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

Throughput Analytics

Hospital-wide · last 30 days · vs. CMS GMLOS baseline
Switch role
MC
Avg LOS
4.8d
−0.6d MoM
vs. CMS GMLOS 5.4d
Avoidable days saved
312
+82 vs. Q1
$1.8M annualized impact
Discharge before 12pm
42%
↑ 12pp
Target 50%
Readmit risk (30d)
11.8%
−1.4pp
HRRP comparator group
Avoidable bed days

By barrier type · last 7 weeks

↓ 38% vs. 7-wk avg
Clinical-ready → actual discharge

Median hours · trending down

14h · target 12h
72-hour discharge forecast

By window · across units

26 forecast · 13 confirmed · 10 at-risk
Case manager workload

Active patients · target ≤ 12

2 CMs over target
Unit-level bottlenecks

Avoidable hours · last 30 days

UnitPrior authSNF bedDME / O₂TransportFamily decisionPharmacy stock
5B Medicine82h41h22h14h24h18h
B6 BMT70h35h18h11h22h28h
S4 Stroke54h28h24h22h36h12h
M3 Ortho44h62h32h18h14h9h
4N Cardio60h30h28h16h22h19h
Cells colored by intensity · click any cell for patient drill-down (coming)
Prior auth TAT

Avg payer turnaround

18h
−47% vs. Q1
Post-acute placement TAT

Referral → SNF accept

34h
−45% vs. Q1
Medication access delay

Rx ordered → patient pickup

8h
↘ trending down
How this screen works
Read-only roll-up of the audit spine — what to show CFO + ops VP
Inputs
What this screen reads
  • AgentInference + Submission audit rows (last 30/90/YTD)
  • Encounter LOS + readiness history (computed from event log)
  • PaCHE barrier resolution timestamps
  • Brain insight history per signalKey
Engine
What it computes
  • Aggregates avoidable days by barrier type
  • Computes TAT (PA, post-acute placement, med access) per encounter
  • Forecasts next-72h discharges by combining EDD + readiness trend
  • Per-CM workload + per-unit bottleneck heatmap
Outputs
What it writes / routes
  • Hospital-wide dashboard for VP Ops + CFO
  • Drill-down to patient (planned: click cell → census filtered)
  • Monthly board pack PDF (planned)
  • Benchmarking vs. CMS GMLOS
Refresh trigger
When it updates
  • Nightly batch on audit spine
  • Real-time on discharge events
  • On-demand when filter window changes
Partners involved:Internal audit spine onlyOptional Datadog / Looker connector for hospital BI
Demo data · no PHI · mocked Epic + payer endpoints