The 40% Deficiency That Wasn't: Micronutrient Gap Data Quality Audit

A district nutrition survey reports 40% of children aged 1–5 have below-RDA intake for iron, vitamin A, and zinc. The story is set to run in the local press. But the analyst notices the survey used two different portion-size methods across two zones — and one method systematically underestimates intake. She must audit the data quality before the headline goes to print.

# Aplly.xyz Case Study Submission

## Title
The 40% Deficiency That Wasn't: Micronutrient Gap Data Quality Audit

## Type
Data Analytics

## Difficulty
Advanced

## Estimated Time
75 minutes

## Overview
A district nutrition survey reports 40% of children aged 1–5 have below-RDA intake for iron, vitamin A, and zinc. The story is set to run in the local press. But the analyst notices the survey used two different portion-size methods across two zones — and one method systematically underestimates intake. She must audit the data quality before the headline goes to print.

## Case Details

Function Focus: Data quality audit, measurement method bias detection, survey methodology reconciliation, public health decision-making under data uncertainty

Scenario:
The district ICDS cell has preliminary survey results: 40% of children (1–5 years) show below 70% of RDA for three critical micronutrients — iron, vitamin A, and zinc. The District Magistrate wants to announce a supplementation programme. But the data analyst notices a footnote in the methodology: Zone North used standardized cup-and-scale measurement (Method A, considered gold standard), while Zone South used visual estimation from photo booklets (Method B, known to underestimate by roughly 20% for micronutrients). The raw numbers for both zones look similar, but if Method B systematically undercounts, the true prevalence may be far lower. The analyst must separate measurement artifact from genuine deficiency before the announcement.

Dataset Structure:
- Two zones (North and South), equal sample size
- Each zone used a different portion-size estimation method
- Per-child estimated daily intake for iron, vitamin A, and zinc (group averages by method)
- Known calibration factor for Method B based on a validation sub-study (20% underestimate, ±5% variance)
- RDA thresholds for each micronutrient

Tasks:
1. Compute the observed prevalence of deficiency (< 70% RDA) for each micronutrient in each zone using the raw (uncorrected) data — report the overall 40% figure
2. Apply the Method B correction factor to Zone South's data — recompute the per-zone and overall deficiency prevalence for each micronutrient after correction
3. Determine whether the 40% headline survives the correction — if not, compute the corrected headline number and state whether the supplementation programme still justifies an emergency announcement
4. Identify what additional data would be needed to confirm the true prevalence with confidence (beyond the available correction factor)
5. Only after submitting your manual analysis, re-run the corrected calculations using a spreadsheet or AI tool and note whether any judgement calls in your manual correction path differ from the computed result

Expected Output:
A data quality audit memo containing: (a) raw prevalence table by zone and micronutrient, (b) corrected prevalence table with Method B adjustment, (c) a before-and-after headline comparison, (d) a recommendation on whether to proceed with the announcement, and (e) specific requests for additional evidence to close the confidence gap.

Evaluation Criteria:
Correct application of the correction factor (applied to Zone South only, not North), correct recomputation of overall prevalence after correction, honest assessment of whether the emergency announcement is justified post-correction, and specificity in identifying missing evidence.

## Data Sources

RDA Thresholds (for children 1–5 years, per day):
| Micronutrient | RDA | 70% of RDA (deficiency threshold) |
|---|---|---|
| Iron | 9 mg | 6.3 mg |
| Vitamin A | 400 mcg RAE | 280 mcg RAE |
| Zinc | 5 mg | 3.5 mg |

Observed Average Daily Intake by Zone:
| Zone | Method | Sample Size | Avg Iron (mg) | Avg Vit A (mcg RAE) | Avg Zinc (mg) |
|---|---|---|---|---|---|
| North | A (Standardized) | 500 | 7.2 | 310 | 4.1 |
| South | B (Visual Photo) | 500 | 5.9 | 265 | 3.4 |
| Combined | | 1,000 | 6.55 | 287.5 | 3.75 |

Distribution Data (estimated from survey — percentage of children below 70% RDA by zone):
| Micronutrient | Zone North (% below threshold) | Zone South (% below threshold) |
|---|---|---|
| Iron | 28% | 52% |
| Vitamin A | 32% | 48% |
| Zinc | 30% | 50% |

Method Validation Sub-Study:
A concurrent sub-study of 100 children in Zone South measured the same meals using both Method B (visual photo) and Method A (standardized cup-and-scale). Results:
- Method B consistently reads 20% lower than Method A for micronutrients across all three measured nutrients
- Variance: ±5 percentage points (95% of paired comparisons fall between 15% and 25% underestimate)
- Correction factor to apply: multiply Method B values by 1.25 (i.e., divide the observed prevalence by... this requires reasoning about how prevalence, not just mean, shifts)

Context:
- Press conference scheduled for 48 hours
- Supplementation programme costs ₹2.4 crore for the district
- Previous year's survey (using Method A only, entire district) found 26% prevalence

## Solution Frameworks
Data quality auditing, measurement method bias detection, survey reconciliation, prevalence rate correction under known bias

## Solver Guidance & Tutorials
Link to: "Data Quality Audits for Nutrition Surveys" tutorial

## What You'll Learn
- Detecting when measurement methodology drives survey results
- Applying known correction factors to prevalence estimates (not just means)
- Making public health recommendations under data uncertainty
- Resisting pressure to announce before the data is validated

## Tags
micronutrients, data quality, nutrition survey, measurement bias, public health

## Registration Links
- Register as Solver
- Register as Evaluator

RDA Thresholds (for children 1–5 years, per day):
| Micronutrient | RDA | 70% of RDA (deficiency threshold) |
|---|---|---|
| Iron | 9 mg | 6.3 mg |
| Vitamin A | 400 mcg RAE | 280 mcg RAE |
| Zinc | 5 mg | 3.5 mg |

Observed Average Daily Intake by Zone:
| Zone | Method | Sample Size | Avg Iron (mg) | Avg Vit A (mcg RAE) | Avg Zinc (mg) |
|---|---|---|---|---|---|
| North | A (Standardized) | 500 | 7.2 | 310 | 4.1 |
| South | B (Visual Photo) | 500 | 5.9 | 265 | 3.4 |
| Combined | | 1,000 | 6.55 | 287.5 | 3.75 |

Distribution Data (estimated from survey — percentage of children below 70% RDA by zone):
| Micronutrient | Zone North (% below threshold) | Zone South (% below threshold) |
|---|---|---|
| Iron | 28% | 52% |
| Vitamin A | 32% | 48% |
| Zinc | 30% | 50% |

Method Validation Sub-Study:
A concurrent sub-study of 100 children in Zone South measured the same meals using both Method B (visual photo) and Method A (standardized cup-and-scale). Results:
- Method B consistently reads 20% lower than Method A for micronutrients across all three measured nutrients
- Variance: ±5 percentage points (95% of paired comparisons fall between 15% and 25% underestimate)
- Correction factor to apply: multiply Method B values by 1.25 (i.e., divide the observed prevalence by... this requires reasoning about how prevalence, not just mean, shifts)

Context:
- Press conference scheduled for 48 hours
- Supplementation programme costs ₹2.4 crore for the district
- Previous year's survey (using Method A only, entire district) found 26% prevalence

Data quality auditing, measurement method bias detection, survey reconciliation, prevalence rate correction under known bias