Smartphone-based Tele-Epigenetic Monitoring in Pediatric Chronic Illnesses

Thesis Abstract

Smartphone-based Tele-Epigenetic Monitoring in Pediatric Chronic Illnesses investigates how mobile health technologies can enable remote epigenetic profiling to inform personalized management in children with chronic conditions. The contextual problem addressed is the gap between fluctuating pediatric disease activity and infrequent in-clinic assessments, which limits timely interventions and may obscure dynamic epigenetic signatures associated with environmental, behavioral, and treatment-related factors. The study aims to determine (1) whether smartphone-delivered epigenetic sampling and remote data capture are feasible in pediatric populations, (2) the association between mobile-teleepigenetic metrics and clinical disease activity across common pediatric chronic illnesses (e.g., asthma, type 1 diabetes, juvenile idiopathic arthritis), and (3) the predictive value of integrated epigenetic biomarkers and digital phenotyping for short-term outcomes over a 12-month period. Specific objectives include assessing feasibility metrics (recruitment, retention, sample quality, data completeness), validating a low-burden cheek swab kit and compatible smartphone app for weekly epigenetic sampling, identifying epigenetic loci responsive to environmental and treatment variables, and developing a multivariate predictive model incorporating epigenetic, device-generated, and clinical data to forecast exacerbations or flares. A concurrent mixed-methods design is employed, combining a longitudinal cohort study with embedded qualitative interviews. The population consists of 280 children aged 6–16 years diagnosed with at least one chronic pediatric condition, recruited from three tertiary pediatric centers. A stratified sampling approach ensures representation across disease categories and socio-economic backgrounds. Data collection instruments include (a) a validated saliva/cheek swab kit for methylation profiling using targeted bisulfite sequencing, (b) a smartphone application for weekly self-reported symptoms, medication adherence, environmental exposures, and real-time physiological proxies (gps-based activity, sleep, heart rate variability), (c) a secure cloud platform for encrypted data transfer, and (d) standardized clinical assessments (e.g., Pediatric Asthma Control Questionnaire, HbA1c for diabetes, juvenile arthritis activity score). Epigenetic endpoints focus on methylation changes in a predefined panel of loci implicated in inflammation, stress response, and metabolism, complemented by genome-wide exploratory analyses to discover novel loci. Reliability and validity are ensured through pilot testing (n=40), calibration studies for sample collection, and cross-validation of epigenetic assays with a certified lab. Quantitative analyses apply longitudinal mixed-effects regression to examine associations between epigenetic marks, digital phenotypes, and clinical outcomes, adjusting for age, sex, socio-economic status, and concomitant therapies. Time-to-event analyses (Cox proportional hazards models) assess risk of flare or escalation events. Machine learning approaches, including LASSO-penalized Cox models and random forest classifiers, identify predictive feature sets for exacerbations, with model performance evaluated via area under the ROC curve, calibration plots, and external validation on a 60-subject holdout subset. Qualitative data from serial semi-structured interviews (n?40) are analyzed using thematic analysis to elucidate user experiences, barriers to engagement, and perceived relevance of epigenetic information in care decisions. Key anticipated findings include (i) high feasibility with >75% of participants completing ?80% of scheduled sampling, (ii) detectable longitudinal epigenetic shifts correlating with documented environmental exposures and treatment changes, and (iii) a robust predictive model integrating epigenetic signatures with digital phenotypes that improves forecasting of short-term clinical deterioration versus standard care alone. The study contributes to knowledge by bridging mobile health, epigenomics, and pediatric chronic disease management, offering empirical evidence on the feasibility, clinical utility, and ethical considerations of tele-epigenetic monitoring in children. It also informs guidelines for patient-centered data governance, pediatric consent processes for genomic-derived data, and scalable frameworks for integrating epigenetic monitoring into routine pediatric care. The main conclusion posits that smartphone-enabled tele-epigenetic monitoring can provide timely, objective biomarkers of disease activity and treatment response in children with chronic illnesses, enabling proactive, personalized interventions. Recommendations emphasize developing standardized protocols for sample collection in home settings, strengthening data privacy protections for minors, expanding panels to include additional disease-relevant pathways, and conducting multicenter trials to generalize findings across diverse pediatric populations.