National Quality Rating Database


The Delta Group’s National Quality Rating Database (NQRD) includes virtually all general, acute, non-federal U.S. hospitals. The database incorporates the most recent three (3) years of federal fiscal year data (October to September) which runs approximately two (2) years behind the current calendar year. The NQRD includes information from the following publicly available datasets:

  • Hospital Quality Alliance (HQA) Hospital Compare Database (All-Payer)
  • Centers for Medicare and Medicaid Services (CMS) Medicare Provider Assessment and Review (MedPAR) File

The NQRD includes the following data elements from the HQA Hospital Compare Database:

  • Core Process of Care Measures for:
    • Acute Myocardial Infarction (AMI)
    • Heart Failure (HF)
    • Pneumonia (PN)
    • Surgical Care Improvement Project (SCIP)
  • Patient Satisfaction Measures for:
    • Overall Hospital Rating
    • Doctor Communication
    • Nurse Communication
    • Staff Responsiveness
    • Medication Explanation
    • Pain Control
    • Hospital Environment
    • Post-Discharge Information

The NQRD also includes the following data elements derived from the CMS MedPAR File:

  • Actual Mortality Rate
  • Actual Complications Rate
  • Actual Patient Safety Indicator (PSI) Rate
  • Actual Inpatient Quality Indicator (IQI) Mortality Rate

Relying on data from the CMS MedPAR file, The Delta Group applied five (5) separate regression models which resulted in the addition of the following NQRD value-added data elements:

  • Expected Mortality Rate
  • Expected Complications Rate
  • Expected PSI Rate
  • Expected IQI Mortality Rate
  • Risk-Adjusted Mortality Index™ (RAMI™)
  • Risk-Adjusted Complications Index™ (RACI™)
  • Risk-Adjusted PSI Index™ (RAPSI™)
  • Global Risk-Adjusted PSI Index™ (G-RAPSI™)
  • Risk-Adjusted IQI Mortality Index™ (RIMI™)

Lastly, the NQRD provides state and national percentiles of performance for each quality measure using a z-value for determining the measure’s level of statistical significance and then calculating a standard score (z-score) to convert all data to a standard normal distribution.