A new approach to control for confounding and reverse causality in radiofrequency electromagnetic field research on health related quality of life

Abstract

Introduction The most relevant sources of radiofrequency electromagnetic fields (RF-EMF) are related to lifestyle such as mobile phone or WLAN use. As a consequence, confounding by unmeasurable psychological factors and reverse causality is a concern, in particular when addressing health related quality of life (HrQoL). To better disentangle the psychological effects of media use from the RF-EMF exposure a new approach is proposed where different usage types are separately considered. Methods We applied latent class analysis to define qualitatively distinct patterns of media usage varying on 11 different use variables in a Swiss sample of 850 adolescents investigated in the years 2012 – 2015. Multivariable regression models were fitted to assess the relationships between the extracted classes and HrQol as measured by the KIDSCREEN questionnaire. Results Five distinct media use profiles could be identified: Gamers, Low users, Medium users, Callers, and High social network users. Group affiliation was dependent on gender and the time of investigation. Affectivity, Self-Perception, parent and school relationship and peer contact scores differed significantly between the five user groups. For instance, high social network users had a lower score on negative affectivity (coefficient: -6.00; 95% CI: (-8.72, -3.29)) and Self-Perception (-3.81; (-6.03, -1.59)) and higher score for peer contact (5.37; (3.05, 7.69)) compared to Low users. Conclusion We found different health profiles in the five user groups which may be due to unmeasured psychological variables and reverse causality. To conduct exposure-response association with RF-EMF within each group separately may be an efficient way to control for such bias and to evaluate potential associations between RF-EMF and HrQol in a robust manner.