Analyzing Medicare Coverage with Tableau

Here is a link to my most recent project on Tableau:

Medicare Inpatient Charge Data

I used the data found here:

Slide one

To begin the presentation, I included a list of the relevant terms that will be discussed along with their definition. This will help the viewer better understand the main ideas and aspects behind the upcoming dashboards.

Slide 2

This is the Summary dashboard. It organizes the data of over 3000 different hospitals by location as well as by the top 100 discharges. On this dashboard, viewers can select their desired state and view the list of hospitals in that state under the “Records” section. The list can be sorted alphabetically by hospital, number of cases per hospital, or by average total and/or covered charges by the hospital. Once a hospital is selected on the list, the map will show where that hospital is located. Furthermore, when a viewer selects a certain state, the top “x” DRG’s will be shown for that state. In the picture above, the “x” is set to 8; this number can be changed to anything between 1 and 20. You can see the Average Total Payments, Average Covered Charges, and Average Medicare payments as well. The main purpose behind this dashboard is to give the viewer a good understanding of the type of data that will be discussed throughout the rest of the presentation.

slide 3

This dashboard shows the total discharges around the nation. The purpose of this dashboard is to give viewers an understanding of where the most discharges tend to happen and what this might mean in terms of the charges. For example, we can see that most of the discharges happen in areas with high rates of employment; at the same time, high rates of discharges tend to happen in areas with low to mid per capita incomes.

Slide 4

Next we examine the providers. This dashboard show how hospitals compare in terms of costs. The parameter is set to look at the top providers in the state (range from 1 to 20). We can see when filtering by highest to lowest costs, there tend to be less total discharges in areas with the highest charges and more in the areas with lower charges. This raises the question: are these hospitals with high covered charges specializing in specific areas, possibly with high expenses? That seems to be a valid explanations as to why areas with higher charges have lower discharges.

Slide 6

This dashboard shows costs over time and how they compare with total discharges over time. As you can see with the use of the motion chart, as the total payments tend to rise, total discharges over time are going down. This can mean many different things; for example, there might be fewer patients getting sick or fewer patients may be receiving coverage. At the same time, the costs for these procedures seem to be rising because the charges and payments are rising. This may be due to a rise in cost for medical equipment or hospital services.

Slide 7

This last dashboard helps the viewer choose a provider based on his or her discharge and location. The viewer can select a discharge and see which providers have the highest covered charges, Medicare payments, or total payments. They can then select a specific state if they want to see which the top in the area is, or just select a hospital right away to see which the top in the country is. The map below will display where that hospital is located. The picture above is examining the cellulitis discharge in Louisiana.


Limitations: Although I was presented a large amount of data, I was not given specific dates for discharges or payments; I was only given the year. Therefore, I was unable to build a visualization that could effectively analyze the discharges and payments over specific time frames. This would have been useful in creating forecasts.

About Pradheeth R

"Data is the new oil? No, data is the new soil." - David McCandless, TEDGlobal 2010 Tableau Data Scientist/Analyst at Novedea Systems. Southern Methodist University graduate with a double major in Economics and Psychology.

Leave a Reply