Abstract
PURPOSE: An increase in excessive Internet use (EIU) among adolescents during the COVID-19 pandemic was suggested in many studies. However, these studies were mostly based on cross-sectional and/or unrepresentative samples. METHODS: Using data from a nationwide Finnish school survey in the years 2017, 2019 and 2021 (N = 450,864; aged 13-20 years), changes in the prevalence of EIU (EIUS, 5-item) were assessed. The effects of COVID-19 (year 2021 vs. 2017/2019 combined) and linear trend were analyzed in logistic regression models. Models were adjusted for loneliness, depression, anxiety, and sociodemographic factors. RESULTS: Among males, EIU prevalence varied minimally (7.8-8.1%) from 2017 to 2021. Among females, the prevalence increased from 6.8 to 11.7% and the effect of COVID-19 was significant (OR = 1.53; p < 0.001). Including the linear trend in the model turned the COVID-19 effect on EIU among females non-significant (p = 0.625), whereas the trend was significant (OR = 1.17; p < 0.001). Adjusting the models with mental health-related factors attenuated the effect of COVID-19 to some extent, but not the effect of linear trend. CONCLUSIONS: There is a sex difference in the way the prevalence of EIU developed from 2017 to 2021 among Finnish adolescents. In males, there was no indication of increased prevalence of EIU and among females, while the effect of COVID-19 was first found, it was also suggested to be a product of a trend already started before the COVID-19 pandemic. These results are in contrast with some earlier studies suggesting an effect of COVID-19 on EIU.