Abstract TH843: Signatures for Tailored Cardiovascular Communication: A Three-Pathway Classification With Chatbot-Enabled Delivery

Abstract

Background: One-size-fits-all counseling often underperforms in long-term cardiovascular prevention where sustained activation is required. We developed Signatures, a practical framework that maps individuals to four communication archetypes—Listener, Motivator, Director, with an Expert overlay—to guide message tone, structure, and shared decision-making at the point of care. The framework operationalizes psychographic segmentation into an implementable taxonomy. Methods: Classification uses three complementary pathways: (1) a 20-item self-assessment that summarizes activation and support-need domains; (2) a clinician 10-domain binary grid scored 0–10; and (3) a supervised NLP classifier that ingests de-identified narrative to estimate archetype probabilities. Discordance is resolved by conservative tie-breaking and barrier-domain overrides (health literacy, trust, access, food security). Intervention (Chatbot): We prototyped a rules-plus-NLP chatbot to (a) administer the self-assessment, (b) collect short narratives for NLP pre-labeling, and (c) deliver Signature-specific counseling (e.g., plain-language, one-step plans for Listeners; option sets and SMART weekly goals for Motivators; concise, data-driven progressions for Directors; synthesis and trade-offs for Experts). Example phrase templates were derived from our library of Signature-aligned responses for common health questions. Results: Formative testing established face validity of the three-pathway workflow and usability of chatbot dialogues. The system consistently generated actionable outputs: an assigned archetype, domain-level flags, and a message kit (tone, structure, and SDM cues) that clinicians can use or edit in real time. The chatbot supported weekly goal-setting, reminders, and teach-back prompts aligned to the assigned Signature. Conclusions: A triaged, multi-method classification combined with chatbot delivery is a feasible approach to precision communication in cardiovascular prevention, offering a practical bridge from psychographic theory to routine encounters and remote interactions. Prospective validation will assess concordance among pathways, equity, and effects on engagement, lifestyle habits, condition management and clinical proxies.