Pre-Procedure Intelligence Layer for the EHR
We don't replace Epic. We read it — and generate the synthesis that currently lives in sticky notes and the surgeon's head.
Everything is documented — but nobody connects the dots across 15+ Epic screens.
Epic screens a nurse must review to prep a surgical patient
Average time a pre-op coordinator spends per patient piecing together the plan
of surgical delays caused by incomplete pre-op preparation (AORN 2023)
The data is all there. The patient's Plavix is in the MAR. The stent is in the Problem List. The biopsy is in the procedure order. But nobody connects those three facts into "hemostatic clips need to be in the room" until someone manually thinks about it.
— This is what Gawande calls an "error of inaptitude" — not a knowledge gap, but a synthesis gap.
We don't compete with the EHR — we complete it.
| Task | What Epic Does | What We Add |
|---|---|---|
| Drug alerts | Fires individual BPA pop-ups (alert fatigue — 90%+ override rate) | Connects medication + procedure + history → specific action with context |
| Pre-op prep | Stores orders as a list. Coordinator manually interprets. | Generates role-specific task lists: what YOU need to do, and why |
| Protocol deviations | Doesn't flag when a patient DOESN'T fit standard protocol | Explicitly states: "Standard protocol won't work because..." |
| Readiness tracking | Each item (labs, consent, escort) is in a different module | Single view: is this patient actually ready? What's still pending? |
| Role-based views | Same chart view for everyone. Nurse sees surgeon's notes. | Anesthesia sees anesthesia concerns. Nurse sees nursing tasks. No noise. |
Launches inside Epic as a SMART on FHIR app. Zero new data entry. Zero documentation burden.
Patient: Margaret Thompson, 70F, scheduled for Total Knee Arthroplasty
What lives in separate Epic screens: MAR: Warfarin 5mg daily (for AFib) • Problem List: History of DVT (2019), Latex allergy, Type 2 DM on insulin • Allergy List: Sulfonamide → Anaphylaxis • Labs: INR 2.3, HbA1c 7.8%
What our system synthesizes (that nobody else assembles automatically):
Source: Sulfa allergy + Surgical prophylaxis protocol
Action: Vancomycin 1g IV instead
Why: Cephalosporin cross-reactivity risk with sulfa anaphylaxis
Source: DVT history + Major joint surgery
Action: Enoxaparin extended protocol
Why: Recurrence risk 3x higher with standard duration
Source: Allergy list + OR setup
Action: Latex-free cart, all supplies verified
Why: Contact dermatitis → potential intra-op reaction
Source: MAR (Lantus 22u) + Endocrine consult
Action: Half dose night before, hold morning of
Why: Prevent perioperative hypoglycemia
None of these are individually hidden. But assembling them requires reading 5 different modules and holding all the connections in your head — for every patient, every day. That's where things get dropped.
Read-only intelligence layer that lives inside the existing workflow.
Technical Integration Path:
1. Register as SMART on FHIR app (Epic App Orchard)
2. OAuth2 authorization - clinician launches from patient chart
3. Read FHIR resources (Patient, Condition, MedicationRequest,
AllergyIntolerance, Observation, ServiceRequest)
4. Cross-reference against procedure protocols
5. Display synthesis in sidebar - clinician stays in Epic
Data Flow:
Epic Hyperspace --> FHIR Server (Interconnect) --> Manifesto AI Engine
|
SMART App (in-browser)
|
External APIs (OpenFDA, RxNorm)
The synthesis gap costs hospitals billions — and exists at every Epic site.
Annual cost of surgical delays and cancellations in the US (Bain & Co, 2022)
Day-of cancellation rate for elective surgeries — most due to prep failures
of hospitals use Epic — and all have this synthesis gap
Target users:
We're not building a new EHR. We're building the intelligence layer that sits on top of the data hospitals already have — and generates the synthesis that currently depends on human memory under stress.
Next step: Pilot with one surgical service line at one hospital.
Measure: Pre-op coordinator time saved, day-of cancellation rate, and missed-item rate.
| FHIR Resource | Epic Module | What We Read |
|---|---|---|
| Patient | Demographics | Name, DOB, contacts, code status |
| Condition | Problem List | Active diagnoses, ICD-10 codes |
| MedicationRequest | MAR / Med List | Active meds, doses, prescribers |
| AllergyIntolerance | Allergy List | Substances, reactions, severity |
| Observation | Labs / Vitals | Recent results, vitals |
| ServiceRequest | Orders / Scheduling | Scheduled procedure, pre-op orders |
| DocumentReference | Notes | Consult notes, nursing assessments |
| Product | What It Does | Gap |
|---|---|---|
| Epic BPA Alerts | Single drug/allergy alerts | No cross-reference, 90%+ override rate |
| Epic OpTime | Schedules OR time, equipment | Doesn't read patient chart for deviations |
| Epic Anesthesia Pre-op | Documents anesthesia assessment | Doesn't synthesize for the anesthesiologist |
| Nurse checklist tools | Generic checklists | Not patient-specific, not adaptive |
| Manifesto AI | Reads ALL modules, generates patient-specific synthesis | This is the gap we fill |