What have we learnt so far?

In a COSMOS study including over 265,000 participants across the UK, Denmark, Finland, The Netherlands, and Sweden, we found no evidence of increased risk of brain tumours associated with mobile phone use - even beyond 15 years (Feychting et al., 2024). However, we need to continue surveillance of this important question to look for possible longer-term effects.  

In a COSMOS study of 78,000 participants from the UK and The Netherlands, we found no link between time spent on calls and headaches. However, frequent texting was associated with more frequent headaches, suggesting that factors such as posture or lifestyle, rather than RF-EMF exposure, may explain the findings (Traini et al., 2024). 

In a COSMOS study of 25,000 participants in Sweden and Finland, people who spent the most time on phone calls reported slightly more headaches, but this was a weak association that was not explained by RF-EMF exposure. Researchers also reported no evidence of association between mobile call time and either tinnitus or hearing loss (Auvinen et al., 2019). 

Among 25,000 COSMOS participants in Sweden and Finland, those who spent 4+ hours per week on calls reported slightly more sleep difficulties. However, overall, it seems likely that sleep issues were linked to behavioural factors, such as excessive or late-night phone use, rather than RF-EMF exposure (Tettamanti et al., 2020). 

Toledano MB, Smith RB, Chang I, Douglass M, Elliott P. Cohort Profile: UK COSMOS-a UK cohort for study of environment and health. Int J Epidemiol. 2017 Jun 1;46(3):775-787.   

 Toledano MB, Smith RB, Brook JP, Douglass M, Elliott P. How to Establish and Follow up a Large Prospective Cohort Study in the 21st Century--Lessons from UK COSMOS. PLoS One. 2015 Jul 6;10(7):e0131521.  

Summary:  These papers describe how participants were recruited and followed in the UK COSMOS Cohort, including the number of participants, response rates, and key characteristics at both the start of the study and during follow-up. The researchers also explored approaches for recruiting and collecting data from large numbers of participants. They found that postal invitation letters were the most effective way to initially contact potential participants, while online systems worked well for obtaining consent and collecting questionnaire data. Sending reminder messages further improved participation and completion rates, helping to keep more participants engaged throughout the study. 

Schüz J, Elliott P, Auvinen A, Kromhout H, Poulsen AH, Johansen C, Olsen JH, Hillert L, Feychting M, Fremling K, Toledano M, Heinävaara S, Slottje P, Vermeulen R, Ahlbom A. An international prospective cohort study of mobile phone users and health (Cosmos): design considerations and enrolment. Cancer Epidemiol. 2011 Feb;35(1):37-43.  

Summary: This paper describes the COSMOS Study, a large, long-term study investigating whether long-term RF-EMF exposure from mobile phones and other wireless devices may affect health. The study aims to find out if using these devices over many years could increase the risk of cancer or other diseases. The study plans to recruit 250,000 adults aged 18 and over across five European countries—Denmark, Finland, Sweden, Netherlands, and United Kingdom—and follow them for 25 years or more. Information on mobile phone use is collected through questionnaires and network operator records, while health outcomes are tracked using disease registries and repeated questionnaires on symptoms and overall well-being. 

Hillert L, Ahlbom A, Neasham D, Feychting M, Järup L, Navin R, Elliott P. Call-related factors influencing output power from mobile phones. J Expo Sci Environ Epidemiol. 2006 Nov;16(6):507-14.  

Summary: To design a prospective cohort study on mobile phone use and health, a pilot study was conducted in Sweden and the UK (2002–2003) to test recruitment feasibility and exposure assessment. The study evaluated factors affecting phone output power, including urban versus rural area, indoor versus outdoor use, stationary versus moving calls, and call length. Results showed that high output power occurs more often in rural areas, while other factors had less impact. Therefore, urban/rural location should be considered in exposure classification, using either the first connected base station or, if unavailable, the participant’s home address.

In studies on mobile phone use and health, researchers often rely on self-reported information about how much people use their phones. However, people may not accurately recall their phone use—for example, they may forget how often they made calls or misestimate the amount of time spent on them. These recall and reporting inaccuracies can lead to measurement errors in the data. More precise information from mobile network operators is available for some participants, but it is not available for everyone and does not capture phone use from earlier years. To understand this issue, COSMOS researchers have undertaken a series of methodological studies.  

Reedijk M, Portengen L, Auvinen A, Kojo K, Heinävaara S, Feychting M, Tettamanti G, Hillert L, Elliott P, Toledano MB, Smith RB, Heller J, Schüz J, Deltour I, Poulsen AH, Johansen C, Verheij R, Peeters P, Rookus M, Traini E, Huss A, Kromhout H, Vermeulen R; COSMOS Study Group. Regression calibration of self-reported mobile phone use to optimize quantitative risk estimation in the COSMOS study. Am J Epidemiol. 2024 Oct 7;193(10):1482-1493.  

Summary: Researchers used data from the COSMOS Study in Denmark, Finland, The Netherlands, Sweden, and the UK. They analysed mobile phone call data collected between 2007 and 2012 and tested several statistical methods designed to correct for errors in self-reported phone use 

The results showed that regression calibration methods, which adjust for measurement errors using the more reliable operator data, can improve estimates of the relationship between mobile phone use and health outcomes.  This approach helps researchers draw more accurate and reliable conclusions about possible health effects related to mobile phone use. 

Toledano MB, Auvinen A, Tettamanti G, Cao Y, Feychting M, Ahlbom A, Fremling K, Heinävaara S, Kojo K, Knowles G, Smith RB, Schüz J, Johansen C, Poulsen AH, Deltour I, Vermeulen R, Kromhout H, Elliott P, Hillert L. An international prospective cohort study of mobile phone users and health (COSMOS): Factors affecting validity of self-reported mobile phone use. Int J Hyg Environ Health. 2018 Jan;221(1):1-8.  

Summary:  The study examined how accurately people report their own mobile phone usage. This study included 75,993 participants (6,229 from Finland, 30,874 from Sweden, and 38,890 from the UK). Participants reported how often they made phone calls and how long they spent on calls over a three-month period. The self-reported data was then compared with objective records from mobile network operators.  The comparison showed that self-reported information was reasonably accurate overall, especially for people who used only one mobile phone. People tended to report low levels of phone use quite accurately. However, those who used their phones very frequently often underestimated how much they used them, meaning they reported fewer or shorter calls than recorded by their phone operators. The study also found that the accuracy of self-reported phone use was slightly lower among women, younger participants, and people who reported symptoms they believed were related to mobile phone use.  These findings highlight the importance of combining self-reported information with objective data from network operators.  Doing this helps researchers measure mobile phone use more accurately and strengthens the reliability of studies investigating health effects. 

 Heinävaara S, Tokola K, Kurttio P, Auvinen A. Validation of exposure assessment and assessment of recruitment methods for a prospective cohort study of mobile phone users (COSMOS) in Finland: a pilot study. Environ Health. 2011 Mar 8;10:14.  

Summary: This pilot study evaluated the agreement between self-reported and operator-derived mobile phone use, based on a three-month monitoring period, and examined consistency of use over time. A total of 5,400 participants were invited from network operators’ subscriber databases, with operator data available for most participants. Results showed moderate agreement between self-reported and operator-derived call times, with participants tending to overestimate their use, particularly among heavier users. Mobile phone use increased steadily between 2007 and 2009. The study reported response rates  ofe 12% which were unaffected by questionnaire length or recruitment method.