Cholesterol Interventions — SCM Causal Engine

In plain wordsThis tool looks at cholesterol and stiff arteries, which drive heart disease. It shows how much treatments help — and that counting them naively overstates the benefit by about a quarter, because they overlap. Combining a statin with add-on drugs works, but not as much as simple addition suggests. Best-case research estimates, not medical advice.

LDL

HDL

eNOS

Infl

OxStr

Evidence tier → causal graph & MACE HR: A–D

A=hard-endpoint RCT/meta · B=RCT on surrogate. Lower grades excluded from the combined effect.

Biomarker States — combined structural equation output

Select interventions to see biomarker changes

Pathway Causal Decomposition — with cross-correlation adjustment

Pathway	Structural Eq. Output	Raw HR	Cross-Corr Adjustment	Adj HR	Evidence Basis
Select interventions

Naïve HR (uncorrected product)

—

multiplicative overcounting

→

Adjusted HR (cross-corr corrected)

—

structural equation model

→

Overcounting Error

—

naive − adjusted

MACE Probability

P(MACE | do(X)) — Adj

—

Select interventions

Baseline 10yr —

ARR —

RRR —

HR (adjusted) —

HR (naïve) —

What-If-Not (counterfactuals)

No interventions selected

PN / PS / PNS — Balke-Pearl

—

PNS bounds

—