Development of a Pediatric Telemedicine Triage and Referral System in Primary Care

Thesis Abstract

This study addresses the persistent inefficiencies and accessibility barriers in pediatric primary care by developing a telemedicine triage and referral system designed to optimize initial assessment, risk stratification, and appropriate escalation to higher levels of care. The aim is to design, implement, and evaluate a scalable telemedicine triage workflow that integrates with existing electronic health records to reduce unnecessary in-person visits, shorten wait times, and improve referral accuracy and timeliness for common pediatric presentations. Specific objectives are (1) to map current triage pathways and identify bottlenecks in pediatric primary care; (2) to co-design a telemedicine triage protocol with clinicians, parents, and information technology specialists; (3) to implement a decision-support–driven triage module incorporating evidence-based red flags and risk stratification using a Bayesian network; (4) to evaluate usability, acceptance, and adoption among primary care providers and families; and (5) to assess impact on referral appropriateness, wait times, and short-term health outcomes over a 12-month period. A mixed-methods design is employed, combining formative, pilot, and evaluative phases under the pragmatism paradigm and guided by the Technology Acceptance Model and Diffusion of Innovations theory. The study setting is three urban primary care clinics serving diverse pediatric populations. The population includes pediatric patients aged 0–18 years, their caregivers, and clinicians (nurse practitioners, pediatricians, and physicians’ assistants). A purposive sample of 150 caregivers and 40 clinicians participates in usability and acceptability assessments, while a larger implementation cohort comprises approximately 2,500 pediatric encounters across 18 months. Data collection instruments include (1) structured clinician and caregiver surveys assessing perceived usefulness, ease of use, and intention to use; (2) semi-structured interviews with 30 clinicians and 30 caregivers for thematic insights; (3) system usage analytics capturing triage decisions, referral outcomes, and wait times; (4) a clinical audit checklist documenting referral appropriateness against established pediatric guidelines; and (5) patient health outcomes extracted from electronic health records. Validity and reliability are ensured through pilot testing of instruments (Cronbach’s alpha >0.70 for scales), triangulation across data sources, and inter-rater reliability checks (Cohen’s kappa >0.6) for chart abstractions. Data analysis includes descriptive statistics and inferential techniques logistic regression to identify predictors of appropriate referral decisions, ANOVA to compare wait times pre- and post-implementation, and time-to-referral analyses using Cox proportional hazards models. The triage algorithm continuity is evaluated through receiver operating characteristic (ROC) curve analysis to balance sensitivity and specificity in red-flag detection. Qualitative data are analyzed via thematic analysis following Braun and Clarke, with coding triangulated by two independent researchers. Key expected findings include (i) high acceptance and perceived usefulness among clinicians (mean system usability score >70 on SUS) and caregivers (net promoter score improvement from baseline to post-implementation); (ii) a statistically significant reduction in non-urgent in-person visits by 25–30% and a decrease in inappropriate referrals by 15–20%; (iii) shorter median wait times for urgent appointments by 40–60% and improved equity in access among socioeconomically disadvantaged families; (iv) enhanced clinician confidence in triage decisions due to the Bayesian network’s probabilistic risk estimates; and (v) favorable integration with existing electronic health record systems without compromising data security or workflow efficiency. The study contributes to knowledge by advancing an evidence-based framework for pediatric telemedicine triage that combines clinical decision support with user-centered design and robust implementation science. It offers a replicable model for primary care settings aiming to optimize pediatric access, ensure timely and appropriate referrals, and reduce system burdens. The main conclusion is that a well-designed telemedicine triage and referral system can improve care efficiency and equity without compromising safety when grounded in validated risk stratification, participatory design, and rigorous evaluation. Recommendations include scaling the system to additional sites with tailored local adaptations, ongoing training and governance for triage protocols, continuous monitoring of referral outcomes, and integration of patient-reported outcome measures to further refine the model.