“Hospital Compare” Cube Implementation Wendi Welch
“Hospital Compare” Datasets WHAT IS HOSPITAL COMPARE? Hospital Compare has information about the quality of care at over 4,000 Medicare-certified across the country. You can use Hospital Compare to find hospitals and compare the quality of their care. The information on Hospital Compare: Helps you make decisions about where you get your health care Encourages hospitals to improve the quality of care they provide Maintained by Centers for Medicare & Medicaid Services (CMS) Official Datasets:
Who Else Cares About Hospital Compare? It’s not just for consumers (patients) This same data is used by Medicare internally to make incentive payments to providers of medical services based on the quality of care: Pay for Performance (P4P) Provides what are typically financial incentives to providers to improve the quality of the care they deliver and/or reduce costs. The model gives health care providers the chance for a financial upside – such as a bonus — but no added financial risk, or downside. By giving an incentive to providers for increasing the quality of care, Medicare reduces base payments for services provided Improving quality of care should reduce the complications/readmissions/deaths being paid for, resulting in a cost savings Competitive Advantages Surrounding or competing hospitals can use this data to compare and benchmark against
Business Objectives Opportunities to analyze the data for various attributes and find: Highest Death Rates by Region/State/Counties in Texas Highest Hospital Acquired Infections Rate by Region/State/Counties in Texas Lowest Patient Satisfaction Measures by Region/State/Counties in Texas Highest Rate of Complications by Region/State/Counties in Texas Highest Rate of Payment for Certain types of procedures by Region/State/Counties in Texas Highest Rate of Outpatient procedures by Region/State/Counties in Texas. Trends of Incentive Payments Etc…
Common Fields Existing between the Scoring Tables Provider Address/Location Score/Measure Measure ID Measure Name/Description National Comparison (for some) In addition there was a Hospital information table Data goes back to 2005 – Only 2013 on was imported Only certain tables/measures were used due to the voluminous nature.
Data Sample
Fields Created for Analysis Key Fields for each table Regions and Counties State Populations Measure Categories Time Fields
Some Required Transformations in the Data Measure IDs/Names changed throughout the years Hospital Names changed New Measures or Measures no longer used Descriptions have changed Some of the IDs had different formatting Measures were in different tables
Transformed Data
Snowflake Schema/Table Relationships
Fact Table Structures (only showing non-repetitive fields)
Dimension Table Structures LocationHospitals TimeMeasures National Comparison
SQL SERVER
Cube Structure and Data Source View
Time Dimension
Measures Dimension
Hospitals Dimension
Hospital Dimension (cont’d) Clusters – Or buckets were used for the population attribute so that analysis could be done on states within a population range.
Reports Just a few examples
Outpatient Volume For Consumers/Patients: Which hospital in Houston does the most of a certain type of outpatient procedure?
Highest Payment Rates
Patient Satisfaction (by hospitals in a particular city) Which Hospital scored the best for these measures?
Patient Satisfaction (by Region) Which Hospital scored the best for these measures?
Death Rates for Certain Measures
Death Rates for Certain Measures Using Population Buckets For those hospitals that received “worse than the national average”, does a state’s population size have any correlation to death rates? Does Ownership Type affect performance?
THANK YOU!