Overview 

This video provides a technical walkthrough of the INA (Initial Nursing Assessment) AI QA Beat feature within the HospiceWorks platform. Designed to mitigate nurse burnout and ensure high-fidelity documentation, this tool acts as a clinical co-pilot during the admission process, ensuring all documentation is audit-ready and compliant with 2026 regulatory standards.

 

Workflow & Performance 

The AI QA review is integrated directly into the assessment interface for seamless clinical use:

  • Activation: Users click the AI Bot icon on the right sidebar to expand the review panel.
  • Initiation: Clicking "New Document Review" triggers a comprehensive analysis of the assessment data.
  • Latency: The engine processes the full data payload in 60–90 seconds, providing real-time feedback before the clinician leaves the bedside.
  •  

Core Functional Modules 

The review is categorized into three actionable segments:

  1. Compliance: Identifies documentation gaps, validates terminal diagnoses against LCD criteria, and reviews ICD-10 code suitability for claim acceptability.
  2. Guidance: Provides a draft Plan of Care (POC) with recommended interventions for pain, nutrition, and safety. It suggests interdisciplinary assignments and DME needs based on structured assessment findings.
  3. Draft Summary: Synthesizes assessment data into professional narrative summaries, including functional decline, symptom burden, and eligibility support.


Technical Integrity: The "Evidence Use" Approach 

Our system utilizes a Rules-Based Prompting Engine to ensure deterministic clinical logic. By processing structured data through a comprehensive, hospice-specific prompt, the AI avoids "hallucinations" and remains grounded in the facts provided by the clinician. Every recommendation includes an "Evidence Used" breakdown, allowing the nurse to see exactly which data points triggered the AI’s logic.


Clinical Accountability & Disclosure 

HospiceWorks adheres to a "Human-in-the-Loop" philosophy. The INA AI QA provides recommendations only. All generated content—including care plans and narratives—must be reviewed, edited, and approved by the licensed clinician. The clinician remains the final authority and author of the medical record.