UC32: Patient Satisfaction (HCAHPS) Analyser

The Problem

HCAHPS (Hospital Consumer Assessment of Healthcare Providers and Systems) scores drive CMS Value-Based Purchasing (VBP) reimbursement adjustments. Low HCAHPS scores can reduce hospital reimbursement by 2 to 3% annually. Survey responses are rich with free-text feedback (e.g., "nurse was rude," "wait time was too long"), but manual thematic analysis is labour-intensive. Many improvement opportunities are missed due to limited review capacity.

What the Agent Does

Processes HCAHPS survey responses in batch mode (monthly or quarterly)
Performs sentiment analysis on free-text responses: identifies positive themes (excellent care, kind staff) and negative themes (long waits, poor communication)
Clusters negative themes by unit/department and shift pattern: identifies if specific units/shifts have higher dissatisfaction
Correlates HCAHPS themes with operational data: e.g., "long waits" theme correlates with ED census >90%, suggesting staffing issue
Generates improvement action items: prioritises by theme frequency and impact on overall HCAHPS score
Tracks outcomes: monitors if interventions (staffing increase, communication training) improve HCAHPS scores in subsequent surveys

Data Requirements

Data Sources:

HCAHPS Survey Responses , Quantitative (1 to 5 scale) and free-text responses, patient demographics, discharge unit, discharge date
Press Ganey or NRC Health Data Feed , HCAHPS data normalised from survey vendor systems
Operational Data (Epic ADT, staffing systems) , Unit census, staffing levels, length of stay by unit and shift
Patient Demographics , Age, acuity, primary diagnosis, insurance type (may correlate with satisfaction)
Historical HCAHPS Scores , 12 to 24 months of data to track trends

Data Classification:

HCAHPS responses: Sensitive (may identify patients), Internal, Unstructured (free-text)
Survey metadata: Internal, Structured
Operational data: Internal, Structured
Patient demographics: Sensitive, Internal, Structured

Data Quality Requirements:

HCAHPS responses: complete and final before analysis (survey period closes 30 days after discharge)
Survey data: verified for missing/anomalous responses (Press Ganey validation)
Operational data: accurate and current (ADT, staffing logs)
Historical data: 12 to 24 months minimum for trend analysis

Integration Complexity: Medium

HCAHPS data feed from survey vendor (Press Ganey, NRC Health API)
Natural language processing (NLP) for theme extraction from free-text
Operational data integration: ADT system, staffing system
Sentiment analysis model (can use pre-trained models or custom tuning)

Score Breakdown

Criterion	Weight	Score (1-5)	Weighted
Time Recaptured	15%	4	0.60
Error Reduction	10%	4	0.40
Cost Avoidance	10%	3	0.30
Strategic Leverage	5%	4	0.20
Data Availability	15%	3	0.45
Process Clarity	15%	3	0.45
Ease of Implementation	10%	3	0.30
Fallback Available	10%	4	0.40
Audience (Int/Ext)	10%	4	0.40
Composite	100%		3.70

Why It Scores Well

HCAHPS analysis directly impacts reimbursement: improving HCAHPS scores by 2 to 3 percentile points can reclaim $100K to $1M in VBP adjustments for large hospitals. The data is rich (free-text feedback); NLP can extract actionable themes. Outcomes are measurable (HCAHPS score trends, improvement action success).

Regulatory Alignment

CMS Value-Based Purchasing (VBP) Program: HCAHPS scores are one of four domains (cost, mortality, readmission, patient experience) affecting reimbursement
HIPAA Privacy: Patient identifiers in surveys must be handled carefully; de-identification recommended before analysis
Patient Privacy: Free-text responses may reveal sensitive information; access controls required

Sprint Factory Fit

Sprint 0 (2 weeks) + 2 build sprints (4 weeks)

HCAHPS analysis runs monthly or quarterly as a batch process: process new survey responses, generate theme reports, track improvement actions. The initial 2-week sprint focuses on HCAHPS data feed integration and basic theme extraction; a second sprint adds sentiment analysis, operational correlation, and action item generation. This is a medium-complexity use case suitable for quality/data analytics teams.

Comparable Implementations

Deploy This Use Case with the Sprint Factory

From zero to a governed, production agent in 6 weeks.

Sprint Factory Schedule a Briefing

Related Use Cases

Batch Claims Operations

Claims Auto-Adjudication Sentinel

Score: 4.3/5.0

Batch Finance & Actuarial

Risk Adjustment (HCC) Capture Agent

Score: 4.2/5.0

Batch Quality & Compliance

HEDIS Quality Score Monitor

Score: 4.15/5.0

Governance Risks to Consider

Before deploying this use case, review these agentic AI risks from the Corvair Risk Catalogue. Each is scored on the DAMAGE framework and mapped to regulatory expectations.