From directive to practice: Operationalizing a system-level systemic anticancer therapy (SACT) oversight program at a large academic cancer center.

503 Background: International guidelines recommend reducing SACT near the end of life (EOL), but cancer centers struggle to translate this into frontline behavior. At MD Anderson Cancer Center, an institutional directive established two simultaneous standards — reducing non-beneficial EOL SACT and transitioning select inpatient (IP) regimens to outpatient (OP) delivery. Three interdependent workstreams were implemented to build clinical consensus, ensure ambulatory readiness, and embed accountability into the EHR workflow. Methods: 1. Methodology Review assessed 787 regimens, classifying each for OP or IP delivery. 2. Ambulatory Operations ensured OP capacity could absorb increased volume, addressing barriers (housing and access). 3. Exceptions Review developed standardized clinical criteria to identify patients at higher risk for poor outcomes or near-term mortality, informing two EHR order sets embedded in the exception request workflow. Prior to giving IP SACT, oncologists submit a request in the EHR with clinical criteria auto populated from the patient's chart. If the regimen is OP-designated or the patient meets criteria for poor prognosis, approval from a department-designated physician reviewer is required. Each department identified 2-3 peer reviewers with disease-specific expertise rather than routing to a central committee, preserving local accountability. All decisions are documented in the EHR, creating an auditable record. Results: Data represent Nov 2025 through May 2026 across 9 solid tumor departments and 386 IP encounters. Of these, 339 exception requests were submitted (88% compliance; range 25%-98%). Of 153 encounters requiring designee review, 107 were completed (70% compliance; range 0%-94%). Departments at 0% — Breast and Gynecologic Oncology — each had 4 encounters each with all regimens defaulting to OP, suggesting early-stage adoption rather than resistance. Exception requests reflected a mix of scenarios including unmet criteria, prior complications, disease progression, emergent needs, and social circumstances. Both metrics reflect increasing adoption and continue to improve. EOL chemotherapy rates declined from 0.10% (Nov 2025) to 0.07% (Apr 2026), a 30% relative reduction (Vizient Definition). Conclusions: A sequenced three-workstream model anchored in clinical consensus, ambulatory readiness, and EHR-embedded accountability translated a system-level directive into measurable frontline compliance. Early compliance with both workflows was high, with ongoing efforts to strengthen the process by linking dispensing to documented approval. The department-level reviewer model recognized disease specialization and preserved clinical autonomy within a centralized framework. Department-level compliance data will drive targeted engagement as the program scales toward 100% compliance.