Product Requirements Document (PRD): RecoverPath

1. Executive Summary

RecoverPath is a specialized Remote Patient Monitoring (RPM) mobile platform designed to bridge the high-risk gap between hospital discharge and the first post-operative follow-up. By combining daily patient-reported outcomes (PROs), automated computer-vision wound analysis, and real-time clinical triage, RecoverPath enables surgical teams to detect complications like surgical site infections (SSIs) or secondary issues early, significantly reducing avoidable 30-day readmissions and improving patient peace of mind.

2. Problem Statement

Post-operative patients often experience "discharge anxiety" and struggle to distinguish normal recovery symptoms from early warning signs of complications. Simultaneously, clinicians lack visibility into the patient's status once they leave the facility, relying on scheduled follow-ups or emergency department visits. This "black hole" in recovery data leads to late-stage interventions, increased healthcare costs, and poor patient outcomes.

3. Goals & Success Metrics

Reduce Readmissions: Target a 25% reduction in 30-day post-op readmissions.
Early Intervention: Decrease the "Mean Time to Detection" (MTTD) for SSIs from an average of 4 days to <24 hours.
Clinician Efficiency: Maintain a false-positive alert rate of <10% to prevent alert fatigue.
Patient Compliance: Achieve an 85% completion rate for daily recovery surveys during the first 14 days post-surgery.

4. User Personas

The Recovering Patient (e.g., John, 68): Underwent total hip replacement; tech-literate but tires easily; needs simple, high-contrast UI and clear instructions.
The Primary Surgeon (e.g., Dr. Aris): High-volume orthopedic surgeon; needs a "at-a-glance" dashboard to identify only the patients requiring immediate intervention.
The Post-Op Nurse (e.g., Nurse Sarah): The primary monitor; handles the triage of "Yellow" alerts and initiates messaging or telehealth calls.
The Family Caregiver (e.g., Mark, 40): Assists John; needs visibility into John’s adherence to meds and mobility goals.

5. User Stories

As a patient, I want to upload a photo of my incision site so that the clinical team can tell me if it looks infected without me driving to the clinic.
As a surgeon, I want to receive a high-priority notification only when a patient’s vitals or pain scores spike significantly, so I can focus on critical cases.
As a nurse, I want to see a trend-line of a patient's mobility and mood over the last 5 days to assess if their recovery is stalling.
As a caregiver, I want to be alerted if my family member forgets to take their prescribed post-op medication.

6. Functional Requirements

6.1 Daily Recovery Surveys

The system shall prompt users daily to rate pain (1-10), mobility, and mood.
Survey logic shall follow "Local-First" architecture to allow offline completion.

6.2 Wound Analysis (Computer Vision)

The app shall guide the user to take a clear photo of the incision.
AI Feature: Integrate YOLO26 for wound localization and TopFormer for tissue segmentation to identify redness, slough, or exudate.

6.3 Automated Triage Engine

Green: Normal recovery; data logged.
Yellow: Minor deviation (e.g., slight temp increase); triggers nurse dashboard notification.
Red: Critical spike (e.g., severe pain + fever); triggers immediate push notification and haptic alert to the surgical team.

6.4 Secure Messaging

End-to-end encrypted (E2EE) chat between patient/caregiver and the clinical team using Seald.io integration.

6.5 Medication & Vitals Tracking

Log scheduled intake of anticoagulants and pain relief.
Manual and wearable entry for Temperature, BP, and HR.

7. Technical Requirements

7.1 Tech Stack (2026 Standards)

Frontend: Flutter v3.38.7 (Stable) using the Impeller rendering engine for high-performance UI across iOS/Android.
Backend: Node.js v24.13.0 (LTS) written in TypeScript.
Database: PostgreSQL 18 (leveraging uuidv7 for time-ordered PKs and JSON_TABLE for log analysis).
Hosting: AWS HealthLake (FHIR R4) with AWS Fargate (Private Subnets).

7.2 Core Integrations

EHR: Epic and Oracle Health (Cerner) via HL7 FHIR R4 and SMART on FHIR 2.1.
Security: Auth0 for identity; AWS KMS CMK for data-at-rest encryption.
Communication: Twilio for emergency SMS; Seald.io for E2EE messaging.

8. Data Model (PostgreSQL 18)

9. API Specification (Sample)

POST /api/v1/checkin

Request:

{
  "patientId": "018d1234-abcd-7000-8000-1234567890ab",
  "vitals": { "temp": 101.2, "heartRate": 88 },
  "painScore": 8,
  "timestamp": "2026-01-20T08:00:00Z"
}

Response:

{
  "status": "success",
  "triageResult": "RED",
  "action": "Clinician notified immediately."
}

10. UI/UX Requirements

Accessibility: Support for 200% text scaling and WCAG 2.2 Level AA compliance.
Offline Indicator: Clear visual state showing when data is saved locally vs. synced to the cloud.
Haptic Patterns: Unique vibration patterns for "Red" alerts on clinician smartwatches.
Wound Capture Overlay: A ghost-frame overlay to ensure the patient holds the camera at the correct distance and angle.

11. Non-Functional Requirements

Compliance: Full HIPAA and GDPR compliance; SOC2 Type II audited.
Latency: Critical "Red" alerts must be delivered within <120 seconds of data submission.
Availability: 99.9% uptime for the triage engine.
Data Integrity: Use of CRDTs (Conflict-free Replicated Data Types) for offline survey merging.

12. Out of Scope

In-app payment processing for medical bills.
General health tracking (weight loss, calorie counting).
Long-term chronic disease management (post-op window only: 0–90 days).

13. Risks & Mitigations

Risk: Alert Fatigue (Clinicians ignore notifications).
- Mitigation: Implement pattern-based logic (5-minute sustained spike) and automated escalation to Charge Nurses if unacknowledged.
Risk: Poor Image Quality (AI cannot analyze blurry photos).
- Mitigation: On-device blurring detection and real-time guidance during the photo capture flow.
Risk: Network Dead-zones.
- Mitigation: Local-first architecture with bi-temporal timestamping to ensure data integrity once reconnected.

14. Implementation Tasks

Phase 1: Project Setup & Infrastructure

[ ] Initialize backend with Node.js v24.13.0 (LTS) and TypeScript.
[ ] Initialize frontend with Flutter v3.38.7 using FVM.
[ ] Provision AWS HealthLake environment and sign BAA with AWS.
[ ] Set up PostgreSQL 18 with uuidv7 support and partition by surgery_date.

Phase 2: Core Data & Sync

[ ] Implement Local-First storage using SQLite/SQLCipher on device.
[ ] Build CRDT-based synchronization logic for daily surveys.
[ ] Develop FHIR-compliant Patient and Observation resources.
[ ] Implement SMART on FHIR 2.1 Auth flow for Epic/Cerner integration.

Phase 3: AI & Triage

[ ] Integrate YOLO26 Nano for on-device wound detection.
[ ] Build the Triage Engine logic (Red/Yellow/Green thresholds).
[ ] Implement Twilio/Push notification service with <120s latency target.
[ ] Develop haptic feedback profiles for clinical alerts.

Phase 4: Secure Messaging & Vitals

[ ] Integrate Seald.io SDK for E2EE messaging in Flutter.
[ ] Build medication adherence logger with local notifications.
[ ] Develop clinician dashboard with "At-Risk" patient sorting.
[ ] Conduct HIPAA penetration testing and SOC2 compliance audit.