NHSN Reporting - Demo & Help

Back to Dashboard

Overview

The NHSN Reporting module helps Antimicrobial Stewardship and Infection Prevention teams track and report:

  • AU (Antibiotic Usage) - Days of Therapy (DOT) and Defined Daily Doses (DDD) by antimicrobial and location
  • AR (Antimicrobial Resistance) - Resistance rates and phenotype tracking (MRSA, VRE, CRE, etc.)
  • HAI (Healthcare-Associated Infections) - Confirmed HAI events (CLABSI, CAUTI, etc.) for NHSN submission

Data is extracted from the hospital EHR (Epic Clarity) and formatted for NHSN submission.

Quick Start - Generate Demo Data

1. Generate Mock Clarity Data

Create demo patients with antibiotic orders and culture results:

# Navigate to nhsn-reporting module
cd /home/david/projects/aegis/nhsn-reporting

# Generate patients with AU/AR data
python -m mock_clarity.generate_data \
    --db-path ./mock_clarity.db \
    --patients 50 \
    --months 3 \
    --all-scenarios \
    --au-ar

This creates:

  • 50 patients with encounters across multiple locations
  • ~100 medication orders with MAR administrations (AU data)
  • ~50 cultures with organisms and susceptibilities (AR data)
  • Various CLABSI scenarios for HAI detection

2. View the Data

After generating data, visit the dashboards:

  • AU/AR Dashboard - Overview of usage and resistance
  • AU Detail - Detailed antibiotic usage by drug and location
  • AR Detail - Resistance rates and phenotype prevalence

AU (Antibiotic Usage) Reporting

Data Pipeline

  1. Extract - Query medication orders and MAR administrations from Clarity
  2. Map - Match medications to NHSN antimicrobial codes (VAN, TZP, MEM, etc.)
  3. Aggregate - Calculate DOT and DDD by location, drug, and month
  4. Export - Generate CSV formatted for NHSN submission

Key Metrics

Metric Definition Use
Days of Therapy (DOT) Count of unique calendar days a patient received any dose of an antimicrobial Primary NHSN measure - reflects duration of therapy
Defined Daily Dose (DDD) Total grams administered / WHO DDD reference dose Alternative measure - reflects total drug volume
DOT/1000 Patient Days DOT normalized by patient census Allows comparison across units of different sizes

Antimicrobial Categories

NHSN groups antimicrobials into categories for reporting:

Carbapenems MEM, ETP, IPM
Glycopeptides VAN
Beta-lactam/Inhibitor TZP, SAM
Fluoroquinolones CIP, LVX
Aminoglycosides GEN, TOB, AMK
Cephalosporins CRO, FEP, CAZ, CFZ

AR (Antimicrobial Resistance) Reporting

Data Pipeline

  1. Extract - Query culture results and susceptibilities from Clarity
  2. First Isolate Rule - Deduplicate: one isolate per patient/organism/quarter
  3. Calculate Rates - Percent resistance for each organism/antibiotic combination
  4. Detect Phenotypes - Identify MRSA, VRE, CRE, ESBL based on susceptibility patterns
  5. Export - Generate CSV formatted for NHSN submission

NHSN First Isolate Rule

To avoid counting the same infection multiple times, NHSN applies the "first isolate rule":

  • Only the first isolate of each organism species per patient per quarter is counted
  • Subsequent positive cultures for the same organism are excluded
  • Different species from the same patient are counted separately

Tracked Phenotypes

Phenotype Definition Pattern
MRSA Methicillin-resistant S. aureus Oxacillin R
VRE Vancomycin-resistant Enterococcus Vancomycin R
CRE Carbapenem-resistant Enterobacterales Meropenem R or Ertapenem R
ESBL Extended-spectrum beta-lactamase CTX R + CAZ R + FEP S (typical pattern)
CRPA Carbapenem-resistant P. aeruginosa Meropenem R or Imipenem R

HAI (Healthcare-Associated Infection) Reporting

The NHSN Reporting module also includes confirmed HAI events from the HAI Detection workflow.

HAI Types

HAI Type Status Description
CLABSI Implemented Central Line-Associated Bloodstream Infection
CAUTI Planned Catheter-Associated Urinary Tract Infection
VAE Planned Ventilator-Associated Events
SSI Planned Surgical Site Infection

HAI Workflow

  1. Detection - Rule-based screening identifies HAI candidates (BSI + device + timing)
  2. Classification - LLM extracts facts from notes, rules engine applies NHSN criteria
  3. IP Review - Infection Preventionist makes final determination in HAI Detection
  4. Reporting - Confirmed HAIs appear in HAI Detail and Submission

CLABSI Criteria (Quick Reference)

  • Central line in place for >2 calendar days on the date of event
  • Line was in place at time of BSI OR within 1 day of removal
  • Laboratory-confirmed bloodstream infection (LCBI)
  • BSI is not secondary to infection at another site

For detailed HAI detection guidance, see the HAI Detection Help page.

