How This Tool Works

The Women's Health Lab Intelligence tool uses a deterministic, rule-based inference engine. There is no AI. No guessing. No probability models. Every result can be traced to an explicit rule and threshold.

Zero AI Policy — This tool deliberately contains no LLM calls, no machine learning, and no generative AI. Interpretation is reproducible: the same lab values with the same context will always produce the same output.

The Threshold System

Each biomarker is classified into one of seven buckets based on numeric thresholds:

Critical Low — requires urgent clinical attention
Low — clinically deficient, intervention typically indicated
Low-Normal — suboptimal, below performance threshold
Optimal — within the evidence-based optimal range for women
High-Normal — elevated but not critical, warrants monitoring
High — above range, clinical evaluation indicated
Critical High — requires urgent attention

Thresholds are women-specific and, where evidence supports it, differ by reproductive status and age. Standard lab reference ranges are not used — those reflect statistical population norms, not performance optima.

The Rule Engine

Rules fire when one or more markers fall into specified buckets. Rules support three logic modes:

AND — all conditions must be true
OR — any condition triggers the rule
MIN_COUNT:N — at least N conditions must be true

When a rule fires, it adds weighted evidence toward symptoms. Compound rules (multi-marker patterns) apply a 1.25x weight multiplier — multi-marker patterns carry more diagnostic significance than single-marker flags.

A symptom appears in results only when its aggregated confidence reaches 20% or higher.

Inverted Markers

For markers where a lower value is worse, the bucket direction is reversed:

hs-CRP, Triglycerides, Homocysteine, HbA1c
FSH (Day 3), Prolactin, Reverse T3, Fasting Insulin
TPO Antibodies, TgAb

For these markers, HIGH means worse — the rule engine accounts for this automatically.

Derived Values

Three derived values are computed from entered markers if sufficient inputs are available:

HOMA-IR = (glucose × fasting insulin) / 405 — insulin resistance index
LH:FSH ratio = LH / FSH Day 3 — elevated ratio (>2) suggests PCOS pattern
FT3:rT3 ratio = Free T3 / Reverse T3 — below 0.20 suggests cellular hypothyroid pattern

Cycle Context

When Last Menstrual Period date is provided, cycle day is calculated and cycle-sensitive markers (estradiol, progesterone, LH, FSH) are interpreted against phase-expected ranges. Day 3 draws are distinguished from Day 21 draws. Progesterone below 10 ng/mL on Day 21 suggests anovulatory cycle or luteal phase defect.

Privacy Architecture

No lab values, no personal information, and no session data are ever transmitted to or stored on any server. The inference engine runs entirely in your browser. The only server call is a one-time load of reference data (thresholds, rules, marker definitions) from a read-only database.

Session continuity works through the report itself: session data is compressed, encoded, and embedded as a QR code and text block in the PDF. Returning users can upload a previous report to restore their session — no accounts, no cookies, no cloud storage.

Substance Interference Detection

If medications, supplements, or substances are reported in context, the tool cross-references a database of known interference patterns. Biotin interference (common in thyroid and other immunoassay tests) is flagged with a specific warning if biotin supplementation was not held for 72 hours before the draw.

Limitations

This tool does not diagnose conditions. Diagnosis requires clinical evaluation, history, and often imaging.
Threshold ranges reflect evidence-based optima for women's health, not universal standards. Individual variation exists.
Results should always be discussed with a licensed healthcare provider.
The rule set reflects the state of evidence as of the tool's authoring date and will be updated as evidence evolves.