4. Exploring Medicare Hospital Readmission Rates with SQL

By /

After preparing the Medicare Hospital Readmission dataset, I explored it using SQL to better understand the data and answer several practical healthcare analytics questions.

Questions I Explored:

  • How do predicted and expected readmission rates compare across hospitals?
  • How many hospitals are included in the dataset?
  • Which states are represented and how many hospitals are in each state?
  • Which readmission measures appear most frequently?
  • What are the average, highest, and lowest readmission rates?

1. Total Number of Hospitals

SELECT COUNT(DISTINCT Facility_ID) AS TotalHospitals
FROM Hospital_Readmissions_Cleaned;

This query counts the number of unique hospitals included in the dataset.

2. Total Number of States

SELECT COUNT(DISTINCT State) AS TotalStates
FROM Hospital_Readmissions_Cleaned;

This query counts the number of states represented in the dataset.

3. Readmission Measures Distribution

SELECT 
	Measure_Name,
	COUNT(*) AS TotalRecords
FROM Hospital_Readmissions_Cleaned
GROUP BY Measure_Name
ORDER BY TotalRecords DESC;

This query shows how many records are available for each clinical readmission measure.

4. Average Excess Readmission Ratio

SELECT
    AVG(Excess_Readmission_Ratio) AS AverageExcessRatio
FROM Hospital_Readmissions_Cleaned;

Calculates the average excess readmission ratio across all hospitals.

5. Hospitals with Highest Excess Readmission Ratios

Identifies the 10 hospitals with the highest excess readmission ratios

SELECT TOP 10
    Facility_Name,
    State,
    Excess_Readmission_Ratio
FROM Hospital_Readmissions_Cleaned
ORDER BY Excess_Readmission_Ratio DESC;

6. Hospitals with Lowest Predicted Readmission Rates

SELECT TOP 10
    Facility_Name,
    State,
    Predicted_Readmission_Rate
FROM Hospital_Readmissions_Cleaned
WHERE Predicted_Readmission_Rate IS NOT NULL
ORDER BY Predicted_Readmission_Rate ASC;

This query returns the 10 hospitals with the lowest predicted readmission rates.

7. Number of Hospitals by State

SELECT
    State,
    COUNT(DISTINCT Facility_ID) AS Total_Hospitals
FROM Hospital_Readmissions_Cleaned
GROUP BY State
ORDER BY Total_Hospitals DESC;

This query groups the data by state and counts the number of hospitals in each state.

8. Average Predicted Readmission Rate by State

SELECT
    State,
    AVG(Predicted_Readmission_Rate) AS Average_Predicted_Rate
FROM Hospital_Readmissions_Cleaned
GROUP BY State
ORDER BY Average_Predicted_Rate DESC;

This query calculates the average predicted readmission rate for each state. The results can later be visualized using maps or bar charts in Power BI.

9. Comparing Predicted vs Expected Readmission Rates

SELECT
    Facility_Name,
    State,
    Predicted_Readmission_Rate,
    Expected_Readmission_Rate,
    Predicted_Readmission_Rate - Expected_Readmission_Rate AS Difference
FROM Hospital_Readmissions_Cleaned;

This query shows the difference between predicted and expected readmission rates for each hospital.

18330 Records

Dataset Summary Using SQL Aggregate Functions

SQL aggregate functions make it easier to summarize large healthcare datasets and identify useful patterns. In this post, we use COUNT, AVG, SUM, MIN, and MAX to better understand hospital readmission patterns.

-- 1. Total number of hospitals and records
SELECT 
    COUNT(DISTINCT Facility_ID) AS Total_Hospitals,
    COUNT(*) AS Total_Records
FROM Hospital_Readmissions_Cleaned;

-- 2. Average, Minimum, and Maximum Readmission Rate
SELECT 
    AVG(Predicted_Readmission_Rate) AS Avg_Readmission_Rate,
    MIN(Predicted_Readmission_Rate) AS Min_Readmission_Rate,
    MAX(Predicted_Readmission_Rate) AS Max_Readmission_Rate
FROM Hospital_Readmissions_Cleaned;

-- 3. Average Readmission Rate by State
SELECT 
    State,
    COUNT(DISTINCT Facility_ID) AS Number_of_Hospitals,
    AVG(Predicted_Readmission_Rate) AS Avg_Readmission_Rate
FROM Hospital_Readmissions_Cleaned
GROUP BY State
ORDER BY Avg_Readmission_Rate DESC;

These examples demonstrate how SQL aggregate functions can be used to summarize healthcare data.

Example:

SELECT State,
       AVG(Predicted_Readmission_Rate)
FROM Hospital_Readmissions_Cleaned
GROUP BY State;

What These Queries Teach:

  • How to summarize large healthcare datasets
  • How to compare readmission rates across states
  • How SQL queries can support Power BI dashboards

11. Key Insights

  • The dataset includes hospitals from across the United States.
  • Readmission rates vary by hospital and clinical measure.
  • Some hospitals have noticeably higher predicted readmission rates than others.
  • Grouping the data by state provides a useful regional summary.
  • Comparing predicted and expected readmission rates helps identify differences that can be explored further.

Scroll to Top