A pattern can justify a low-risk experiment or a question for a professional. It does not prove a diagnosis, mechanism, treatment response, or peptide selection.
Six labels PEPRA keeps separate
- Direct measurement
- Device estimate
- Self-report
- Descriptive personal correlation
- Model-generated hypothesis
- Clinician-confirmed conclusion
Each future insight should expose provenance, time window, missing data, confidence, alternative explanations, evidence strength, and whether the source permits AI processing.
The Evidence Passport
PEPRA will not publish one blended confidence score. Each insight keeps five independent facets visible:
- Measurement (M0–M3): how directly and reliably the source measures the signal.
- Population evidence (E0–E4): the strength and relevance of research beyond one person.
- Personal evidence (P0–P3): repeated observations or prespecified experiments in the member's own history.
- Applicability (X0–X3): fit to ancestry, population, device version, age, and context.
- Actionability (A0–A3): whether the output is only educational, supports a low-risk experiment, or requires a qualified professional.
Additional sources can change the available context; they do not establish accuracy, causation, intervention benefit, clinical utility, or authority to provide care.
Current evidence appendix
Digital measurement validation
A wearable measure should be verified, analytically validated, and clinically validated for its intended purpose before usefulness is inferred from it.
Wearable sleep
Consumer devices can estimate sleep and wake patterns, but device performance varies and stage agreement with polysomnography is inconsistent.
Activity trackers
Tracker-supported programs can increase activity on average, while physiological effects are generally small and only occasionally statistically significant.
HRV-guided training
HRV-guided training has not consistently produced better fitness or performance outcomes than predefined training.
Personalized nutrition
Food4Me showed modest dietary-behavior changes; genotype did not improve effectiveness beyond diet-based personalization or establish clinical outcomes.
Polygenic scores
Performance depends on the populations and reference data used, and predictive accuracy can decline across ancestry groups.
30× whole-genome data
Supported whole-genome files must pass source, format and quality checks. Potentially important findings require appropriate clinical confirmation before medical use.
Personal experiments
The CENT reporting guideline calls for transparent reporting of the design, periods, outcomes, harms, carryover, and analysis in clinical N-of-1 trials. A simple personal before/after comparison is not an N-of-1 trial and does not establish cause. PEPRA's current synthetic observation plans are demonstrations, not clinical trials.
Genetics and intervention boundary
Genetic association is context, not destiny. PEPRA would not turn a consumer variant, polygenic score, wearable score, or questionnaire response into an independent medication or peptide selection. Potentially important findings require an appropriate confirmation and clinical pathway before medical use.
As of PEPRA's July 13, 2026 evidence review, PEPRA has not identified a validated evidence chain establishing that consumer genomic, wearable, and questionnaire data can safely select a peptide for an individual. PEPRA therefore separates educational analysis from any future independently controlled clinical decision.
Editorial controls
Evidence records should include DOI, access status, population, endpoint, effect size when relevant, limitations, correction or retraction status, last review date, reviewer, and next review date. Material claims must be revised when the underlying evidence changes.
Evidence review date for this preview: July 13, 2026.