The Centers for Medicare and Medicaid Services (CMS) collects data from hospitals and other healthcare facilities nationwide which results in a good problem—a wealth of information without an effective way to use it to assess quality at the hospital level. The state of Maryland had publicly reported data which they need to consolidate for a longitudinal data study where measure, program, policy, and requirements change over time.

Lantana’s analytics team gathered data from CMS’ Hospital Compare data archive from 2015 to July 2019, separated by fiscal year where applicable (i.e., per quarterly data update). They compiled hospital-level data as well as state- and national-level data for the mortality, readmission, patient safety and adverse event measures, healthcare associated infections (HAI), and patient satisfaction. The team compiled the data and created a file and data dictionary for each measure category, explaining the data including the policy and program context.

Our team then created a database for longitudinal analysis to detect trends at the facility, state, and national levels by transforming the raw, hospital-level data received in multiple formats. This project produced very large datasets (several million rows) on a tight timeline. Using SAS, we combined, transposed, and manipulated data to create downloadable database files documenting quality measure results and trends across a range of years.

The table above shows hospital performance data for five reporting periods.

Key outcomes:

  • Simplified tracking of hospital quality over time
  • Created a database from large datasets in multiple formats
  • Documented each file

Tools and Methods Used:

  • Transition scope matrix
  • Gap analysis
  • SAS