Bayesian Causal Atlas · Vol. Endocrine · Pearl Structural Causal Model

Thyroid Disease — Structural Causal Analysis

Thyroid dysfunction affects an estimated ~20 million United States adults (hypothyroidism most common, with hyperthyroidism, nodular and autoimmune disease). Mostly morbidity, but untreated dysfunction is linked to atrial fibrillation, cardiovascular events, fracture and, in extremes, mortality. This oracle estimates the causal effect of restoring euthyroid control — the shared mediator — across replacement, anti-thyroid, ablative and surgical options, using effect sizes from named guidelines and trials. For education, not individual medical advice.

Method. Structural Causal Model (SCM) with backdoor adjustment (Pearl). Interventions are not assumed independent: each acts by moving thyroid function toward the euthyroid set-point, so overlap is removed by an eigenvalue-corrected equicorrelation model at an adjustable mean cross-correlation ρ̄ (default 0.30). Robustness to unmeasured confounding is quantified with the E-value. PN / PS / PNS are reported under monotonicity. Every estimate is cited — no effect size is invented. Note: hypo- and hyper-thyroid therapies are for opposite disorders and are not co-administered. The front door is resolved through an EXPLICIT mediator cascade (restore thyroid → symptomatic → adjunct → disease state), not one lumped node: each intervention acts on a specific node, so same-node interventions are substitutes that saturate against each other, while different-node interventions are d-separated given the intermediate node and compose in series. The cross-correlation removal thus follows from the graph structure; the residual ρ̄ cleans up only the mediator-independent (direct) effects.
ρ̄ = 0.30
A–D (all)
A high (RCT/meta) · B cohort · C case-series/modelled · D consensus/provisional. Lower-grade interventions are excluded from the DAG, front-door pooling, Pareto, Monte‑Carlo & sensitivity.

Interventions

Tick the interventions to combine. Each shows its trial effect estimate, 95% confidence interval (CI), E-value, mechanism, and citation. ★ = in the current Pareto effective set but not yet ticked.

Combined causal estimate

Headline is the front-door estimate: shared thyroid-control overlap removed via dose-response saturation; residual direct-effect overlap removed via the eigenvalue model at ρ̄.

1.00
Combined RR
0%
Relative risk ↓
Pooled E-value
Interventions selected (k)
0
Effective independent dimensions (n_eff)
0
Redundancy discount applied
0%
Baseline risk (illustrative anchor)
Absolute risk after intervention
Absolute risk difference (RD)
Backdoor-only RR (no front-door)
Mediator (control) overlap removed
Number needed to treat (NNT)
ρ-sensitivity band (ρ 0 → 0.6)
Interpretation

Causal attribution

Under monotonicity + exogeneity (E-value bounds the exogeneity assumption).

Probability of Necessity (PN)
Probability of Sufficiency (PS)
Prob. of Necessity & Sufficiency (PNS, lower bound)
Causal DAG
Cross-correlation
Pareto (threshold)
Monte Carlo
Front-door mediation
What-if / If-not-for
Sensitivity
Antithesis

Faithful causal directed acyclic graph (DAG). Each therapy is a node; each acts through the shared mediator euthyroid restoration, which drives atrial-fibrillation/cardiovascular risk, bone loss and symptoms — the components of the morbidity/mortality burden Y. Beta-blockade acts independently of thyroid state (symptomatic). Named confounders — especially disorder type — open back-door paths (adjusted). Mediator cascade: interventions attach to the node they act on (restore thyroid → symptomatic → adjunct), which converge on the disease state and thence the endpoint — drawing the intermediate mediators explicitly is what exposes d-separation and per-channel saturation.

Confounders U:age · thyroid disorder type · autoimmunity · iodine status · cardiac comorbidity · pregnancy → back-door paths (adjusted)Levothyroxine(hypothyroid)Antithyroid drug(hyperthyroid)Radioactive iodineThyroidectomyBeta-blockerSelenium (autoimmune)Restore thyroidfunctionSymptomatic(beta-blocker)Adjunct(selenium)ThyroiddysfunctionCV / outcomeburdenFront-door: through euthyroid restorationThyroid-state-independent (symptomatic)Back-door confounding (adjusted)

Eigenvalue diagnostics for the selected interventions under an equicorrelation matrix (off-diagonal ρ̄). A large λmax relative to k signals redundancy; n_eff is the effective number of independent interventions actually contributing.

k (selected)
0
λmax
λmin
n_eff = (Σλ)² / Σλ²
Condition number

On mechanistic grounds ρ̄ ≈ 0.30 is defensible within a disorder: replacement, anti-thyroid, ablative and surgical options all converge on the euthyroid set-point, so combining two definitive therapies is redundant. ρ̄ is user-adjustable, but the more important caveat is that hypo- and hyper-thyroid therapies are never stacked. Most of the overlap is now handled structurally by the mediator nodes (same-node substitutes saturate); ρ̄ governs only the residual correlation among direct effects.

