile phones more than adults for reading e-mail, SMS, Web access, and receiving music files or pictures. In a subsequent study of exposure assessment using dose phones conducted as part of the INTERPHONE study and involving 58 participants aged 30-56 years from Bielefeld, Germany, the researchers assessed whether variables such as call duration or number of calls could be used as a reliable surrogate for power output (and by further extrapolation, human exposure). The researchers found good agreement between average number of calls per day and self-reported data, and suggest it presents a fairly accurate assessment of actual phone use sufficient to consider the number of calls when estimating exposure. In a recent study in children (European J Epidemiol., 2004), mobile phone was again surveyed in 1,933 nine to ten years old attending 34 different schools in Mainz, Germany. The authors report about a third of the 4th graders (n = 671, or 34.7%) owned a mobile phone, and on average more city kids than rural kids owned phones. Young mobile phone users are more likely to be older, male, an only child, spend more time watching TV and playing computer games, be picked up by parents from school by car, and go to bed later than non-users. Again, the authors note the proportion of cell phone owners was somewhat higher in classes with socially disadvantaged children. There was no correlation with performance in school. A recent study compared exposure assessment in mobile phone studies using questionnaire (self-reported) vs. network subscriptions. The authors re-evaluated data from two prior studies replacing billing records for questionnaire exposure classification and visa versa, and conclude that the two methods are in agreement with each other and give comparable results. Gender was identified as a factor that could lead to exposure misclassification (males more likely to use mobile phones without a private subscription, females more likely to not be regular users). In a related (2009) study looking at general agreement between questionnaire and job exposure matrix data (using acoustic noise level above 80 dBA as a metric), the authors reported modest agreement in controls (n = 191) and acoustic neuroma patients (n = 94) with kappa values of 0.62 and 0.48, respectively.
Findings		Not Applicable to Bioeffects
Status		Completed With Publication
Principal Investigator		University of Mainz, Germany - joachim@cancer.dk
Funding Agency		Private/Instit.
Country		GERMANY
References		Schlaefer, K et al. Eur J Epidemiol., (2009) 24:469-475 Schuz, J et al. Bioelectromagnetics, (2006) 98:1707-1713 Berg, G et al. J Expo Anal Environ Epidemiol., (2005) 15:217-224 Schuz, J Bioelectromagnetics, (2005) 26 suppl 7:S45-S50 Bohler , E et al. Eur J Epidemiol, (2004) 19:1043-1050 Samkange-Zeeb , F et al. J Exp Anal Environ Epidemiol, (2004) 14:245-248 Schuz , J et al. Am J Epidemiol, (2003) 158:710-716
Comments