Improving Performance of a Pediatric TeleMedicine and Medication Delivery Service through mHealth Technology

点击翻译按钮获取中文摘要

PROJECT SUMMARY / ABSTRACT Acute respiratory infection and diarrheal disease are the two leading causes of pediatric death between 1 month and 5 years of age globally. These common problems have low-cost treatments, but these treatments are most effective when administered early. This is difficult in resource-limited settings, especially at night. Based on over 5 years of formative NIH-funded implementation science research, our team has built and deployed a telemedicine and medication delivery service (TMDS) called MotoMeds to improve nighttime access to care for children. The rationale is that a TMDS generates a multiplier effect to reduce mortality and morbidity to a greater extent than the provision of in-person emergency medical services (EMS) alone. A TMDS differs from EMS in that it mobilizes resources from a centralized location and transports these resources to households. An EMS, on the other hand, identifies patients at households and transports the patients to centralized resources. EMS logistics, clinical guidelines and decision-support tools do not readily apply to a TMDS. Novel methods are needed to assure that TMDS models of healthcare delivery safely reach their potential. MotoMeds was launched in Haiti in 2019 as a pre-pilot focused on safety and logistical feasibility and configured for scale in a pilot study in 2021. We developed paper-based decision-support tools for virtual call center exams that were derived from in-person World Health Organization (WHO) guidelines. Our research exposed a critical need for electronic clinical decision support (eCDS) to assure guideline adherence at scale and a need to confidently assign triage levels despite limitations in virtual telemedicine environments. Accurate triage is essential to determine TMDS pathways of care: delivery alone (mild cases), medication delivery plus an in-person exam (moderate cases) and EMS/hospital referral (severe cases). In the R21, we will address these needs by concurrently 1) designing an alpha prototype of the eCDS tool through human centered design followed by qualitative research analysis, while 2) using existing data from our prior two studies to derive and internally validate a disease severity prediction (DSP) algorithm for integration into the eCDS tool. In the R33, we will externally validate the accuracy of the DSP and evaluate the performance of a beta eCDS prototype in a pilot stepped wedge cluster randomized trial. The performance of the beta eCDS prototype will be determined by comparing rates of guideline adherence (primary outcome measure) and logistical metrics among providers using paper (control) vs electronic (intervention) decision support. This single-site pilot will provide essential experience and metrics for a future multi-site randomized controlled trial of the eCDS integrated with the DSP.