Abstract
BACKGROUND: Internet Addiction (IA) is often shown to be associated with health issues, but no study explicitly examined a possible gradient in the association between different levels of IA and health. This study aimed to examine if the levels of IA had a graded relationship with poor sleep quality, psychological distress, and self-rated health among university students in Bangladesh. METHODS: In this cross-sectional study, a sample of 625 students from six universities/colleges responded to an online survey that contained measures of internet addiction test (IAT), general health questionnaire (GHQ-12), sleep quality, and self-rated health. Modified Poisson regression models were fitted to estimate the adjusted risk ratios (RR) and confidence intervals (CI) of the associations between IA and health outcomes. RESULTS: The IA levels were associated with each of the three health outcomes in a linear fashion. Compared to the lowest IA quintile, the highest quintile remained associated with an increased risk of poor-quality sleeping (RR: 1.77; 95% CI: 1.26, 2.48), psychological distress (RR: 2.09; 95% CI: 1.55, 2.82), and worse self-rated health (RR: 1.46; 95% CI: 1.09, 1.96) after adjusting for socio-demographic covariates. There were also dose-response associations between IAT z-scores and health outcomes. The association between IAT z-scores and psychological distress was significantly stronger in males compared to females (p-value for interaction<0.05). CONCLUSIONS: The study found strong gradients between levels of addiction to internet and health outcomes, suggesting that increased health risks may exist even at lower levels of internet addiction. The findings highlight the need for departure of current research from a focus on the classic dichotomy of problematic versus not problematic internet use and a move toward recognizing the potential hierarchical effects of IA on health.