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

Osteoporosis — Structural Causal Analysis

Osteoporosis and low bone mass affect ~54 million United States adults aged 50+, and a hip fracture carries a ~20–30% one-year mortality. This oracle estimates the causal reduction in fracture / mortality achievable by combining bone-directed therapies, using hazard ratios (HRs) from named published trials — flagging where the endpoint is a surrogate or the evidence is weak.

Method. Structural Causal Model (SCM) using cited, confounder-adjusted effect sizes (backdoor adjustment is the source study’s) (Pearl). Interventions are not assumed independent: their heavy mechanistic overlap (act partly through bone mineral density and partly through bone quality, turnover and fall reduction) is removed by an eigenvalue-corrected equicorrelation model at an adjustable mean cross-correlation ρ̄ (default 0.30). Robustness to unmeasured confounding is quantified per intervention with the E-value. Probabilities of Necessity / Sufficiency (PN / PS / PNS) are reported under a monotonicity assumption. Every hazard ratio is cited to its source trial — no effect size is invented.
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 hazard ratio, 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 bone-density-mediator overlap removed via dose-response saturation; residual direct-effect overlap removed via the eigenvalue model at ρ̄.

1.00
Combined HR
0%
Relative risk ↓
Pooled E-value
Interventions selected (k)
0
Effective independent dimensions (n_eff)
0
Redundancy discount applied
0%
Baseline 10-yr risk (illustrative anchor)
Absolute risk after intervention
Absolute risk difference (RD)
Backdoor-only HR (no front-door)
Mediator (weight) overlap removed
Number needed to treat (NNT)

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 intervention is a distinct node; effects flow through named mediators — the front-door path via bone mineral density (BMD), the BMD-independent path via bone quality/turnover and the muscle→falls chain — into fracture, which itself mediates mortality. Estrogen carries a competing-harm path (breast cancer / venous thromboembolism / stroke → mortality). Named confounders open back-door paths (adjusted; robustness bounded by the E-value). This graph is illustrative of structure, not yet the identification engine.

Confounders U:age · menopausal status · glucocorticoid use · prior fracture · renal function · smoking/alcohol · body-mass index → back-door paths (adjusted)BisphosphonateDenosumabAnabolic(romo / teriparatide)Calcium +vitamin DExerciseEstrogen / HRTBone resorption/ turnover ↓Vitamin-D statusMuscle strength ↑Bone micro-architectureBone mineraldensity (BMD)Fall risk ↓FractureMortality (Y)Breast cancer /VTE / strokeFront-door: through BMDBMD-independent: bone quality / fallsCompeting harm → mortalityBack-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 for this domain: incretin drugs, surgery, diet, and activity overwhelmingly share the weight-loss / insulin-sensitivity pathway, so ~30% of their nominal effects overlap. A mediation analysis of semaglutide found ~80% of its MACE benefit is not mediated by weight loss, which is why ρ̄ is user-adjustable rather than fixed.

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-hazard-ratio 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 HR
95% simulation interval
Standard deviation of combined HR
P(combined HR < 0.90)

Antithesis — challenging this oracle's own conclusions

Bone density explains little of the benefit. Only ~11\u201318% of risedronate's anti-fracture effect is mediated by BMD change (Watts); most acts through turnover suppression and microarchitecture. A BMD-only surrogate therefore understates efficacy \u2014 the front-door layer corrects for this.
Estrogen was net harm. The Women's Health Initiative confirmed estrogen cuts hip fractures (HR 0.67) but increased breast cancer, stroke and venous thromboembolism, so it is NOT recommended for osteoporosis alone. A protective HR on one endpoint can hide net harm.
Romosozumab carries a cardiovascular signal. ARCH showed more serious cardiovascular events on romosozumab than alendronate, producing a boxed warning. Its superior fracture reduction is traded against cardiovascular risk \u2014 not captured by a fracture-only HR.
Denosumab cannot simply be stopped. Discontinuation triggers rapid bone loss and a rebound of multiple vertebral fractures; the on-treatment HR overstates the real-world benefit of a therapy that must be continued indefinitely or transitioned to a bisphosphonate.
Fractures come from falls, not only bone. Much fracture risk is fall-driven; exercise/fall-prevention act on that BMD-independent pathway, while calcium + vitamin D is largely NULL in community-dwelling adults. An oracle mediating only through density misses half the causal graph.

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(mortality | do(∅)) — baseline
P(mortality | 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 hazard ratio would be…". Isolates each intervention's marginal causal contribution after mediator-overlap removal, so shared-pathway agents are not double-credited.

If not for…HR without itHR with full setmarginal RRR lost

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

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

Front-door (mediation) decomposition. Bone therapies act partly through one shared mediator — bone mineral density (BMD) gain. Each log-hazard-ratio is split into a BMD-mediated (indirect) and a BMD-independent (direct) part (bone quality/turnover, and the fall pathway). Indirect parts are pooled through the mediator with dose-response saturation (BMD gain cannot stack without limit), removing the mediator cross-correlation; direct parts keep the residual eigenvalue correlation at ρ̄. This is why exercise/fall-prevention (which acts on falls, not density) is NOT discounted for overlap with bone drugs.

Mediator saturation cap = 12% BMD gain
Sum of standalone BMD gain (naive)
Combined BMD gain after saturation
Mediator overlap removed (1 - saturation)
Direct-effect redundancy removed (1 - n_eff/k)
Front-door combined HR
Backdoor-only combined HR (comparison)
InterventionHR%BMDBMD-medindirect logHRdirect logHR

Which % of cross-correlation is appropriate? Not one number. The mediator overlap is fixed empirically by the BMD dose-response saturation (currently removing of the summed BMD-mediated effect when bone drugs are stacked); the direct overlap is the residual ρ̄. BMD-mediated fractions are cited from published mediation analyses (risedronate ~11–18%; denosumab ~35–51%) — a striking finding that most anti-fracture benefit is NOT explained by density.

Front-door caveat (antithesis): BMD is a weak mediator here — it explains a minority of the anti-fracture effect, so a BMD-only model badly understates efficacy. The BMD→fracture path is also confounded (bone quality co-varies with density). This is a natural direct/indirect-effect decomposition in the front-door spirit, not a clean front-door-structured mediation; indirect estimates are bounds.

Executive summary

Select interventions to generate a plain-language summary.