FASTR Denominator Calculation & Coverage Estimation

Rapid-Cycle Analytics for HMIS-Based Service Coverage Monitoring

PART 1: Denominator Calculation and Selection

Harmonize multiple data sources to calculate and compare denominator options

1 Gather data sources
HMIS Data
(Module 2)
Survey Data
(MICS/DHS)
Population Data
(UN WPP)
↓
2 Process and harmonize data

Data preparation steps

  • Aggregate monthly data to annual level
  • Prioritize DHS data over MICS where both available
  • Forward-fill survey data gaps
  • Filter population estimates to country level
  • Align all datasets by time period
↓
3 Calculate multiple denominator options

Denominator options by source

HMIS-Based:
• From ANC1 contacts
• From Deliveries
• From SBA procedures
• From BCG vaccinations
• From Penta1 doses
Population-Based:
• From birth rate estimates
• From total population
• From under-1 population
• With demographic adjustments
↓
4 Calculate coverage for each denominator

Coverage calculation formula

  • Coverage = (Service Volume / Denominator) × 100
  • Apply demographic adjustments to each denominator option
  • Generate coverage estimates for all service indicators
↓
5 Compare each coverage to survey benchmark

Error calculation

Error Metric:
Error = (HMIS Coverage - Survey Coverage)²
Evaluation:
Identify denominator with minimum error
↓
6 Select best denominator

Selection criteria

  • ✓ Primary: Minimum error compared to survey
  • ✓ Secondary: Prefer HMIS-based over population-based
  • ✓ Tertiary: Data quality and completeness considerations
↓
7 Save Part 1 outputs

Output files

  • M4_denominators: All calculated denominator options and coverage estimates
  • M4_combined_results: Coverage values + best denominator selection by indicator
  • M4_selected_denominator_per_indicator: Summary of which denominator was selected as "best" for each indicator

PART 2: Denominator Selection and Coverage Projection

Select denominators and project survey estimates forward using HMIS trends

1 User configures denominator selection

Configuration options

  • Use "best" denominators (automatically selected from Part 1)
  • Override with custom denominator specifications per indicator
  • Specify different denominators for different health service areas
↓
2 Filter Part 1 results to selected denominators

Data subsetting

  • Extract selected denominator option for each indicator
  • Retain associated coverage calculations
  • Prepare data for projection analysis
↓
3 Calculate year-over-year coverage deltas

Delta calculation

Formula:
Δ Coverage = Coverage(t) - Coverage(t-1)
Purpose:
Identify trends in service delivery over time
↓
4 Identify most recent survey as baseline

Baseline selection

  • Locate most recent DHS or MICS survey
  • Extract survey coverage estimates for all indicators
  • Use as reference point for projection
↓
5 Project survey estimates forward

Projection formula

  • Projected Coverage = Last Survey + (HMIS_now - HMIS_then)
  • Uses HMIS trend to update survey baseline forward
  • Bridges gap between survey year and current period
↓
6 Combine final results

Result integration

• HMIS-based coverage estimates
• Original survey values
• Projected survey coverage
• By administrative level
↓
7 Save final outputs

Final deliverable

  • M5_coverage_estimation_national.csv: National-level coverage estimates including HMIS-based values, survey data, and forward projections
  • M5_coverage_estimation_admin2.csv: Admin area level 2 coverage estimates
  • M5_coverage_estimation_admin3.csv: Admin area level 3 coverage estimates