Background: Greater than 90% of the global 6.4 million deaths in
children under 14 years occur in resource-limited settings. Acute pediatric
critical illnesses – sepsis, pneumonia, trauma – are leading causes of death
and disability outside of the neonatal period, yet critical care services are
not universally available(1-4). A significant number of
these lives could be saved by
proven, simple critical care and supportive interventions(5-7) despite challenging environments and fewer pediatric
available critical care resources as compared to high-income countries(8). However, there is a gross disparity not only in
available pediatric critical care resources, but also of available data, and,
thus, evidenced-based guidelines, between high income counties (HIC’s) and low
and middle income countries (LMIC’s)(9). Without region-specific data that captures the
burden of disease, outcomes, and resource utilization of pediatric populations
in resource-limited settings, we cannot develop context-appropriate,
evidence-based interventions, or appropriately allocate limited but available
resources to hospitals. Even accepted practices such as aggressive fluid
resuscitation for patients in septic shock, which is standard of care in HIC’s,
can lead to increased mortality compared to controls in LMIC’s(10).
We propose to undertake a prospective, observational, multicenter, multinational point prevalence study to
measure the burden of acute pediatric critical illness in LMIC’s. Our proposed project is a crucial first step in
setting future research and health delivery priorities for LMIC’s.
Specific
Aim 1: Determine the etiology and prevalence of pediatric acute
critical illness amongst children presenting to participating hospitals in
resource-limited settings.
Specific Aim 2: Measure hospital outcomes (hospital mortality, length
of stay) in children with acute critical illness in resource-limited settings.
Specific Aim 3: Determine hospital resource utilization by children
with acute critical illness
Specific Aim 4: Determine the current resources available to provide
acute critical care across LMIC’s.
Study Design. We will leverage the Pediatric
Acute Lung Injury and Sepsis Investigators’ (PALISI) Research network to
recruit hospitals in LMIC’s in Central and Latin America, Africa, and South
Asia. Sites will include hospitals in Mali, Malawi, Tanzania,
Kenya, Pakistan and Peru and others
as enrolled. All sites will be asked
to fill out a survey assessing the above
aspects of resource availability, the presence of a basic research
infrastructure including ethical and/or IRB approval mechanisms, and the
availability of a local site PI. The study population is acutely ill or
injured children (aged 28 days to 14 years) presenting to an ED or directly
admitted to an inpatient unit at a participating hospital.
Inclusion criteria include:
1. Children (aged 28 days to 14 years)
2. Children evaluated in the ED for an acute illness
or injury or admitted to an inpatient unit at a participating study site, who
are discharged home after ED evaluation, who died in the ED or are transferred
to a higher level of care.
Exclusion criteria include: Children presenting for a follow-up visit,
vaccinations, suture removal (or other non-acute complaint) or children with a
corrected gestational age less than 42 weeks.
All pediatric patients will be screened upon
presentation to the emergency department (ED), or equivalent acute hospital
receiving unit, over a 24-hour time period on a given day, and children
admitted to the hospital, even if boarded in the ED, will be followed for
outcomes. Patients admitted to the hospital (general pediatric ward, high-dependency
unit [HDU], or intensive care unit [ICU]) will be followed daily to determine
resource utilization in the 24 hours preceding the time of discharge or death
to determine in-hospital outcomes. Participants will receive routine standard
of care per local standards and resource availability.
The primary outcome is prevalence of acute critical illness, defined as
death within 48 hours of presentation to the hospital, including ED mortality;
OR admission/transfer to an HDU or ICU; OR transfer to another institution with
a higher level of care; OR receiving critical care-level interventions (vasopressor infusion, invasive mechanical
ventilation, non-invasive mechanical ventilation) regardless of location in the
hospital.
Secondary outcomes include: etiology of critical illness, in-hospital mortality,
early mortality (death within 48 hours of presentation), cause of death,
resource utilization, and change in neurocognitive status from premorbid state
from admission to discharge
(Pediatric Overall Performance Category [POPC] and Pediatric Cerebral
Performance Category [PCPC]).
A member of the research team will screen the medical
records of all children presenting to the ED or direct admits to inpatient
units for inclusion criteria. Once included, study personnel will then follow
participating patients prospectively to determine in-hospital outcomes and
resource utilization by reviewing their medical records. The local research
team will collect data from the medical chart on the day of admission and then
once a day until death or discharge.
For Specific Aim 4, a seperate survey will be sent out to be
filled by physicians working in different hospitals across the world to
understand the resources available at these facilities (human resources, types
of beds, technologies, medication availability). This survey will not include
any patient information or identifiers. Data
collection will include: 1) patient characteristics; presenting symptoms;
severity of illness; anthropometrics (weight, height, mid-upper arm
circumference); comorbidities (HIV status, congenital heart disease,
malnutrition); presenting vital signs; available laboratory test results;
imaging results; and the POPC and PCPC prior to the current illness. 2)
hospital resource utilization: We will collect data on use of blood
transfusion, fluid bolus, vasoactive agents, non-invasive positive pressure
ventilation, oxygen, mechanical ventilation, ICU admission, and antibiotic
administration. 3)
outcomes: including discharge home, transfer to a higher level of care within
the hospital, transfer to another hospital, and death; final hospital
diagnosis, length of stay and documented cause of death (if applicable), and
the POPC and PCPC at the time of discharge. 4) Hospital Characteristics: including average number
of patient encounters and admissions, types of inpatient units, human
resources, infrastructure including healthcare devices, medications and laboratory
availability Data
Analysis: Data will be analyzed on all subjects who meet inclusion criteria. Once the
selection period ends, we will conduct power calculations to assess the minimal
detectable differences at sufficient powers with a minimum of 80%. Once data is
collected we will use statistical software (R, STATA) to first conduct
univariate descriptive analyses to summarize population-level information. We
will test for an association between patient characteristics and outcomes with
bivariate analyses. We will fit a mixed effects model to explore risk factors
associated with critical illness (primary outcome), in-hospital mortality and
early mortality (secondary outcomes). The mixed effects will account for
clustering (facility-level, region-level) time-varying variables, and
time-invariant variables. Variables will be chosen for inclusion in the
multivariate models based on their empirical significance in the literature
(age, sex, severity of illness, HIV status, anemia, malnutrition), and based on
their significance at the bivariate level to account for cohort-specific
variation. The final model with the best fit will be evaluated through
goodness-of-fit tests for nested models and Bayesian Information Criterion for
non-nested models. Duration of study: The study duration is over
two years. This includes site recruitment, pilot testing of data collection
tool, IRB approval and data collection at four separate time points. |