Denominators

NHSN requires denominator data to normalize rates across different unit sizes:

Denominator Used For Data Source
Patient Days AU (DOT/1000 PD), AR rates Census from ADT data
Central Line Days CLABSI rates Flowsheet documentation
Urinary Catheter Days CAUTI rates Flowsheet documentation
Ventilator Days VAE/VAP rates Flowsheet documentation

View and verify denominators on the Denominators page.

NHSN Submission

AU Submission (Monthly)

  1. Go to Submission page
  2. Select the reporting month
  3. Review the data summary by location and drug
  4. Click "Export CSV" to download the NHSN-formatted file
  5. Upload to NHSN via the Antimicrobial Use and Resistance (AUR) module

AR Submission (Quarterly)

  1. Go to Submission page
  2. Select the reporting quarter
  3. Review first isolates, resistance rates, and phenotype counts
  4. Click "Export CSV" to download the NHSN-formatted file
  5. Upload to NHSN via the Antimicrobial Use and Resistance (AUR) module

Export Format

The exported CSV files follow NHSN specifications:

# AU Export columns:
orgID, locationCode, summaryYM, antimicrobialCode, daysOfTherapy, patientDays

# AR Export columns:
orgID, locationCode, year, quarter, organismCode, firstIsolates,
antibioticCode, susceptible, intermediate, resistant

Architecture

+----------------------+     +----------------------+
|   Mock Clarity DB    |     |  Production Clarity  |
|   (SQLite demo)      |     |  (Oracle/SQL Server) |
+----------+-----------+     +----------+-----------+
           |                            |
           +-------------+--------------+
                         v
              +---------------------+
              |   AU/AR Extractors  |
              |   (SQLAlchemy)      |
              +----------+----------+
                         |
         +---------------+---------------+
         v               v               v
+----------------+ +--------------+ +----------------+
| AU Aggregation | | AR Analysis  | | Denominators   |
| (DOT, DDD)     | | (First       | | (Patient Days, |
|                | |  Isolate)    | |  Device Days)  |
+-------+--------+ +------+-------+ +-------+--------+
        |                 |                 |
        +--------+--------+---------+-------+
                 v                  v
        +----------------+  +----------------+
        |  Dashboard UI  |  | NHSN Export    |
        |  (Flask)       |  | (CSV/CDA)      |
        +----------------+  +----------------+

Demo Data Generation Options

Command-Line Arguments

python -m mock_clarity.generate_data [options]

Options:
  --db-path PATH       Database file path (default: ./mock_clarity.db)
  --patients N         Number of random patients (default: 50)
  --months N           Months of historical data (default: 3)
  --all-scenarios      Generate CLABSI test scenarios
  --au-ar              Generate AU/AR data (orders, cultures, susceptibilities)
  --au-encounters N    Number of encounters with AU data (default: 50)
  --ar-encounters N    Number of encounters with AR data (default: 30)

Example Scenarios

# Minimal demo data
python -m mock_clarity.generate_data --patients 10 --months 1 --au-ar

# Full demo with all scenarios
python -m mock_clarity.generate_data --patients 100 --months 6 --all-scenarios --au-ar

# Only AR data (resistance reporting focus)
python -m mock_clarity.generate_data --patients 30 --au-ar --au-encounters 0 --ar-encounters 50

Generated Organisms

The demo data includes realistic organism/susceptibility combinations:

  • S. aureus - MSSA and MRSA variants
  • E. coli - Susceptible and MDR variants
  • K. pneumoniae - Susceptible and CRE variants
  • P. aeruginosa - Susceptible and CRPA variants
  • E. faecalis/faecium - Susceptible and VRE variants
  • E. cloacae - With typical AmpC resistance

Troubleshooting

No data showing on dashboards

# Verify database has data
sqlite3 ./mock_clarity.db "SELECT COUNT(*) FROM ORDER_MED;"
sqlite3 ./mock_clarity.db "SELECT COUNT(*) FROM CULTURE_RESULTS;"

# Regenerate if needed
python -m mock_clarity.generate_data --au-ar

Database connection error

# Check environment variable
echo $CLARITY_DB_URL

# For demo, use SQLite:
export CLARITY_DB_URL="sqlite:////home/david/projects/aegis/nhsn-reporting/mock_clarity.db"

NHSN mapping missing

# Check antimicrobial mappings
sqlite3 ./mock_clarity.db "SELECT * FROM NHSN_ANTIMICROBIAL_MAP LIMIT 5;"

# Check phenotype definitions
sqlite3 ./mock_clarity.db "SELECT * FROM NHSN_PHENOTYPE_MAP;"

Restart the Flask application

# Kill existing process
pkill -f "gunicorn.*dashboard"

# Restart
cd /home/david/projects/aegis/dashboard
gunicorn -w 2 -b 0.0.0.0:8082 app:app

Links

Demo Environment: All patient data displayed is simulated. No actual patient data is available through this dashboard.