Smartphone Dependence, Sleep Quality, and Suicidal Ideation Across Adolescence: A Six-Year Longitudinal Study

Suicide is a leading cause of death during adolescence, and both smartphone dependence and poor sleep have been associated with suicidal ideation. However, it remains unclear whether these associations reflect stable differences between adolescents or year-to-year changes within adolescents. This study examined the developmental trajectories of smartphone dependence, sleep quality, and suicidal ideation over six years spanning the onset of the COVID-19 pandemic and tested whether sleep quality accounted for the association between smartphone dependence and suicidal ideation at the between-person and within-person levels. Data were drawn from six annual waves (2018 to 2023) of the Korean Children and Youth Panel Survey 2018 middle school first-grade cohort (N = 2590; 54.2% boys). A parallel piecewise latent growth curve model (LGCM) and random-intercept cross-lagged panel model (RI-CLPM) were estimated. Smartphone dependence increased before the pandemic onset (B = 1.47 per year) and declined slightly thereafter (B = −0.16), sleep quality declined after the onset (B = −0.05), and suicidal ideation decreased before the onset (B = −0.10) and increased gradually thereafter (B = 0.03), with no immediate shift at the onset. In the LGCM, the indirect effect of smartphone dependence on suicidal ideation through sleep quality was significant for usual levels (B = 0.005, p < 0.001) and pre-onset change (B = 0.008, p = 0.001) but not for post-onset change (B = −0.001, p = 0.570); Wald tests indicated that the path from sleep quality to suicidal ideation (p = 0.001) and the indirect effect (p = 0.004) differed between the two periods. In the RI-CLPM, higher-than-usual smartphone dependence predicted higher-than-usual suicidal ideation one year later (β = 0.05, p < 0.001), whereas no within-person lagged paths involving sleep quality were significant. Smartphone dependence, poor sleep quality, and suicidal ideation thus co-occurred between adolescents, but annual within-person changes did not show a pathway through sleep quality. Because period differences cannot be separated from age-related development in a single cohort, they should not be interpreted as pandemic effects. These findings support attending to adolescents with high smartphone dependence and poor sleep within existing school-based screening systems, whereas the mechanisms through which risk is transmitted require studies with shorter measurement intervals.