Evidence Library

Scientific References

The research behind Metabolicum's evidence-based approach

How We Evaluate Evidence

Not all research is equal. We assign confidence grades to each source based on study design, sample size, replication status, and methodology. This helps you understand how confident you can be in each finding.

Click any grade to learn more

Grade A

High Confidence

Replicated findings across multiple well-designed studies

Grade B

Good Confidence

Well-designed single studies with strong methodology

Grade C

Moderate Confidence

Observational data with consistent patterns

Grade D

Emerging Evidence

Mechanistic or theoretical - interpret cautiously

Grade E

Clinical Consensus

Practitioner experience without formal trials

Understanding the Reproducibility Crisis

Why are we so careful about evidence grading? Because a significant portion of published research fails to replicate.

Study	Finding	Grade
Begley & Ellis, 2012	Only 21% of landmark cancer studies could be replicated	A
Open Science Collaboration, 2015	Only 36% of psychology studies replicated	A
Ioannidis, 2005	Theoretical framework explaining why most findings are false	A

This doesn't mean you should distrust all research. It means:

•Single studies warrant caution — Replication matters
•Peer review is necessary but not sufficient — Many peer-reviewed findings are later overturned
•Effect size matters more than p-values — Statistical significance ≠ practical importance
•Longer studies are generally more reliable — Short-term biomarker changes may not reflect long-term outcomes

We grade evidence honestly so you can calibrate your confidence appropriately.

Research Library

In-depth analysis of foundational studies with translated summaries

Showing studies cited on: Kraft Patterns

Topic

Paradigm

Evidence Grade

Study Type

Cited On

Showing 2 studies

Cohort Study

Crofts 2016: Hyperinsulinemia in the Kraft Database

Crofts et al. • Diabetes Research and Clinical Practice • 2016

Over 50% with normal glucose have hyperinsulinemia - earliest metabolic disease marker

Insulin Resistancekraft patternshyperinsulinemia+1 more

Cohort Study

Hayashi 2013: Insulin Patterns During OGTT Predict Diabetes Risk

Hayashi et al. • Diabetes Care • 2013

Pattern 4 (120-min peak, lower at 30 min): 47.8% diabetes incidence, OR 12.55 (95% CI 4.79-32.89) vs early peak patterns; Pattern 5: 37.5% incidence, OR 8.34

Insulin ResistanceDiabeteskraft patterns

How We Use References

In Calculators

Primary thresholds come from Grade A sources. When Grade A evidence suggests a range rather than a single value, we present the range with context.

In Educational Content

We draw on Grades A-C for educational claims, clearly labeling evidence quality. Grade D and E content is marked as theoretical or practitioner-derived.

In Paradigm Frameworks

Standard Medical: Grade A (established guidelines)
Research Consensus: Grade A-B (outcome research)
Metabolic Optimization: Grade B-E (intervention research + clinical consensus)

Evidence Grades

Grade A: High ConfidenceReplicated findings across multiple studies Grade B: Good ConfidenceWell-designed single studies Grade C: Moderate ConfidenceObservational data with consistent patterns Grade D: Emerging EvidenceMechanistic or theoretical Grade E: Clinical ConsensusPractitioner consensus

Other Pages

Our MethodologyHow we approach health information Four Clinical FrameworksWhy we show multiple interpretations

Our Methodology

This bibliography contains the research supporting Metabolicum's evidence-based approach.

Last comprehensive review: December 2025