Demo environment — not for clinical use — real patient records when signed in

Pathway control center

Eight approved pathways

Nexvani is not an unrestricted chatbot. Every patient statement is routed into one of these predefined pathways, each with fixed language, permitted actions and escalation criteria.

Pathway detail

Treatment Burden

Active

Patient-language examples

  • “I hate dialysis.”
  • “I'm sick of dialysis.”
  • “I don't want to go.”
  • “I'm tired of doing this.”
  • “It takes over my entire life.”
  • “I can't stand sitting there for four hours.”

Classification logic

Language matching this pathway is classified as Treatment Fatigue / Burden. Classification is deterministic and confined to the approved pathway set — unmatched language routes to Unknown Barrier, never to free-form generation.

Permitted Nexvani response

“What about dialysis is making it hardest to go back right now?”

“Is that making you think about missing your next treatment?”

Permitted automated actions

  • Capture the specific burden in the patient's own words
  • Record barrier without escalation when attendance is not at risk
  • Flag social worker follow-up when attendance risk is stated

Human escalation criteria

  • Patient indicates they may miss the next treatment
  • Repeat burden reports across three consecutive check-ins

Staff responsible

Clinic Social Worker

Guardrail: AI may interpret patient language but cannot independently make clinical decisions.