Exploring the Depth of EDPS Data: A New Frontier for Accurate Risk Adjustment and Research Applications

Risk adjustment is the backbone of accurate payment models and population health management, especially in Medicare Advantage (MA) and other risk-based arrangements. While traditional claims data has been the go-to source for diagnoses and healthcare utilization, Encounter Data Processing System (EDPS) data from CMS represents a paradigm shift in both richness and accuracy. EDPS provides a more comprehensive view of patient diagnoses by incorporating incremental findings from retrospective chart reviews and encounter-level details that often go beyond what is coded in standard claims submissions.

This essay explores the technical and operational reasons why EDPS data offers superior diagnostic accuracy compared to claims data, how CMS provides deidentified EDPS data for research purposes, and the transformative use cases this dataset enables across market segments.

The Fundamental Differences Between EDPS and Claims Data

Claims data, though widely used, has inherent limitations. It is primarily designed for billing purposes, reflecting the procedures, services, and diagnoses deemed necessary to justify reimbursement. However, this process often excludes nuanced or secondary diagnostic information not required for billing. EDPS, in contrast, offers a more comprehensive lens into a patient’s health by combining claims data with retrospective chart reviews and encounter-level assessments.

Key Features of EDPS Data:

Comprehensive Diagnostic Capture:

• EDPS includes diagnoses not initially coded on the claim but identified through retrospective risk adjustment chart reviews.

• Diagnoses captured in EDPS must meet specific criteria: they must be monitored, evaluated, assessed, or treated (MEAT) as documented in unstructured clinical notes. This ensures that the data reflects meaningful clinical activity, not merely billing artifacts.

Inclusion of Chart Review Data:

• Chart reviews provide insights into unstructured encounter notes, enabling the identification of conditions missed in original claims coding. Examples include:

• Chronic conditions like diabetes or COPD that were discussed but not billed for during the encounter.

• Comorbidities identified in lab results, imaging, or consultations but omitted from claims submissions.

Encounter-Level Granularity: