CASE STUDY
AI-Powered Referral Intake & Plan of Care Automation
A secure, HIPAA-compliant document intelligence platform extracting clinical faxes, structuring patient referral packets, and generating program-specific Plan of Care (PoC) documents across PCA, PPAC, and PDN programs.
Cognic IntakeAI
Search referral faxes, patient names or doctors…
Clinical Referral Document Parsing & Plan of Care Synthesis
1,200
Monthly Referrals Processed
< 15 Min
Intake to PoC Generation
99.1%
Clinical Extraction Accuracy
100%
Structured Denial Capture
| Referral ID | Referring Facility | Patient Name / Program | Clinical Status | AI Extraction | Confidence |
|---|---|---|---|---|---|
| REF-5510 | Baptist Health Hospital | Robert C. (PDN Skilled) | Accepted – High Priority | PoC Draft Ready | 99% |
| REF-5514 | Jackson Memorial Hospital | Elena S. (PCA Personal) | Accepted – Routine | Medications Extracted | 98% |
| REF-5519 | Mount Sinai Medical | Arthur M. (PPAC Pediatric) | Under Clinical Review | Missing MD Signature | 94% |
| REF-5524 | Kendall Regional | Lucia G. (PCA Personal) | Denied – Out of Coverage | Structured Denial Logged | 100% |
| REF-5530 | University of Miami Health | David W. (PDN Nursing) | Accepted – Complex Care | 485 CMS Form Pre-filled | 97% |
1,200+ Monthly Clinical Referrals
Extracted from multi-page PDF faxes and electronic intake emails
85% Faster Plan of Care Creation
Reduced preparation time from 2 hours per packet to < 15 minutes
Zero Retyping Demographic & Rx Extraction
Automatic parsing of medications, diagnoses, and physician orders
100% HIPAA Private VPC Processing
Compliant AI pipeline with zero training on sensitive patient data
What is automated clinical referral intake?
Automated clinical referral intake uses Document AI, OCR, and medical entity extraction to parse incoming referral faxes, validate patient insurance eligibility, extract diagnostic and medication lists, and automatically populate program-specific Plan of Care (PoC) documents (e.g., CMS-485 forms) for clinician sign-off.
A Smarter Way to Onboard Patients into Care
The Challenge
Clinical intake staff manually reading 20-30 page scanned referral faxes, re-typing patient details, and building care plans by hand.
The Solution
HIPAA-compliant document parsing pipeline with automated medical entity extraction and pre-populated Plan of Care templates.
The Outcome
Turnaround accelerated from hours to minutes, medical re-typing errors eliminated, and referral acceptance capacity tripled.
The Challenge
Home health agencies face severe intake bottlenecks when medical referral faxes arrive as multi-page, low-quality scanned PDFs with messy clinical handwriting.
Manual review of 20–40 page clinical referral faxes from hospitals and physicians
Hours spent manually searching for medications, ICD-10 diagnoses, and physician orders
Repetitive manual re-entry of patient demographics into Plan of Care (PoC) software
Distinct formatting and compliance requirements for PCA, PPAC, and PDN care programs
Lack of structured tracking for referral denials and inability to analyze lost referrals
Delays in obtaining physician signatures on completed Plan of Care documents
High risk of HIPAA non-compliance if clinical staff use unauthorized consumer AI tools
Care delivery delays while patients wait for administrative intake clearance
Manual Referral Processing vs. Cognic AI Intake Platform
| Workflow Stage | Traditional Manual Workflow | Cognic AI-Powered Engine |
|---|---|---|
| Fax / Email Ingestion | Manual download and paper printing of 30+ page faxes | Automated digital fax & email intake with Azure Document Intelligence OCR |
| Data Extraction | Intake nurses scanning pages with highlighters to find diagnoses & Rx | Semantic Clinical AI extracting demographics, vitals, allergies, and medications |
| Plan of Care Drafting | Typing clinical summaries into CMS-485 forms and Word templates from scratch | Auto-generation of program-specific PoC documents pre-filled from extracted data |
| Denial Management | Denials communicated verbally or lost with no structured tracking | Structured denial reason taxonomy capturing coverage, staffing, and clinical grounds |
| Physician Signatures | Physical faxing back-and-forth for required physician sign-offs | Integrated digital e-signature workflows with automated physician reminders |
| HIPAA Security | Physical papers left on desks or uploaded to unvetted cloud tools | Encrypted private enclave architecture with strict BAA compliance and audit trails |
The Cognic Clinical Referral Engine
Three specialized modules accelerating patient transition from referral to active care.
01
Clinical Document AI Parser
Processes low-resolution faxes, clinical narratives, and handwritten physician notes to extract structured medical entities.
