ID Number		1838
Study Type		Epidemiology
Model		INTERPHONE - IARC exposure assessment during 900, 1800 MHz (NMT, GSM, W-CDMA) mobile phone use and in support of studies to examine correlations with brain, acoustic nerve, parotid gland tumors
Details		Volunteers (n = 672 in 11 countries) were evaluated for recall ability 6 months after a period of use, and this information was compated to data collected on software modified mobile phones as well as billing records. A 2007 dosimetry paper reported subject recall of number of calls was more accurate than recall of time of calls. It also described the methods, description of the study population, and potential problems that might confound the study results. These included individual exposure assessment (dynamic power control, DTx, phone model, use patterns), recall error, selection bias, and effects of region, age, sex, and educational / socioeconomic status. A 2008 paper on recall bias compared recall in nested subpopulations and showed a general trend for under-estimation of call number, over-estimation of call duration, and a tendency for over-estimation of use in brain tumor cases with longer mobile phone use - this has the potential to cause over-estimation bias. Specifically, the authors report correlations were moderate to high (avg ~ 0.7) between recall and recorded phone records over the most recent 6 month period, although subjects on average underestimated the number of calls per month (mean ratio = 0.81, 95% CI 0.85-0.99), and overestimated call duration (mean ratio = 1.42, 95% CI 1.29 to 1.56). Underestimation increased in light users and overestimation increased in heavy users. There was significant variability between countries. The authors conclude that subjects recall recent phone use with moderate systematic error and substantial random error (which can be expected to reduce the power of the Interphone study to detect an increase in risk of brain, acoustic nerve, and parotid gland tumours with increasing mobile phone use, if one exists). Specifically, they suggest that up to 60% of subjects may be misclassified in terms of exposure category due to systemic recall error. A modeling study using a hypothetical user base was also shown to underestimate tumor correlations when substantial recall bias was introduced. In a 2009 study, the Interphone group reported an estimated 10% downward bias in the overall data as a result of selection bias in the control sample. Controls that agreed to participate in the study had significantly higher self-reported mobile phone use. Non-users were less likely to participate in the study than regular mobile phone users. A study by Lahkola et al relevant to exposure assessment in the Finish INTERPHONE arm and based upon a subset of the cases reported on errors (random bias and systemic / confounding bias) and suggested that selection bias (regular mobile phone users were more likely to participate as controls than non-users) may have influenced some findings and suggest a reverse correlation (protective effect) between mobile phone use and high grade gliomas. In another 2009 study, the authors followed 500+ volunteers in 12 countries (60,000+ calls) using software modified mobile phones and reported that mobile phone transmit power was approximately 50% of maximum on average and varied by a factor of 2 to 3 between study sites and different network operators. Maximum transmit power was used 39% of the time, and output power decreased with increasing call duration. There was little reported variation between use when driving vs. stationary or inside vs. outside. There were higher average transmit powers seen in very sparse-rural sites vs. urban use. In other studies, members of the INTERPHONE team reported on the ability to estimate 3D SAR distributions in head models using the limited data obtained from compliance testing measurements.
Findings		Not Applicable to Bioeffects
Status		Completed With Publication
Principal Investigator		University Barcelona (CREAL) Spain - ecardis@creal.cat
Funding Agency		EU, INTERPHONE (IARC) coordinated studies, MMF, GSM Association
Country		FRANCE
References		Nicolle-Mir, L Envir Risques Sante, (2010) 9:11-12 Wake, K et al. Phys Med Biol, (2009) 54:5695-5706 Vrijheid, M et al. Occup Environ Med, (2009) 66:664-671 Vrijheid, M et al. Ann Epidemiol, (2009) 19:33-41 Vrijheid, M et al. J Expo Sci Environ Epi., (2009) 19:369-381 Vrijheid, M et al. Occup Environ Med, (2006) 63:237-243 Auvinen, A et al. Eur J Cancer Prev, (2006) 15:516-523 Vrijheid, M et al. J Exposure Sci Environ Epidemiol, (2006) 16:371-384 Lahkola, A et al. Annals of Epidemiology, (2005) 15:321-325
Comments		Many of the validation participants may have been academic colleagues and aquantances not entirely representative of the actual case population