Project Summary
Sepsis, defined as life-threatening organ dysfunction caused by a dysregulated host response to infection,
encompasses a continuum that ranges from sepsis to severe sepsis, septic shock, multiple organ dysfunction
syndrome (MODS) and eventually death if untreated. Sepsis is the leading cause of child mortality worldwide,
with most of these deaths occurring in low and middle-income countries (LMICs) yet few clinical tools have
been developed for identifying, monitoring, or managing septic children in LMICs. There is immense potential
for novel clinical tools that can help clinicians more rapidly identify children with advanced stages of
sepsis (severe sepsis, septic shock and MODS), who are at highest risk for decompensation and death.
Mobile health (mHealth) tools, wearable devices, and artificial intelligence techniques have rapidly proliferated
for a multitude of medical applications and could serve to bridge the gap in care of critically ill patients in
LMIC settings. By enabling the detection of subtle physiologic changes indicating clinical deterioration, these
tools may allow clinicians to intervene earlier, better direct care, and allocate scarce resources, all without the
need for advanced laboratory diagnostics or critical care infrastructure. Furthermore, remote monitoring
capabilities may also prove highly valuable in improving patient care and protecting the safety of healthcare
workers during times of infectious disease outbreaks such as from novel coronavirus 2019 (COVID-19).
This proposed research will develop a context-appropriate mHealth tool linking continuous physiologic
data obtained from a wearable device with a novel machine learning approach known as personalized
physiologic analytics (PPA) run on a standard smartphone to provide clinicians with accurate assessments
of sepsis severity and mortality risk in septic children admitted to the Dhaka Hospital of the International
Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b). Formative research among clinicians at icddr,b
will be used to develop this mHealth tool incorporating the PPA algorithm with a clinical decision support and
alert system for use by front-line clinicians. Finally, the tool’s feasibility, usability, and accuracy for detection of
sepsis severity and MODS will be validated in a new population of pediatric patients with sepsis.
Knowledge gained from this study will greatly advance the evidence base for the use of mHealth tools and
artificial intelligence techniques to help clinicians worldwide better care for critically ill children in LMIC settings
earlier in the course of their disease, thereby reducing morbidity and mortality from sepsis. The results of
this investigational research will be used to inform a multi-center clinical trial which would seek to assess the
impact of using this mHealth tool on clinical outcomes as well as the cost-effectiveness of this tool. This tool
may also provide an effective means of assessing patient responses to various therapeutic interventions via
continuous physiologic monitoring in future clinical trials. The proposed initiatives will also build a base of
technical and professional expertise at icddr,b in mHealth research capacity and user-centered design.
