J Med Internet Res. 2026 Sep 8;28:e92310. doi: 10.2196/92310.
ABSTRACT
BACKGROUND: Problematic internet use (PIU) is an important public health concern among college students because of its associations with adverse psychological, social, and academic outcomes. Stressful life events may contribute to PIU, but previous studies have mainly relied on linear models and have rarely examined nonlinear, threshold, or plateau effects while accounting for sociodemographic and lifestyle factors.
OBJECTIVE: This study aimed to examine the linear and nonlinear dose-response associations between stressful life-event scores and probable PIU among Chinese college students, and to determine whether these associations remained after adjustment for potential confounders.
METHODS: This cross-sectional study used convenience sampling to recruit students from 8 higher education institutions, including comprehensive universities, medical universities, and vocational colleges, in Shanghai, Jiangsu, and Zhejiang, China, between March 1 and June 30, 2025. Stressful life events during the preceding 12 months were assessed using the 26-item Adolescent Self-Rating Life Events Checklist (ASLEC). Probable PIU was screened using the 8-item Internet Addiction Diagnostic Questionnaire (IADQ), with a score of 5 and above indicating probable PIU. Descriptive statistics, independent-samples 2-tailed t tests, chi-square tests, multivariable logistic regression, and restricted cubic spline (RCS) analyses were conducted. The primary analyses used a complete-case sample of 2805 participants.
RESULTS: Among the 2805 participants, 436 (15.5%, 95% CI 14.3%-16.9%) met the screening criterion for probable PIU. Participants with probable PIU had higher mean ASLEC scores than those without probable PIU (30.26, SD 28.10 vs 17.75, SD 20.84; P<.001). In the fully adjusted model, each 1-point increase in the ASLEC score was associated with 2% higher odds of probable PIU (adjusted odds ratio 1.02, 95% CI 1.01-1.02; P<.001). RCS analysis showed significant overall and nonlinear associations (both P<.001). The estimated odds changed little below a score of approximately 14, increased markedly between approximately 14 and 60, and appeared to plateau above 60. Because relatively few participants scored above 60, estimates in this range were less precise.
CONCLUSIONS: This study is innovative in applying RCS analysis to characterize the nonlinear dose-response association between stressful life-event scores and probable PIU. Unlike previous studies based mainly on conventional linear models, it identified potential inflection points at approximate scores of 14 and 60 and provided a more detailed account of how the odds of probable PIU varied across exposure levels. These findings extend the existing evidence and may provide preliminary information for developing tiered screening, stress-management support, and referral pathways in university mental health services. However, the identified values are exploratory statistical points rather than validated clinical cutoffs. Longitudinal and intervention studies are needed to confirm their practical use and determine whether stress-reduction and coping-focused interventions can reduce the onset or persistence of probable PIU.
PMID:42709974 | DOI:10.2196/92310
