Salutogenic Approach to Early Kidney Health Promotion: Comparing Deep Learning Facial Analysis and Questionnaire Screening in Adolescents
Authors:
Open Access
- 9 January 2026
- journal article
- Published by Universitas Airlangga in Jurnal PROMKES
- Vol. 14 (SI1) , 50-58
Abstract
Background: Chronic kidney disease (CKD) is increasingly reported in adolescents and is driven by modifiable lifestyle factors such as inadequate hydration, frequent consumption of processed foods, obesity, and poor sleep quality. Because early CKD is typically asymptomatic, routine laboratory screening is often impractical in school or community settings. Self-reported questionnaires can help estimate lifestyle-related risk but rely on subjective recall. Objective: To compare the effectiveness of a DL-based facial analysis model with self-reported questionnaire screening for early CKD risk in adolescents. Methods: A cross-sectional study was conducted among 100 adolescents aged 16–19 in Karawang, Indonesia. Participants completed validated questionnaires assessing hydration, dietary habits, and sleep quality using the Pittsburgh Sleep Quality Index (PSQI). Standardized facial photographs were analyzed using a convolutional neural network (CNN) trained to detect CKD-related facial markers. Agreement between methods was assessed using Cohen’s kappa, and diagnostic performance was evaluated using ROC analysis. Results: Questionnaire screening classified 88% of adolescents as low-risk, 10% moderate-risk, and 2% high-risk, while the AI model classified 95%, 4%, and 1%, respectively, demonstrating moderate agreement (κ = 0.61, p < 0.001). The AI model achieved 91.0% accuracy, 88.0% sensitivity, 92.0% specificity, and an AUC of 0.904. Overweight adolescents had higher odds of being at risk (OR = 2.35). Conclusion: Combining AI-based facial analysis with questionnaire assessment provides a rapid, scalable, and non-invasive strategy for early CKD risk screening in adolescents, particularly in resource-limited settings.Keywords
This publication has 20 references indexed in Scilit:
- A noninvasive model for chronic kidney disease screening and common pathological type identification from retinal imagesNature Communications, 2025
- Advance care planning for patients with end-stage kidney disease on dialysis: narrative review of the current evidence, and future considerationsJournal Of Nephrology, 2024
- Cost-effectiveness of screening for chronic kidney disease in the general adult population: a systematic reviewClinical Kidney Journal, 2023
- Facial Depth and Normal Estimation Using Single Dual-Pixel Camera2022
- Factors associated with self-care behavior in patients with pre-dialysis or dialysis-dependent chronic kidney diseasePLOS ONE, 2022
- Artificial intelligence in learning and development: a systematic literature reviewEuropean Journal of Training and Development, 2022
- Machine Learning–Aided Chronic Kidney Disease Diagnosis Based on Ultrasound Imaging Integrated with Computer-Extracted Measurable FeaturesJournal of Digital Imaging, 2022
- Sample size and performance estimation for biomarker combinations based on pilot studies with small sample sizesCommunications in Statistics - Theory and Methods, 2020
- Validity and reliability of the Indonesian version of the Pittsburgh Sleep Quality Index in adolescentsInternational Journal of Nursing Practice, 2020
- Chronic kidney disease as a global public health problem: Approaches and initiatives – a position statement from Kidney Disease Improving Global OutcomesKidney International, 2007