Chart Abstraction

Sources and review notes

These links support definition-level research and do not establish the regulatory status, safety, or suitability of any product.

CMS describes chart-abstracted measures in Hospital Quality Reporting as patient-level data used to calculate and validate program-specific quality results. CDC explains that cancer incidence data begin with staff abstracting information from patient medical records into registries, and NCI's SEER training emphasizes that abstractors must apply the reporting and coding rules of the registry or standard setter for the case. These sources demonstrate record review, defined data elements, downstream reporting, and validation in specific programs; they do not establish one universal abstraction schema, prove that AI extraction is accurate, or authorize reuse for coding, payment, prior authorization, research, or clinical decisions. Teams must define the source record set, criteria and version, provenance, missing and conflicting information handling, qualified reviewer role, sampling and re-abstraction method, error thresholds, PHI controls, and downstream correction process for each use.

• Source 1: security.cms.gov/pia/hospital-quality-reporting
• Source 2: www.cdc.gov/united-states-cancer-statistics/technical-notes/incidence-data-sources.html
• Source 3: training.seer.cancer.gov/abstracting/intro/