Stanford Med · Unit 5B Medicine·Day shift·Mon · 9:42 AM

Patient Discharge Cockpit

David Chen · Acute psychotic episode + ESRD on HD (F23 + N18.6)
Switch role
MC
DC
David Chen
38 M · MRN-58518 · Acute psychotic episode + ESRD on HD (F23 + N18.6)
BarrieredUnit 5BBehavioral HealthAnthem Blue Cross CA (Medi-Cal MCO)LOS 6d
ComorbiditiesEnd-stage renal diseaseSchizoaffective disorderAlcohol use disorderHypertensive CKD with HTN
Estimated discharge
Wednesday 3:00 PM
HIGH delay risk
Six readiness dimensions

All six must align for discharge to actually happen

Open discharge map →
♡
Clinical readiness
75%Blocked
⚒
Functional & equipment readiness
50%Blocked
℞
Medication readiness
67%Blocked
$
Financial readiness
67%In progress
▦
Care coordination readiness
50%In progress
✿
Patient agreement & support readiness
50%Blocked
Unified readiness model

Three axes → one “ready” verdict

Not ready · 6 blockers
Axis 1 · Feasibility
Can this disposition happen? (engine = authority)
4 blocker(s)
Target: HOME
Binding: Clinical acuity · VIOLATED
Hard blockers: Behavioral & psychiatric stability
Axis 2 · Progress
How far along is the work? (tasks, not a score)
0/2 required
0% of execution tasks complete
Clinical readiness0/1
Medication readiness0/2 ●
Patient agreement & support readiness0/5 ●
Axis 3 · Execution-staging
Is the bed actually secured? (acceptance pipeline)
No bed needed
Stage: Insurance Authorized
Home-track disposition · no external bed to confirm.
  • ●3 engine constraint(s) violated: CLINICAL_ACUITY, FUNCTIONAL, SOCIAL_SDOH.
  • ●Hard category blocker(s): BEHAVIORAL_STABILITY.
  • ●T2 gate not cleared: Hospitalist 'medically ready' attestation on file.
  • ●T2 gate not cleared: Patient/family OOP acceptance documented (if residual cost).
  • ●2 required execution task(s) still open.
  • ●1 readiness gate(s) unmet: PT evaluation completed and documented.
Scalars (display-only): clinical 72 · operational 18 — non-authoritative; feasibility + progress above are the truth.
30-d readmission risk
LACE+ + HOSPITAL · not HRRP-tracked
MODERATE

Drives pathway intensity tier — HIGH score adds Coleman / Naylor touchpoints automatically.

Active pathwayelevated

Behavioral + dialysis combined discharge

disposition=HOME · services=BEHAVIORAL_OUTPATIENT+DIALYSIS_OUTPATIENT+NEMT_RECURRING+OUTPATIENT_PHARMACY · rules=rule-ckd-esrd-continuation

Open patient workflow →
0 / 13 steps
0 done6 in progress7 blocked0 pending

Intensity tier: LACE+ 9 (MODERATE) + HOSPITAL 3 (LOW) → MODERATE.

Engine-derived barriers

Top blockers · with one-click playbooks

1 barrier
  • Actively blockingOwner: Attending· Open 6d
    Internal clinical

    Active behavioral risk (SI/psychosis/withdrawal) — unsafe for home or unlicensed settings.

    via extendedReadiness blocker (CLINICAL_ACUITY)

    → drafts go to HITL queue, not sent until approved
Execution layer · HITL queue

Pending actions · awaiting clinician approval

0 open

Drafted by playbooks / AI · NOT sent until you approve. Approve fires the integration mock + writes a Submission audit row.

No pending actions. Click Run playbook on a barrier to draft one.
From the family · patient portal

What the caregiver has told us

Not started
Recovery preferenceWants to talk it through
TransportNot sure — wants a call
Home supportNeeds extra help
Preferred pharmacy—
Preferred languageEnglish
Discharge teaching0/3 videos watched
Workflow swim lanes

Tasks by role · cross-functional discharge orchestration

4 tasks
Hospitalist2
Assigned
Resolve: Internal clinical
In progress
Hospitalist 'medically ready' attestation on file
Nursing0
No active tasks
Pharmacy1
In progress
Discharge medication reconciliation completed by pharmacist
PT / OT1
In progress
PT evaluation completed and documented
Case manager0
No active tasks
Event simulator · demo only

Fire a realistic event to watch the engine recompute

src/lib/engine/dispatcher.ts
Event feed · dynamic engine

What moved the pillars

0 total
No events yet. Use the simulator below to fire a payer decision, FHIR observation, or vendor callback and watch the pillars move.

AI chart summary · live via Anthropic Claude

Runs through runAgent · writes an AgentInference audit row with cost, latency, prompt hash.

Click Run chart summary to call Claude Sonnet 4.6 on this patient's record. The audit row will be visible in Neon's AgentInference table within ~3s.

Prior auth drafter · Claude → HITL → CoverMyMeds

Draft → review → approve → submit

Live wire

Drafts via src/lib/ai/agents/prior-auth-drafter.ts → submits via src/lib/integrations/cmm.ts (mock-contract). Both write audit rows.

Demo data · no PHI · mocked Epic + payer endpoints