Currency:
Trust & Transparency

Data Methodology

How we collect, verify, normalise, and model medical cost data so you can compare hospitals across the globe with confidence.

Last updated: February 2026

Our 4-Step Data Pipeline

1

Collection

We collect procedure cost data from hospital billing offices, international patient coordinators, verified patient reports, published healthcare indices (WHO, OECD, local ministries), and licensed third-party data providers. Sources are logged and date-stamped for traceability.

2

Verification

Each data point is cross-referenced against at least one secondary source. Data points with no corroboration are either flagged as 'estimated' or excluded. Patient-submitted costs go through an additional photo verification and medical team review before being incorporated.

3

Normalisation

All costs are converted to USD using live exchange rates fetched daily from a reliable forex API. We apply purchasing power parity (PPP) adjustments for research publications, but display simple converted prices on public-facing pages for clarity.

4

Modelling & Refresh

Hospitals without direct pricing data receive AI-modelled estimates based on country cost indices, procedure complexity scores, and peer hospital pricing. These are clearly labelled '* Est. based on AI modeling'. Data is reviewed quarterly or when significant currency/market shifts are detected.

AI Safeguards — How We Prevent Hallucinations

Our AI content generation system uses Retrieval-Augmented Generation (RAG). This means the model never invents statistics — all numerical claims are pulled directly from our Postgres database and injected into the prompt context.

Database-bound figures

Cost ranges, accreditation counts, bed capacity — all sourced from structured records, not model weights.

Confidence thresholds

If a data field has insufficient verification confidence, the model is instructed to omit it rather than estimate.

Post-generation validation

Generated content is scanned for numeric patterns and cross-checked against source data before storage.

Human audit loops

Our editorial team reviews flagged pages monthly and after any major database refresh.

Coverage & Limitations

We currently cover 50+ countries and 200+ medical procedures. However, direct hospital records are only available for a subset of hospitals. The following labels appear on the site to indicate data quality:

Price Verified

Direct cost confirmed with hospital billing. Most reliable.

Cost Estimated

Derived from regional indices or peer hospital data using our modelling pipeline.

Contact for Pricing

No pricing data available. We recommend contacting the hospital directly.