← Home
Iodine Global Network · Namibia

Informing national fortification policy

Multi-source analysis of what Namibian households actually eat — by region and by supply chain — to turn a blanket fortification mandate into a targeted, evidence-based policy.

10×
regional spread in staple intake (wheat 30–292 g/c/d)
14
regions analysed
7
fortification vehicles assessed
199
food items harmonised into one FCT
Client
Iodine Global Network (IGN)
Role
Analysis & modelling
Data
NHIES 2015–16 · DHS 2000–2013 · 4 FCTs
Scale
14 regions · 7 vehicles
Timeline
Jul 2020 – Apr 2021
At a glance
The problem
Namibia committed to mandatory staple-food fortification — but setting standards needs evidence: which foods households eat, how much, through which supply chains, and how that varies across 14 regions.
What I built
A modular Python system over NHIES 2015–16, three DHS rounds, and four food-composition tables — computing weighted per-capita intake for seven candidate vehicles and a harmonised 199-item nutrient database.
The outcome
Regional fortification vehicle profiles showing ~10× variation in staple intake and >45% own-production share for some grains — evidence that a blanket commercial mandate alone would not reach everyone.
01 Challenge

You cannot set fortification standards on a national average

Namibia’s Food and Nutrition Security Policy committed the government to mandatory fortification of staple foods. But before setting standards, policymakers needed answers: which foods do households actually consume, in what quantities, and through which supply chains? How does that vary across 14 regions? And what is the baseline for nutrients — like iodine — where fortification is already underway?

The Iodine Global Network contracted A3DI to build the evidence around the 2015–16 National Household Income and Expenditure Survey (NHIES) — a 12-month, nationally representative survey with seven-day food diaries. The raw microdata had not been analysed at the detail needed for vehicle selection, dosage setting, or supply-chain targeting.

The real requirement

Region-level evidence for vehicle selection, dosage setting, and supply-chain targeting — not a single national figure that hides where the policy would and would not reach.

02 Approach

Three data sources, seven vehicles, fourteen regions

A3DI built a modular analysis system in Python across 15+ Jupyter notebooks, with a custom analytics module for weighted survey calculations (weighted means, weighted medians, and a generalised disaggregation function). The work ran in four interconnected phases.

Input
3 data sources
NHIES, DHS salt, 4 FCTs
Baseline
Grain consumption
own-production vs market
Compute
Intake
g/c/d × 7 vehicles × 14 regions
Harmonise
199-item FCT
→ vehicle profiles
Why the sourcing distinction matters

Only commercially processed grain can be fortified, so the pipeline tracked own-production vs market purchase, not just intake. Four food-composition tables (West African 2012, Kenyan, USDA SR24, South African 2017 — the last PDF-extracted with Camelot) were standardised into one 199-item dataset that fed FAO’s ADePT-Food Security Module for micronutrient adequacy modelling.

03 Result

Consumption varies up to tenfold across regions

Staple intake ranges across 14 regions

g/c/d, min–max on a common 0–460 scale
Wheat
30–292
Mahangu
4–452
Maize
67–338
Salt
2.8–7.5

Min–max across the 14 regions, straight from the NHIES analysis.

Wheat intake ranged from 30 to 292 g/c/d across regions; mahangu from 4 to 452; maize from 67 to 338. Own-production consumption exceeded 45% for certain grains in specific regions — meaning a substantial share of grain intake would not be reached by mandatory commercial fortification alone.

Salt consumption averaged 5.4 g/c/d nationally (median 4.9), ranging from 2.8 to 7.5 across regions — essential for setting dosage that delivers adequate micronutrients without exceeding safe limits. The harmonised Food Composition Table gave government and IGN a standardised, locally relevant nutrient database, delivered as reproducible Jupyter notebooks with Excel workbook outputs the team can re-run as new data arrives.

The takeaway

Fortification policy needs more than knowing what people eat — it needs how much, where, and through which supply chains.

Turning multi-source data into policy evidence?

I harmonise messy, multi-source survey and composition data into reproducible analysis your stakeholders can explore and update.