Header
- Model name
- Metabolic Syndrome
- Developer
- Zenlo LLC
- Release stage
- Research Tool (not FDA-cleared)
- Version
- 1.0
- Availability
- United States
- Regulatory status
- Not applicable — academic and transparency positioning
- Pattern slug
- metabolic_syndrome
- Biomarkers
- GLU, INSULIN, TG, HDL, HBA1C
This document follows the CHAI Applied Model Card format (v0.1).
Summary
The Metabolic Syndrome pattern evaluates a cluster of metabolic laboratory abnormalities aligned with ATP-III–style criteria available from standard fasting panels: HOMA-IR–derived insulin resistance, elevated triglycerides, low HDL cholesterol, and dysglycemia (fasting glucose or HbA1c). A deterministic detector counts qualifying criteria; sufficient abnormalities fire the pattern. Claude Haiku 4.5 provides narrative synthesis for physician review. For licensed functional-medicine physicians, this is supportive clinical decision support — it does not measure waist circumference or blood pressure, which are required components of full ATP-III metabolic syndrome diagnosis. Outputs highlight metabolic clustering in laboratory data for clinical correlation, not standalone syndrome diagnosis.
Uses & Directions
Intended use
Clinical decision support for licensed physicians evaluating metabolic clustering in adult fasting laboratory panels.
Primary users
Licensed functional medicine physicians and similarly qualified clinicians.
How to use
Review flagged metabolic criteria with blood pressure, waist circumference, weight, and lifestyle context.
Target population
Adults aged 18 and older in the United States.
Out of scope
- Direct patient use without physician oversight
- Pediatric populations
- Standalone ATP-III metabolic syndrome diagnosis without clinical measures
- Type 1 diabetes management
Warnings
Clinical risk level
Low — supportive tool; the treating physician retains full clinical judgment and responsibility.
Known limitations
- NHANES audit used a 4-lab subset only — blood pressure and waist circumference were unavailable, so detected prevalence is lower than full ATP-III.
- HOMA-IR requires fasting state; non-fasting samples reduce reliability.
- HDL thresholds are sex-specific; missing sex defaults conservatively.
Trust Ingredients
AI system facts
- Deterministic pattern detector (metabolic_syndrome) plus Claude Haiku 4.5 for narrative synthesis
- Primary inputs: Fasting glucose, fasting insulin, triglycerides, HDL cholesterol, and HbA1c from structured extraction.
- Output: Pattern flag plus narrative on metabolic criteria met.
Security & compliance
- Anthropic Business Associate Agreement with zero-data-retention configuration
- HIPAA-aligned design; no patient data used for model training
Ongoing maintenance
Versioned, transparent, and reproducible via a public independent audit harness (see Resources).
Transparency
Self-funded development; no third-party sponsor for this pattern card.
Key Metrics
Usefulness / Efficacy
Zenlo's detection approach was benchmarked across five models in a separate study; see the medRxiv preprint in Resources. No per-pattern efficacy metric is published for this pattern.
Source: 5-model benchmark, medRxiv MEDRXIV/2026/346284
Fairness / Equity
Overall prevalence detected: 10.3% (615 / 5,992)
By sex
| Subgroup | n | Prevalence detected |
|---|---|---|
| Female | 3,105 | 9.6% |
| Male | 2,887 | 11.0% |
By age group
| Subgroup | n | Prevalence detected |
|---|---|---|
| 18–39 | 2,226 | 7.1% |
| 40–59 | 1,865 | 12.4% |
| 60+ | 1,901 | 11.8% |
By race / ethnicity
| Subgroup | n | Prevalence detected |
|---|---|---|
| Mexican-American | 1,064 | 12.7% |
| Other Hispanic | 798 | 12.3% |
| Non-Hispanic White | 1,914 | 10.9% |
| Non-Hispanic Black | 1,265 | 6.7% |
| Non-Hispanic Asian | 726 | 8.3% |
| Other / Multiracial | 225 | 12.4% |
This audit applied a 4-laboratory subset (fasting glucose, triglycerides, HDL, HbA1c) without blood pressure or waist circumference — two of five ATP-III criteria were unavailable in NHANES 2015–2016 for this run. Detected prevalence is therefore lower than full ATP-III metabolic syndrome rates in the literature.
These are descriptive population rates from the audit harness — not model error rates or benchmark accuracy metrics.
Source: Independent NHANES 2015–2016 audit harness (N=5,992, commit c10afe8)
Safety / Reliability
- Supportive-only; not intended as a standalone diagnostic
- Physician authorization required before clinical use
- Deterministic detector is reproducible for inputs: GLU, INSULIN, TG, HDL, HBA1C
Resources
- medRxiv preprint MEDRXIV/2026/346284 — 5-model benchmark (system-level detector evaluation)
- JAMIA Open submission JAMIO-2026-0120 (under review)
- Independent audit harness: github.com/dimashibakov/zenlo-audit — reproducibility manifest, NHANES 2015–2016 cycle, harness commit c10afe8
- HbA1c biomarker reference sheet — related biomarker-level NHANES distribution
Footer
Note: The mention or sharing of any examples, products, organizations, or individuals does not indicate any endorsement of those examples, products, organizations, or individuals by the Coalition for Health AI (CHAI).