New 0604 Pthc Valya Irisa Laura Vanessafinal Version

The New 0604 PTHC project has successfully:

Given these outcomes, the evidence strongly supports moving from pilot to full deployment, with the recommended governance and maintenance structures in place to sustain performance over time.

| Metric | Baseline (pre‑PTHC) | Post‑Implementation (pilot) | % Change | |--------|---------------------|------------------------------|----------| | Median TTT (days) | 21 | 17 | ‑19 % | | High‑risk patients identified correctly (sensitivity) | — | 92 % | — | | Low‑risk false‑positive rate | — | 8 % | — | | Average number of unnecessary tests per patient | 2.3 | 1.1 | ‑52 % | | Patient satisfaction (5‑point scale) | 4.2 | 4.6 | +9 % | | Cost saving (per 1,000 patients) | — | $145,000 | — | New 0604 Pthc Valya Irisa Laura Vanessafinal Version

Model Performance

| Model | AUC | Calibration (Hosmer‑Lemeshow p) | |-------|-----|-----------------------------------| | Logistic Regression | 0.84 | 0.62 | | Random Forest | 0.88 | 0.71 | | XGBoost (final) | 0.89 | 0.78 | The New 0604 PTHC project has successfully:

The XGBoost model was selected for deployment due to its superior discrimination and acceptable calibration.

| Milestone | Target Date | Owner | Deliverable | |-----------|-------------|-------|-------------| | Final sign‑off on SOPs | 15 Apr 2026 | Vanessa | SOP package (PDF) | | System go‑live (all sites) | 1 May 2026 | Vanessa | Live EHR integration | | Staff training completion | 31 May 2026 | Vanessa | Training attendance logs | | First KPI review | 30 Jun 2026 | Valya | KPI report | | Model recalibration (12 mo) | 30 May 2027 | Laura | Updated model version | | Expansion feasibility study | Q3 2027 | Irisa | Feasibility report | Given these outcomes, the evidence strongly supports moving

| ID | Risk / Issue | Likelihood | Impact | Mitigation / Resolution | |----|--------------|------------|--------|--------------------------| | R1 | Data‑quality inconsistency across sites | Medium | High | Implement automated data‑validation scripts; quarterly audits. | | R2 | Clinician resistance to workflow change | Medium | Medium | Conduct hands‑on training, champion program (Vanessa). | | R3 | Regulatory compliance (HIPAA) | Low | High | Privacy impact assessment completed; encryption of API calls. | | I1 | Delay in EHR vendor API release (Sept 2025) | High | Medium | Secured interim data‑feed via HL7 interface; updated schedule. | | I2 | Model drift after 12 months | Medium | Medium | Set up continuous‑learning pipeline; quarterly recalibration. |