• Medical OCR & handwriting recognition
• ICD-10 & medication parsing
• Automated document classification
→
02
Dynamic Plan of Care Generator
Maps extracted clinical data into standardized Plan of Care templates tailored for PCA, PPAC, and PDN home health programs.
• Program-specific template mapping
• CMS-485 compliance validation
• Care frequency & goal drafting
→
03
Intake Workbench & Signature Hub
Empowers intake coordinators to verify extracted data, resolve missing items, and route documents for digital physician sign-off.
• Side-by-side fax review UI
• Automated physician signature alerts
• Structured denial tracking analytics
→
Intake Nurse Review
Side-by-side verification against original fax
CMS-485 Auto-Fill
Pre-filled standardized care plans
Physician E-Sign
Automated digital signature reminders
Referral Processing Architecture & Pipeline
An automated, HIPAA-compliant document intelligence pipeline.
1 Fax / Email Ingestion
› Digital Fax Listener
› Secure Email Intake
› Multi-Page Splitter
› Image Enhancement
› Document Archiving
2 Document AI Parsing
› Azure Document AI OCR
› Handwriting Recognition
› Entity Extraction (Rx, Dx)
› Insurance Validation
› Confidence Scoring
3 PoC Synthesis
› PCA / PDN Template Map
› CMS-485 Generation
› Frequency & Goal Rules
› Missing Item Detection
› Intake Nurse Workbench
4 Signatures & EHR
› Physician E-Signature
› Automated SMS / Email Alerts
› EHR Direct Write
› Structured Denial Store
› HIPAA Audit Trail
From Referral Fax to Authorized Care Plan
A verified clinical data journey from hospital discharge to active home care delivery.
Inbound Fax
→
AI Extraction
→
Nurse Check
→
PoC Generation
→
Physician Sign
→
Active Care
Clinical Document Intelligence Layer
Specialized AI capabilities powering patient referral intake.
Medical Handwriting & Fax OCR
High-precision recognition of handwritten clinical notes and low-res faxes
Medication & Diagnosis Parsing
Extracts full medication schedules, dosages, and corresponding ICD-10 codes
Dynamic CMS-485 Form Population
Automatically fills Medicare/Medicaid Home Health Certification forms
Structured Denial Reason Engine
Classifies and archives denial causes to optimize hospital referral relationships
Physician Digital Signature Routing
Coordinates secure e-signatures with automated reminders and compliance logging
Clinical Intake Coordinator Workbench
Intake nurses review extracted patient data side-by-side with original referral faxes before generating final care plans.
| Referral ID | Hospital & Patient | Extracted Diagnoses | Plan of Care Target | Action | ||
|---|---|---|---|---|---|---|
| REF-5510 | Baptist Health / Robert C. | I50.9 Heart Failure, E11.9 Diabetes | PDN Skilled Nursing (CMS-485) | PoC Ready | 99% | |
| REF-5514 | Jackson Memorial / Elena S. | M16.1 Bilateral Osteoarthritis | PCA Personal Care Plan | PoC Ready | 98% | |
| REF-5519 | Mount Sinai / Arthur M. | G80.0 Spastic Cerebral Palsy | PPAC Pediatric Care Plan | Pending MD Sig | 94% | |
| REF-5524 | Kendall Regional / Lucia G. | Z96.651 Right Knee Arthroplasty | PCA Care – Out of Coverage | Denied & Logged | 100% |
HIPAA & Medical Document Security
Enterprise safeguards protecting protected health information.
HIPAA BAA Compliant
SOC 2 Type II
Private VPC Processing
Zero PHI LLM Training
AES-256 Bit Encryption
Granular Access Control
Technology Stack
Key technologies powering the production deployment.
Azure AI Document Intelligence
Python FastAPI
OpenAI Private Enterprise (BAA)
React.js
PostgreSQL
Docker Containers
PDFKit & Docx Engine
Twilio Signature Alerts
TailwindCSS
What Changed
Key operational and efficiency improvements.
Referral intake and Plan of Care drafting time dropped from 2 hours to < 15 minutes
Clinical intake coordinators expanded capacity from 400 to 1,200+ monthly referrals
Zero manual data-entry errors in patient medication schedules and diagnoses
Structured tracking of 100% of referral denials to pinpoint growth opportunities
Physician signature turnaround accelerated by 65% via automated digital routing
Built for Home Health Leadership
Engineered for clinical directors, intake supervisors, and agency executives.
Clinical Intake Directors
Accelerate referral-to-admission cycles and eliminate administrative charting backlogs.
Home Health Agency Executives
Capture more hospital referrals and scale agency revenue without adding intake headcount.
Chief Nursing Officers (CNOs)
Ensure rigorous clinical accuracy and standardized care planning across all nursing programs.