How HRD works
Methodology
HRD maps consumer health claims to public research metadata. It is an evidence-discovery system, not a clinical guideline or automated medical verdict.
Pipeline
From source record to claim receipt
- Retrieve human-research candidates from PubMed and registered studies from ClinicalTrials.gov.
- Require intervention-specific terms in high-signal metadata.
- Classify study design from authoritative publication types, then methods-language heuristics.
- Match claims only when outcome terms appear inside an already-matched intervention topic.
- Publish counts, exact terms, dates and source links so readers can reproduce the result.
Research Receipt
What a count means
A Research Receipt is a reproducible count within HRD's current indexed snapshot. It is not a systematic review, risk-of-bias assessment or proof of efficacy. Records can overlap topics and claims.
Trials
Registry records are not results
Trial Radar reports registry status, dates and interventions. Registration does not show that a study succeeded, and estimated completion dates can change.
Limitations
Known boundaries
PubMed is not the whole literature. Metadata can be incomplete. Keyword matching can miss eligible records or include marginal ones. Claim answers are cautious editorial syntheses and should be reviewed as the corpus changes.