Navarro EA, Segura J, Portoles M, Gomez-Perretta C (2003)

This is a study somewhat similar to that of Santini et al. (2003). The authors carried out a survey of people to determine if there were any differences in the symptoms of those living close to base stations, compared with those living further away. They delivered questionnaires (a Spanish language adaptation of Santini's) to people in La Nora, in the Murcia region of Spain. They introduced this as a study to evaluate the impact on the area of the cellular phone base stations. The questionnaires asked about a variety of symptoms, and also gathered data about age, gender, and distance to the antennas. The rate of response was "about 70%". Results were based on 101 questionnaires. The survey was supplemented by measurements of electromagnetic fields (EMF) in the bedrooms of the participants.

EMF measurements were lower than 0.2µW/cm². Spanish legislation established a maximum level of 450µW/cm² at a single frequency (900 MHz). Two groups were compared - those with high exposure, averaging 0.11µW/cm², who declared themselves to be less than 150 metres from the base station, and those with low exposure, who were at a distance greater than 250 metres and with an average exposure of 0.01µW/cm². The high exposure group had a significantly higher severity rating of some symptoms, including irritability, headache, nausea, appetite loss, discomfort, sleep disturbance, depression, difficulty in concentration, dizziness, and skin alterations. There was also a statistically significant correlation between most symptoms and the measured power density.

There were differences between the groups - the high exposure group used cell phones and personal computers more than the low exposure group.

It is possible that the results are biased, in that the declared intent of the study was to "evaluate the impact on the area of the cellular phone base stations". Those subjects living closer to the station may have perceived themselves to be at greater risk of health effects. Since the questionnaire depended on a personal statement of symptom severity, rather than on any objective measures, it is difficult to exclude respondent bias.