Innovative Monitoring in PAediatrics in Low-resource settings: an Aid to save lives? (IMPALA)
Loading...
Date
2022-05-19
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Kamuzu University of Health Sciences
Abstract
Type of Research:
An observational cohort study of 1000 children (28 days-60 months) admitted to the high dependency
areas of Queen Elizabeth Central Hospital (QECH) and Zomba Central Hospital (ZCH).
The problem:
The number of children dying in African hospitals remains too high. A large part may be prevented
if children can be observed more closely allowing timely life-saving treatments. Continuous
monitoring of vital signs such as heart rate and oxygen saturation is applied for this reason in highincome
countries but these techniques have not been adapted for low resource settings. New
techniques of monitoring may make it possible to predict potential deterioration (and not just detect),
these include new vital signs sensors, bedside blood tests (biomarkers) and artificial
intelligence/machine learning.
The objectives:
The study aims to evaluate the predictive potential of the machine learning algorithms among
critically ill Malawian children for improved healthcare using a continuous vital signs monitoring
system specifically improved for use in low-resource settings. Additional objectives include
assessing if biomarkers and demographic data can predict critical illness events alongside vital signs.
Furthermore we will evaluate assess if study site and age affects the composition and performance
of the predictive potential of the model. To confirm, explain and compare potential associations
identified by machine learning, conventional uni- and multivariate models will be used.
The methodology:
In this clinical observational study we will use the IMPALA monitoring system to gather the data
required to develop the predictive algorithms that will be incorporated into a new version of the
IMPALA system after this study. The data that will be collected will consist of vital signs (e.g. blood
pressure, heart rate and respiratory rate) of 1,000 children admitted to the high dependency areas of
two centres in Malawi (QECH and ZCH). Moreover, we will collect blood samples for bedside
detection of biomarkers (RNA or protein) that may predict critical illness and sepsis. Additionally,
anonymised sociodemographic data of the patients will be collected. These data and/or the biomarker
data will optionally be used to strengthen the predictive power of the algorithms.
Expected findings:
The clinical-study firstly aims to construct several risk predictive models for both outcome and
specific critical care actions using 1) continuously recorded vital signs alone and 2) including
biomarkers and sociodemographic data. Secondly the clinical-study aims to define the accuracy of
biomarkers predicting critical illness and bacterial sepsis.
The overall aim of the entire IMPALA project is to optimise monitoring of hospitalised children in
low resource settings by integrating an algorithm that predicts critical illness events allowing early
detection and treatment. The ultimate aim is to reduce amendable in-hospital deaths in children living
in low-resource settings.
Dissemination:
The project will be part of three PhD programmes and its results will be shared through open- access
peer-reviewed publications. Additionally, specific study results will be disseminated at the KUHeS
Research Dissemination Conference, the involved departments and institutes in and to particular
interest groups, both in meetings, workshops and conferences, nationally and internationally. All
publications will be shared with KUHES REC. Additionally, the IMPALA consortium and partners
will disseminate the findings through their websites and media channels.
Description
Keywords
Innovative monitoring in peadiatrics