Minimum-effective-set analysis. Set a target combined risk reduction; the model finds the smallest set of interventions — accounting for front-door mediator overlap — that reaches it, and highlights them. If the target exceeds what all interventions together can achieve, the full set is shown (never an empty one). "Apply" ticks exactly that set.

Target combined risk ↓ ≥ 50%

Monte Carlo propagation. Each selected intervention's log-effect is sampled from a normal distribution implied by its 95% CI; samples are combined with the same eigenvalue overlap discount. 5,000 draws.

Median combined RR
95% simulation interval
Standard deviation of combined RR
P(combined RR < 0.90)

Antithesis — challenging this oracle's own conclusions

Hypo and hyper are opposite disorders. Levothyroxine treats an underactive thyroid; antithyroid drugs, radioiodine and surgery treat an overactive one. They are never combined in the same patient — a pooled estimate only makes sense within one disorder, so read the subtype selector as fixing the arm.
Subclinical treatment is largely unproven. The randomised TRUST trial found levothyroxine gave no symptom or cardiovascular benefit for subclinical hypothyroidism in older adults. Normalising a mildly abnormal TSH is not the same as improving outcomes; over-treatment risks atrial fibrillation and bone loss.
The mortality/AF link is inferred, not trial-proven. Overt thyroid dysfunction is associated with atrial fibrillation, cardiovascular events and fracture, but few randomised trials show that treating it reduces hard endpoints. The downstream arrows are observational bounds.
Definitive therapy trades one state for another. Radioiodine and surgery reliably control hyperthyroidism but usually create lifelong hypothyroidism; the favourable control relative risk ignores the burden of permanent replacement and, for surgery, operative complications.
Over-replacement is its own harm. Excess levothyroxine (suppressed TSH) causes iatrogenic thyrotoxicosis with atrial fibrillation and accelerated bone loss — a competing harm not captured by a “restore euthyroidism” relative risk.

What-if — the do-operator: P(Y | do(S))

Intervening on the selected set S with Pearl's do-operator (setting the interventions, not merely observing them). Contrast against do(∅) = no intervention.

P(outcome | do(∅)) — baseline
P(outcome | do(S)) — intervened
Absolute risk reduction (ARR)
Number needed to treat (NNT)

If-not-for — but-for counterfactual (leave-one-out)

For each intervention: "if not for this one, the combined front-door effect would be…". Isolates each intervention's marginal causal contribution after mediator-overlap removal, so shared-pathway agents are not double-credited.

If not for…RR without itRR with full setmarginal RRR lost

One-at-a-time sensitivity. Each intervention's effect is swung across its 95% confidence interval (others held at point estimate); the bar is the resulting swing in the combined front-door effect. A long bar means the combined estimate leans heavily on that single trial's precision.

Optimistic bound (all at CI-low)
Point estimate
Pessimistic bound (all at CI-high)
Pooled E-value (confounding robustness)

Front-door (mediation) decomposition. Each therapy acts through one shared mediator — restoration of euthyroid control. Each log-effect is split into a control-mediated (indirect) and an independent (direct) part. Indirect parts are pooled through the mediator with dose-response saturation (you cannot become more than euthyroid), removing the mediator cross-correlation; direct parts keep the residual eigenvalue correlation at ρ̄. Beta-blockade, a thyroid-state-independent symptomatic bridge, is not discounted against the disease-modifying options. Here mediated effects are pooled WITHIN each cascade node (dose-response saturation of substitutes) and composed in SERIES across nodes (d-separated channels), with the per-node reductions reported so the channel structure is visible.

Mediator saturation cap = 55% control ↓
Sum of standalone thyroid-control reduction (naive)
Combined thyroid-control reduction after saturation
Mediator overlap removed (1 - saturation)
Direct-effect redundancy removed (1 - n_eff/k)
Front-door combined RR
Backdoor-only combined RR (comparison)
InterventionRR%ctrlmed-fracindirect logdirect log

Which % of cross-correlation is appropriate? Not one number. The mediator overlap is fixed empirically by the thyroid-control saturation (currently removing of the summed mediated effect when interventions are stacked). Note the overriding domain caveat: hypothyroid and hyperthyroid therapies treat opposite disorders and are never combined; and treating subclinical hypothyroidism is contested (TRUST was null). Control-mediated fractions are transparent, adjustable priors.

Front-door caveat (antithesis): the endpoint is thyroid control — a process surrogate. The atrial-fibrillation, cardiovascular and fracture benefits are largely inferred from association, not proven in randomised trials, and over-replacement is itself harmful.

Executive summary

Select interventions to generate a plain-language summary.