Computer Assisted Coding Services

The Vendor is required to provide enterprise computer-assisted coding (CAC) and core coding infrastructure solution to modernize and streamline medical coding and documentation improvement processes across the health system.
- Computer-assisted coding (CAC) automation
1. NLP-driven code suggestion:
• Use natural language processing or similar ai to analyze clinical documentation and suggest appropriate icd-10-cm diagnosis codes, icd-10-pcs procedure codes, CPT/HCPCS codes (outpatient/professional), and e/m levels.
• Emphasis on accuracy and completeness – the system should capture both primary and secondary codes and any applicable modifiers
2. Autonomous coding capability:
• Provide a foundation for autonomous coding of straightforward cases (with the option to auto-approve certain encounters).
• The system should allow configuration of confidence thresholds or rules that determine when an encounter can be automatically coded vs. routed to a human coder.
• Human-in-the-loop controls are essential: e.g., ability to set those complex cases (trauma, transplants, etc.) Always require manual review, regardless of AI suggestion
3. Capture-to-code workflow:
• Solution supports a capture-to-code workflow, including how clinical documentation is ingested, interpreted, converted into suggested codes, reviewed, accepted, rejected, or routed for follow-up.
4. Code sequencing and grouping:
• Automate code sequencing logic (selection of principal diagnosis, principal procedure, etc., per official guidelines).
• The system must handle grouping of codes into MS-DRG/APR-DRG for inpatients, but within the CAC engine, it should already prioritize correct principal diagnosis/procedure to drive grouping.
• The solution must include logic to ensure the most impactful codes are prioritized and sequenced such that they are included within the claim submission limit (e.g., top 25 codes), with emphasis on reimbursement, compliance, and clinical accuracy.
• There must be transparency in how the sequencing is determined.
5. Continuous learning:
• The solution should improve over time. Describe how the system learns from coder corrections or new data. If ml-based, clarify how model updates are delivered (e.g., periodic retraining, client-specific tuning). The feedback loop from coding outcomes (including denial feeds or audit results if available) would be a plus to fine-tune accuracy.
- Clinical documentation integrity (CDI) support
• Integrated CDI features and workflow, including case prioritization, auto-suggested DRG Assignments, query alerts, collaboration with coding, and reporting of CDI intervention impact.
• Included CDI reference content, best-practice guidance, and whether the solution supports adult and pediatric documentation concepts.
• Functionality for case prioritization, including factors used, scoring logic, user visibility, and whether prioritization can be configured by health.
• Solution uses NLP or other methods to assign working MS-DRG or APR-DRG without requiring CDI staff to manually code the case concurrently.
• Solution supports concurrent review and improves CDI efficiency in chart review.
- CDI-coding collaboration and physician query support
• Offer a mechanism for CDI and coding to reconcile code sets and DRGs before billing.
• Physician query routing, query templates, pediatric-specific query content, pended-case tracking, provider response capture, escalation, and reporting.
• Automated or AI-suggested query content, including how suggestions are reviewed, governed, and integrated into epic physician workflow.