GENERATIVE AI-ENABLED OBJECTION HANDLING TO SUPPORT PAYER ENGAGEMENT IN ADJUVANT MELANOMA THERAPY

Author(s)

Turgay Ayer, PhD1, Sumeyye Samur, PhD1, Ismail F. Yildirim, MSc1, Mine Tekman, PhD1, Jag Chhatwal, PhD2;
1Value Analytics Labs, Boston, MA, USA, 2Massachusetts General Hospital/Harvard Medical School, Boston, MA, USA
OBJECTIVES: Payer and HTA coverage decisions are frequently influenced by objections. Our objective was to evaluate if generative AI can support evidence-based objection handling to improve readiness for payer engagement in adjuvant pembrolizumab for resected high-risk melanoma.
METHODS: We used ValueGen.AI, a "Deep Agent" architecture using the LangGraph framework that orchestrates over 20 specialized sub-agents to decompose queries and coordinate extracting evidence from multiple sources. Using pembrolizumab as adjuvant therapy for resected high risk melanoma (stage IIB/IIC and stage III), ValueGen.AI simulated common payer objections and compiled counterarguments supported by citable evidence. Sources synthesized included KEYNOTE 716 and KEYNOTE 054 trial publications/updates, cost effectiveness analyses, HTA documents (NICE and IQWiG), HRQoL reports, and clinical guidelines (ASCO 2023 and ESMO 2024).
RESULTS: The tool generated five payer objection domains and, for each, produced a structured objection-response playbook. For cost/budget impact, the tool advised citing multiple cost‑effectiveness evaluations (e.g., ~$68,736/QALY in US stage IIB/IIC; CHF 27,424/QALY in Switzerland) and reinforcing affordability narratives using supportive HTA conclusions (e.g., NICE TA766/TA837) and transparent payer cost tables (IQWiG). For immature OS, it recommended positioning RFS/DMFS as decision-relevant adjuvant endpoints and referencing durable improvements in KEYNOTE‑716 and persistent long‑term benefit in KEYNOTE‑054 follow‑up. For toxicity/HRQoL, the tool advised pairing standard immune‑related AE management approaches with HRQoL evidence indicating no meaningful long‑term decrement versus placebo. For watch‑and‑wait, it recommended emphasizing that crossover at recurrence did not eliminate population‑level RFS/DMFS advantages and citing timing evidence (SWOG S1801) supporting earlier checkpoint inhibition. For patient selection/biomarkers, it recommended highlighting consistent subgroup effects and guideline-based clinicopathologic selection rather than biomarker gating. The output concluded with a concise payer briefing organized into clinical benefit, cost‑effectiveness, safety/HRQoL, and guideline alignment.
CONCLUSIONS: Generative AI can systematically anticipate payer objections and generate structured, evidence-based counterarguments aligned with clinical trials, HTA precedent, economic evaluations, and clinical guidelines.

Conference/Value in Health Info

2026-05, ISPOR 2026, Philadelphia, PA, USA

Value in Health, Volume 29, Issue S6

Code

HPR48

Topic

Health Policy & Regulatory

Topic Subcategory

Reimbursement & Access Policy

Disease

SDC: Oncology

Your browser is out-of-date

ISPOR recommends that you update your browser for more security, speed and the best experience on ispor.org. Update my browser now

×