## Jun 3 Understanding U.S hospital billing practices - part 2 - chest pain

I have chest pain, but not because of what you might think. Although if it were that serious, I might go to the hospital and get it checked out. And the cost would vary wildly - depending upon where the hospital is located. And that's how you segue. Today we're going to take a gander at the cost of treating patients with Chest Pain in New York State. As a refresher, we're analyzing recently released Medicare cost data. I previously introduced the topic, ad nauseam, so I'll spare you the deets. However, I also surreptitiously inserted various useful links in this paragraph that you can click on to learn more.

Before continuing, I'd like to note that for the medically inclined out there, when I say "chest pain" (or "diabetes" or "respiratory"), I'm referring to the Medicare Severity Diagnosis Related Group (MS-DRG). These medical categories can be searched on here.

With those caveats out of the way, let's visualize chest pain - datafied. Below we investigate every hospital in NY that treated chest pains in 2011 and check out the amount that the hospital billed (either to the patient or Medicare) and the amount that Medicare covers. The difference is presumably covered by the patient, and the role insurance plays in this is not clear.

Without pulling an Elaine (we'll do that below), it looks like there's quite a bit of variation in the amount hospitals are billing for treating chest pains. For some perspective, the first quartile (the lower 25%) charges below $8,300, the upper quartile charges above$15,000, and median hospital charge is roughly $11,000. It looks like the hospitals that charge more, charge about 2 times more on average than the less expensive ones. For the same procedure. In the same state. What are the Medicare trends like? This chart looks deceptively similar. I say deceptively because if you notice the y-axis, the spread is from a minimum of$3,000 or so, to a maximum of $8,000 (we call this difference the range). Medicare, therefore, pays out much more evenly than what hospitals are charging. In dollars, Medicare varies up to$5,000 depending on the hospital, whereas hospital charges vary up to \$21,000 - or about 4 times that of Medicare. Wow. Crazy. Superlative.

This is roughly what one would expect from government payments - that is, to be less variable than individual hospital charges. However, is 4 times reasonable? My sense is that no one really knows given all the noise that's in this data. Still, something to think about when people complain that Medicare preferentially treats people from certain areas more than others...judged by how much hospital charges vary, Medicare doesn't rate that poorly!

Finally, to emphasize the discrepancy let's position each hospital in the above charts side by side.

The difference is palpable.

Next, let's name names. Specifically, let's identify the hospitals (and corresponding cities) on the extremes. That is, the top 10 and bottom 10 Medicare coverage to hospital payment discrepancies.

Interesting - the top 3 cities - Mineola, Patchogue and Smithtown are located around Long Island. So certain hospitals around these cities tend to charge the most compared to what Medicare covers.

What about the other side of the coin - the hospitals and cities which charge less (or as much) as Medicare covers.

(Note that a negative bar indicates that Medicare is paying MORE than the hospital charges)

Huh. 7 of these 10 hospitals are located in the Greater New York area, and the other 3 are near the Canadian border - that is, the other side of the state. So we see hospitals in a similar geographic area (Long Island isn't far from New York City) on opposite ends of the cost/coverage spectrum and hospitals on opposite sides of the state, near each other on the cost/coverage spectrum. Loosely speaking, I think this evinces the arbitrariness of hospital spending and Medicare coverage. I'm simplifying of course, but the point remains.

That's all for part 2. Any suggestions for the next installation?