HandRehab+ addresses gaps in outpatient rehabilitation by pairing guided home exercise with daily symptom tracking, AI-assisted range-of-motion (ROM) analysis, and a secure provider dashboard. The goal is to improve adherence, enable timely interventions, and support data-driven care between clinic visits. Target users are:
- Patients recovering from hand, wrist, or elbow surgery (e.g., thumb CMC arthroplasty)
- Occupational/Physical Therapists who monitor progress, adjust plans, and communicate with patients
Evidence from recent mHealth literature shows that interactive apps with multimedia exercises, tracking, feedback loops, and clinician oversight improve adherence and outcomes versus paper instructions alone. HandRehab+ implements these evidence-based features end-to-end. :contentReference[oaicite:1]{index=1}
- Daily assessments: pain, grip strength, sleep, fatigue
- Video-guided exercises with clear pacing and form cues
- AI ROM analysis (on-device) for joint angle estimation (privacy-preserving)
- Progress dashboards and badge rewards to motivate adherence
- Educational micro-lessons & quizzes to reinforce self-management
- Caseload dashboard with real-time metrics and trends
- AI-generated alerts for missed sessions, rising pain, or stalled progress
- Plan management: assign/adjust exercise programs, message patients
- Secure data sharing into clinical workflows
Early usability sessions (therapists + post-op patients) informed revisions such as clearer pain-scale guidance, slow-motion looping exercise videos, and streamlined quizzes. :contentReference[oaicite:2]{index=2}
- Patient Home: schedule, % recovered, reps completed, shortcuts to plans
- Exercise Tracker: status by activity (Not Started/In Progress/Completed)
- Provider Dashboard: caseload view, alerts, drill-downs to patient metrics
- Progress View: ROM, grip, and pain trends pre/post exercise
(Wireframes created in Figma; parallel patient/provider swim-lane flows guided IA and navigation.) :contentReference[oaicite:3]{index=3}
- Mobile client: iOS (Swift + SwiftUI) with TensorFlow Lite for on-device ROM analysis
- Backend: HIPAA-ready AWS stack (e.g., Lambda/EC2/Kubernetes) with Node.js or Python services
- APIs & Auth: OAuth 2.0 + OpenID Connect, Single Sign-On (SSO) with hospital IdP
- EHR Interop: FHIR APIs
- Pull: provider assignments, surgery details
- Push: PGHD metrics (ROM scores, pain levels) as FHIR Observation/Communication resources
- Data: validated on client/server; stored in encrypted cloud DB (e.g., RDS/DynamoDB) with timestamps and plan linkage
A data architecture supports secure PGHD capture, therapist review, and EHR feedback loops. :contentReference[oaicite:4]{index=4}
- Encryption: TLS 1.3 in transit; AES-256 at rest (AWS KMS for key management)
- Access Control: RBAC by role (patient vs. therapist), MFA for providers
- Session Security: timeouts, secure token storage
- Auditing: immutable logs for PHI access/changes
- Retention: ephemeral handling of sensitive media (e.g., exercise videos) per policy
- Compliance posture: HIPAA-aligned safeguards and periodic security reviews
:contentReference[oaicite:5]{index=5}
- Literature scan shows superior engagement/outcomes when apps combine guided exercises, tracking, feedback, and therapist oversight; simplistic, non-interactive apps underperform.
- Task and cognitive analyses clarified requirements (e.g., basic smartphone skills, need for progress feedback, safe recording environment).
- Accessibility from the outset: high contrast, scalable typography, large touch targets, skip options for unavailable measures, and planned voice-over/audio guidance.
:contentReference[oaicite:6]{index=6}
- Participants: 2 OTs, 1 PT, 3 post-op patients
- Tasks: sign-in, pre-exercise survey, guided session (optional recording), post-session quiz
- Scores: avg ease-of-use ≈ 4.3/5
- Top revisions:
- Contextual pain anchors (0–10 descriptors)
- Slow-motion MP4 loops to clarify hand/finger positions
- Simplified quizzes (one Q per screen, immediate feedback)
:contentReference[oaicite:7]{index=7}
- Clinical validation and pilots across diverse patient populations
- Expanded accessibility: voice-to-text and text-to-voice controls
- Model hardening: robust, multi-context ROM estimation with on-device ML
- Deeper EHR integration: broadened FHIR resources and automated alerts
- Android client parity after iOS prototype maturity
:contentReference[oaicite:8]{index=8}
- Mobile: Swift, SwiftUI, TensorFlow Lite (on-device inference)
- Backend: AWS (Lambda/EC2/K8s), Node.js/Python services
- APIs/Interop: OAuth2, OIDC, FHIR (Observation, Communication)
- Data: RDS/DynamoDB (encrypted), PGHD intake & validation
:contentReference[oaicite:9]{index=9}
- Austin Cherian, Tsegie Kassahun, Maxwell Lewis — Design & Development
- Clinical & faculty collaborators referenced in the project report
:contentReference[oaicite:10]{index=10}
This README summarizes the HIDS 7007 Final Project documentation for HandRehab+. :contentReference[oaicite:11]{index=11}