Perio Causal Control Tower

Causal Simulator · Clinical Edition 1.0 · in silico counterfactual engine

Pearl Level 3 · Counterfactual n = 100

1 · Cohort scope

2 · Intervention — do(·)

3 · Comparator arm

The engine contrasts Ydo(protocol)(u) against Ydo(reference)(u) on the same exogenous draw u per patient — abduction → action → prediction, not two independent samples.

Endpoint by phenotype

mean 12-month ΔCAL — negative = attachment gain

  • Simulated protocol
  • Comparator arm

Individualised treatment effect

per-patient τi = ΔCALprotocol − ΔCALcomparator

  • Benefit (attachment preserved)
  • Harm (further loss)

Causal graph — GEM 21S regeneration DAG

edges into an intervened node are severed by the do-operator

Patient-level counterfactual ledger

  • AMD · Anterior-Mobility
  • MFD · Molar-Furcation
  • GSV · Generalized Severe
  • MNT · Maintenance
PatientPhenotypeDefect Sys loadHbA1cSPTCAL V1 CAL comparatorCAL protocolτi Prognosis
Synthetic cohort — not real patient data. 100 records generated from a structural causal model implementing the four architectural phenotypes and their DAG topologies. Product and sequence effects are parameterised to be directionally consistent with the published regenerative literature and with Application of Directed Acyclic Graphs (DAGs) to GEM 21S; magnitudes are illustrative and are not validated treatment-effect estimates. Intended as a demonstration of counterfactual reasoning mechanics — identification, effect modification, and unit-level contrast — not as clinical decision support. GEM 21S is a registered product of its manufacturer; used here as a named exemplar of a growth-factor + scaffold intervention node.