Neonatal module

1.0 Overview 

The neonatal component of the MNCNH portfolio simulation is comprised of one single module. This module functions more similarly to standard vivarium simulations and is comprised of cause models, risk models, and intervention models (all of which are linked below).

2.0 Module Diagram and Data 

2.1 Module Inputs 

Module required inputs
Input	Source module	Application
Birth outcome	Pregnancy module	Only live births proceed through the neonatal module
Gestational age	Pregnancy module	LBWSG risk effects
Birth weight	Pregnancy module	LBWSG risk effects
Child sex	Pregnancy module	Neonatal mortality rates
Antenatal corticosteroid coverage	Intrapartum intervention module
RDS intervention propensity	Initial attributes module	Determines which simulants receive each RDS interventions (CPAP and ACS)
Hemoglobin exposure at birth	Hemoglobin module	Affects neonatal sepsis risk

2.2 Cause models 

Neonatal Mortality Model

2.3 Risk models 

The following risk factors affect neonatal causes:

Low birth weight and short gestation
- Risk exposure: informed as an output from the pregnancy module
- Risk effects: instructions for how to apply LBWSG risk effects are described on the Neonatal Mortality Model. Additional information can be found on the LBWSG risk effects document
Hemoglobin at end of pregnancy
- Risk exposure: informed as an output from the hemoglobin module
- Risk effects: affects neonatal sepsis according to the instructions on the hemoglobin risk effects document

2.4 Intervention models 

Antibiotics for treating bacterial infections
CPAP for treating Preterm with RDS
Neonatal probiotics
Antenatal corticosteroids (coverage assumed equal to CPAP coverage, see below)

2.5: Module Outputs 

See observer/outputs section on main concept model document.

3.0 Assumptions and limitations 

In GBD, LBWSG impacts all-cause mortality, which overlaps with the other neonatal causes. The method for handling this is complex, since preterm birth is a PAF-of-one cause, that we want to split into preterm with and without RDS, and other causes must have a RR with LBWSG to make the all-cause RR calibrate.
In this phase of model building, we are not including lung surfactant or kangaroo care which are closely tied to the CPAP/NICU intervention. We might add these to the model in a later phase.

4.0 Verification and Validation Criteria 

Confirm ACMR in sim matches ACMR in artifact
Confirm LBWSG exposure match
Confirm LBWSG RR on ACMR matches
Confirm CSMR matches for preterm, sepsis, encephalopathy
Confirm that RDS incidence and mortality match expectations
Confirm that interventions have expected efficacy and coverage rates