Visualizing Medicare - Comparative Analysis

In the United States, Medicare is a national health insurance program to which all Social Security recipients who are either over 65 years of age or permanently disabled are eligible. In addition, individuals receiving railroad retirement benefits and individuals suffering from end stage renal disease are eligible to receive Medicare benefits.

As number of Medicare beneficiaries touch the mark of 58.4 million (as of 2017) and number of hospitals providing Medicare services rise over 4000 across the country, it becomes important to analyze them.

Although an in-depth analysis can be done on a topic as important and interesting as Medicare; however in this article; I will touch only few but important data points to deduce a meaningful interpretation.

I chose geographical map for visualization of data over numbers in tabular form, because of 4 reasons:

  • Readers get a broader insight of data without going much into details
  • Easier for readers to grasp the gist of data
  • Simple to make comparisions with multiple data points
  • Spot outliers and data-points of interest, with ease

Besides, a block full of numbers is boring to read, right?

Goal of analysis in this article is to try answering questions about Medicare services. Questions like:

  • Are enough patients receiving benefits of Medicare services?
  • Which regions need more and immediate attention to improve Medicare services for patients?
  • Which counties need more hospitals with Medicare services?
  • Where are patients least satisfied with the hospitals that provide Medicare services?

It is easy to get lost in data as huge as we have, hence we choose 6 following target locations for comparative analysis:

  • Los Angeles (CA) and San Diego (CA) counties from the West coast
  • Fermont and Carbon in Wyoming
  • Burlington (NJ) and Queens (NY) from the East coast

All of these locations are very interesting from analysis point of view and we will see why. Anyone interested in analysing county of their interest, can do so using the geographical visualization in the same manner as we do here.

Geographical visualization has helped us locate regions which could be intersting. Now, in our next step, I will summarise the data related to these locations in a tabular form.

County # of medicares Population per medicare hospital EDV (patients annually) Left unattended Rating (out of 5)
CA, Los Angeles 82 123,945 High (40,000 - 59,999) 2.6% 2.9
CA, San Diego 16 208,605 High (40,000 - 59,999) 3.5% 2.7
WY, Fermont 1 11,889 Medium (20,000 - 39,999) 1% 1
WY, Carbon 1 2,674 Low (0 - 19,999) 23% 1
NJ, Burlington 4 112,149 Very High (60,000+) 3% 3.5
NY, Queens 7 336,940 Very High (60,000+) 2.4% 2.4

Comparing LA and San Diego
From table, we observe that a lot of patients visit San Diego medicare hospitals. This is also confirmed by high volume of patients visiting Emergency Department.

A whooping 3.5% of patients are left unattended. Considering large population of San Diego, 3.5% comes out to be a big number. No surprises that average rating of hospitals with medicare services in this region are low.

We can easily conclude that San Diego urgently needs additional hospitals providing medicare services.

Comparatively, LA has more Medicare centers (5 times more than San Diego) and hence population per medicare center is comparatively less. Compared to San Diego, LA has same volume of patients visiting Emergency Department (ED), yet a low percent of patients leave without being attended. Although a moderate yelp rating, LA seems to have right amount of medicare centers to handle the population in that region.

Now you can see, why we chose San Diego and LA as our two location of interest. Although neighbouring counties, yet data shows big difference in medicare services in these locations. LA has right number of medicare centers for its population, which results into more patients being attended, while San Diego lacks medicare centers for its population and hence more number of patients are left unattended.

Comparing Fermont and Carbon in Wyoming
We observe that although Carbon has a very low population for its only medicare center, it still has a very low rating and shockingly has a very high percentage of patients (23%) leaving without being attended.

Certainly Carbon is an outlier, and hence a deeper analysis and investigtion is needed to figure out the reasons of such low performance of this particular medicare center.

Comparatively, Carbon's neighbouring county, Fermont seems to have a decent performance. Fermont too has only 1 medicare center, however, its population is 5 times more than in Carbon.

We also observe that ED volume is higher in Fermont, yet only 1% of patients are left unattended, which is a good performance. This is completely different than its neighbouring county Carbon, and hence makes it a much more interesting location for investigation.

The yelp ratings of Fermont are low, but it could be because patients are not socially active and only rate hospital on Yelp when they are not happy with the services.

Comparing Burlington (NJ) and Queens (NY)
On East coast, we compare performance of medicare hospitals at Burlington; a county of New Jersey; and Queens; located in New York.

Both counties are extremely busy and we expect a lot of patients attending medicare centers. It will be interesting to see how medicare centers fare in these locations.

We observe that 4 medicare hospitals in Burlington are handling a fairly large number of population(112,149). The ED volume is very busy, which indicates that a lot of patients attend medicare hospitals on regular basis, as expected from a busy county.

However, what's interesting is, inspite of being very busy with patients, medicare hospitals at Burlington seems to be highly rated and only 3% of patients are left unattended. Although converting 3% into number, gives us a lot of patients unattended.

Hence we can conclude that medicare hospitals at Burlington have decent performance; although adding 1 more medicare center will result into more patients being attended with more care.

On the other hand we have Queens, which consists of 7 medicare centers handling even larger population per hospital (336,940, approx 3 times more than Burlington, which is already high enough).

One might expect staff and services at medicare centers in Queens to be extremely busy, and consequently more patients left unattended. However, only 2.4% of patients are left unattended (around 0.6% lower than Burlington).

This is a good indicator; however, we must consider that when converted into numbers, 2.4% of patients in Queens is a lot of patients unattended.

Hence, Queens does need addition of more medicare centers. A lower rating of hospitals with medicare services at Queens is expected, as it is not possible to attend each patient with best services when number of patients visiting hospital are so high.


The analysis gave us great insights into Medicare services at our target locations. One can further go into depths of analysis, by analyzing other features of a Medicare center. Features like; quality of technology used, budget allocated, readmission ratio etc. can be of greater value to draw interpretations.

As I mentioned earlier, feel free to use the visualization and analyze your county!!

Note, References & Links:
  • Data is for the year of 2017
  • Population per medicare is calculated by dividing total population of county with number of medicare in the county
  • Medicare services are for individuals above 65 years of age, hence total population cannot be a sole indicator for number of patients visiting medicare centers. However, it can be used to make a rough assumption, as exact data is not available.
  • Yelp ratings are ratings by individuals and hence cannot be used as a sole criteria for judging performance of a medicare center. However, yelp ratings do provide an idea of patient's satisfaction in general.
  • Medicare data
  • Yelp ratings data
  • Population data