Epidemiological framework · Version 1.0

Methodology & Sources

Three evidence-based pathways translate CMIP6 climate projections into neonatal mortality estimates. All coefficients are derived from peer-reviewed literature; no proprietary data are used.

No projection loaded — run a city projection to see live calculated values.

4-tab spreadsheet: README · Control Panel · Core Engine · Constants & Provenance

Pathway 1 — Direct Neonatal Hyperthermia

— Threshold 32°C ▾

Neonates lack functional eccrine (sweat) glands for the first 2–3 weeks of life and cannot vasodilate effectively, making them almost entirely dependent on the ambient thermal environment. Above 32°C, the risk of hyperthermia escalates rapidly.

Deaths = Births × (NMR/1000) × AF × (HW_days/365) where: AF = (RR − 1) / RR (attributable fraction) RR = exp(β × ΔT) (relative risk) ΔT = max(0, Tx5d − 32°C) (temperature excess) β = 0.1133 (Chersich et al. 2020, SSA pooled)

Peak Temp (Tx5d)

—

°C · 5-day rolling max

ΔT (excess above 32°C)

—

°C

Relative Risk (RR)

—

exp(0.1133 × ΔT)

Attributable Fraction

—

(RR−1)/RR

Heatwave Days

—

days/yr ≥3-day runs above 32°C

Excess Deaths

—

neonatal/yr attributable

The beta coefficient (0.1133) is the sub-Saharan Africa pooled estimate from Chersich et al. (2020) — deliberately steeper than the adult all-cause coefficient (0.0801, Gasparrini 2017) to reflect neonatal thermal vulnerability.

Pathway 2 — Heat-Induced Preterm Birth

— Threshold 29°C ▾

Maternal heat stress elevates serum cortisol and oxytocin, triggering myometrial contractions and early cervical ripening. Excess preterm births then face a substantially elevated neonatal mortality rate (150 per 1000 live preterm births, Blencowe et al. 2013).

PTB_excess = Births × PTB_rate × (RR_ptb − 1) × (HW_days/365) Deaths = PTB_excess × (PTB_NMR / 1000) where: RR_ptb = exp(β_ptb × ΔT_ptb) β_ptb = ln(1.05) = 0.0488 (Chersich 2020) ΔT_ptb = max(0, Tx5d − 29°C) PTB_rate = 0.12 (12% baseline, WHO SSA 2023) PTB_NMR = 150/1000 (Blencowe et al. 2013)

ΔT (excess above 29°C)

—

°C

RR Preterm

—

exp(0.0488 × ΔT)

Excess PTB Events

—

extra preterm births/yr

PTB Deaths

—

neonatal/yr via prematurity

Pathway 3 — Excess Stillbirths

— Reported separately ▾

Stillbirths are tracked separately — they do not contribute to the neonatal mortality count but represent an important component of the total reproductive burden of extreme heat.

SB_excess = Births × SBR × (RR_sb − 1) × (HW_days/365) where: RR_sb = exp(β_sb × ΔT_ptb) β_sb = ln(1.08) = 0.0770 (Chersich 2020) SBR = 0.024 (24/1000, WHO GHO 2021)

RR Stillbirth

—

exp(0.0770 × ΔT)

Excess Stillbirths

—

events/yr above baseline

Summary Totals

— DALYs ▾

Total Neonatal Deaths

—

heat-attributable/yr

Excess Stillbirths

—

reported separately

DALYs

—

28 life-years/neonatal death

Climate Data — Open-Meteo CMIP6

IPCC AR6 ▾

Temperature projections are fetched from the Open-Meteo Climate API using CMIP6 model ensembles. Two model sets are used depending on the target year:

2015–2039: MRI_AGCM3_2_S (~20 km) · EC_Earth3P_HR (~50 km)
2040–2100: MPI_ESM1_2_XR (~50 km) · FGOALS_f3_H (~50 km)

All models are selected for sub-Saharan Africa performance per IPCC AR6 Chapter 10. A 10-year window centred on the target year is averaged to reduce inter-annual variability.

Tx5d

—

max 5-day rolling mean (°C)

Heatwave Days

—

days/yr above 32°C in ≥3-day runs

Climate Source

—

model set

Mitigation Physics

— Santamouris 2015 ▾

Cooling effects of urban interventions are modelled using coefficients from Santamouris (2015), a meta-analysis of 169 studies on urban heat island mitigation. Mitigation is applied to the peak temperature before running the mortality model.

Every 10% increase in canopy cover → ~1.2°C peak temperature reduction
Every 10% cool roof deployment → ~0.8°C peak temperature reduction
Heatwave days reduce proportionally at ~5% per °C of cooling

Canopy Cover

—

% target coverage

Cool Roofs

—

% deployment

Temp Reduction

—

°C from interventions

Lives Saved

—

neonatal deaths averted/yr

Limitations & Caveats

▾

Research estimates only. These projections are designed for analytical and planning purposes — not for clinical decision-making or epidemiological reporting.

The Chersich framework assumes a log-linear exposure-response; actual dose-response may flatten or steepen at extreme temperatures.
Heatwave-day scaling (HW/365) is a simplification — heat also kills outside formal heatwave periods.
The preterm birth RR (1.05/°C) extrapolated over large temperature excesses may overestimate at extreme tails; treat worst-case scenarios as upper-bound estimates.
City population growth to the target year is not modelled — all projections use current demographic baselines.
Adaptation already underway (existing shade infrastructure) is not subtracted from the baseline.

References

▾

Source	Use in NeoHeat
Chersich MF et al. (2020) Int J Epidemiol 49(6):1997–2010	Neonatal β (0.1133), preterm birth RR (1.05/°C), stillbirth RR (1.08/°C) — SSA pooled
Blencowe H et al. (2013) Lancet 382(9890):417–425	Composite preterm NMR (150/1000 live preterm births)
WHO Global Health Observatory (2023)	Regional PTB rate baseline (12%), stillbirth rate baseline (24/1000)
Santamouris M (2015) Energy Build 98:119–124	Mitigation cooling coefficients (canopy 1.2°C/10%, cool roofs 0.8°C/10%)
Open-Meteo CMIP6 API (2024)	Temperature projections: MRI_AGCM3_2_S, EC_Earth3P_HR, MPI_ESM1_2_XR, FGOALS_f3_H
DHS Program (various years)	City-level NMR for 43 seeded African cities
UNICEF/WHO (2022)	Country-level NMR fallbacks for geocoded city search