Digital eye strain among library users: Prevalence and ROC-derived cut-off point for the DESRIL-27

Digital eye strain (DES), also known as computer vision syndrome, has become a growing public health issue among students due to prolonged screen exposure. Reliable, locally validated screening tools are essential to estimate prevalence and guide prevention. This study aimed to establish a receiver operating characteristic (ROC)-based cut-off for the Digital Eye Strain and Risk Level Questionnaire (DESRIL-27) against the Computer Vision Syndrome Questionnaire (CVS-Q) as the reference standard, and to estimate the prevalence of DES among university library users.

A cross-sectional study was conducted between June and September 2024 among users of Universiti Malaysia Sabah libraries. Participants completed the DESRIL-27 and CVS-Q via Microsoft Forms. Data were exported to Excel, cleaned, and analysed in RStudio (version 2025.05.1). Reliability was assessed using Cronbach's alpha. ROC analysis determined the optimal DESRIL-27 cut-off, and diagnostic performance was quantified using sensitivity, specificity, predictive values, and likelihood ratios. Prevalence was estimated based on the optimal threshold with 95% confidence intervals (CI). Ethical approval was obtained from the Medical Research Ethics Committee, Faculty of Medicine and Health Sciences, Universiti Malaysia Sabah (JKEtika 3/24 (8)).

A total of 277 participants were included (mean age 24.2 years, 67% female). Internal consistency was excellent (α=0.964 for symptoms; α=0.921 for risk factors). ROC analysis yielded an area under the curve of 0.982 (95% CI: 0.970, 0.994). The optimal cut-off was ≥14.5, demonstrating sensitivity of 90.9% and specificity of 96.7%, with a positive predictive value of 98.3% and a negative predictive value of 83.8%. The likelihood ratios were LR+ =27.6 and LR- =0.09. At this threshold, the prevalence of DES was 62.1% (95% CI: 56.1, 67.8).

The DESRIL-27 demonstrated excellent reliability and diagnostic performance, with an empirically derived cut-off of ≥14.5. These findings support its use as a screening tool for digital eye strain in university settings.