How to start using Hospital Compare to report on hospital performance near you

If you’re like me, you are alarmed by the rates of complications and deaths from sepsis worldwide. You want to shine a light on the problem in the hopes we can better address the underlying causes of sepsis and better deploy known and effective treatments.

As a reporter, where do you start?

The total picture of how your health system handles — or, in some cases, causes — the full range of fatal and debilitating illnesses and injuries can be hard to grasp. There are a variety of datasets, though, that can give you clues and glimpses of how well hospitals near you are doing, both in terms of the underlying causes of sepsis and efforts to combat it once it is diagnosed.

In this post, I will walk you through a comparison of two hospitals using data from Medicare. And I will share some important caveats along the way.

First, if you have not used Hospital Compare before, it’s worth giving yourself a good hour to just play around with the site and read some of the introductory and explanatory material. Like all data query tools, it has its flaws, but prior to its existence there was no simple way to make sense of this massive data set. (One flaw is that it seems to work better with some browsers than others — I would use Google Chrome.)

Depending on your data skills, you may want to connect with the Centers for Medicare and Medicaid Services directly about gaining access to the larger data set for you to conduct your own custom queries. The Arizona Republic, for example, took Medicare data and built its own interface for readers to view hospital quality scores.

For this exercise, I went to Hospital Compare and plugged in a ZIP code: 06501. (I had been talking with a colleague about their Yale undergraduate days and so thought I would check out hospitals in the Yale area.)

I hit “search” and the site returned 13 different hospitals. Which to choose? The first thing you should do is decide what you’re trying to compare. Because I knew I was interested in hospitals that were likely to deal with a lot of infections and sepsis, I knew that I wasn’t as interested in behavioral hospitals or psychiatric hospitals. So I skipped those. I also knew that I wanted to be able to have a good apples-to-apples comparison at the end, so I skipped any hospital that did not have an overall quality score. Those are the little one- to five-star ratings you will see alongside the hospitals’ names. (If they don’t have an overall quality score, it usually means that Medicare did not have enough information to be able to give them a score.)

That left a subset of nine hospitals. Because Hospital Compare only allows you to compare three hospitals at a time, I decided I would try out three different hospitals with three different overall quality scores. So I chose Waterbury Hospital (one star), St. Vincent’s Medical Center (two stars), and Midstate Medical Center (four stars).

I hit “Compare Now” and was taken to a page of results.

I wanted to explore the causes of sepsis first and so I clicked on the “Complications & Deaths” tab. Then I clicked on “Infections.” As Hospital Compare helpfully explains:

Healthcare-associated infections, or HAIs, are infections that people get while they’re getting treatment for another condition in a healthcare setting. HAIs can occur in all settings of care, including acute care hospitals, long term acute care hospitals, rehabilitation facilities, surgical centers, cancer hospitals, and skilled nursing facilities. Many of these infections can be prevented through the use of proper procedures and precautions.

Then I clicked on “Show Graphs.”

Here’s the first caveat. You will receive a comparison of hospitals that in all likelihood will show that all of the hospitals you have chosen are kind of the same. That’s because Medicare takes pains to downplay any variation in these outcome measures. At times it can seem as if a hospital would have to be killing people in the street and burying them in the hospital parking lot in order to get a bad score. In this case, each of the three hospitals is labeled as “No Different than National Benchmark.”

The charts you will see are all standardized on a one to five scale. The actual measures tell a different story. Click on the “View More Details” button. There you get the underlying data. For Waterbury, there were seven health care-associated infections in the reporting period over 5,549 patient days, meaning the number of days patients were catheterized in the sample population studied. Midstate, by contrast, had one infection for 4,340 patient days, and St. Vincent’s had five for 3,886.

How do you make sense of all that? Hospital Compare puts the numbers in terms of a ratio: the actual number of infections compared to the predicted number of infections. Midstate’s ratio was the lowest, and the other two were quite comparable. This same kind of comparison can be made across other types of infections, including: central line-associated bloodstream infections, surgical site infections from colon surgery or abdominal hysterectomy, methicillin-resistant staphylococcus aureus infections, and clostridium difficile intestinal infections. These can all be causes of sepsis.

Now let’s look at what happens when someone develops sepsis. In 2018, Hospital Compare added sepsis to its measures of “timely and effective care.” Dr. Steve Claypool wrote a good piece for STAT News about the sepsis data in the Hospital Compare site when it was launched, and the site has been updated since his piece ran:

To help ensure timely, consistent, and high-quality care for sepsis patients, the Centers for Medicare and Medicaid Services adopted in 2015 the Sepsis National Hospital Inpatient Quality Measure (SEP-1) that had been developed by the National Quality Forum. This metric assesses hospitals’ timely treatment of sepsis, which costs more than $27 billion annually.

So how do these three Connecticut hospitals stack up on this metric?

Up at the top of the results, click on the “Timely & effective care” tab. Then click on “Sepsis care.” As Hospital Compare explains:

Sepsis is a complication that occurs when your body has an extreme response to an infection. It causes damage to organs in the body and can be life-threatening if not treated. Sepsis can sometimes turn into septic shock, which has a higher risk of death. Identifying sepsis early and starting appropriate care quickly increase the chances of survival.

The graphs here tell two stories.

The first story is that there is a pretty big difference among these hospitals in how they handle sepsis. St. Vincent’s Medical Center treats a higher percentage of patients with timely and effective care than the other two hospitals, than the state average, and than the national average. Waterbury Hospital’s rate of providing “timely and effective care” is fully half that of St. Vincent’s. Not a great place to be if you get sepsis.

But here’s the second story. None of these hospitals are winning the day. Even at St. Vincent’s, one out of every four patients is not receiving timely and effective care. At Waterbury, nearly two thirds of all patients do not receive “timely and effective care,” according to Hospital Compare. In Connecticut, just about half of all patients do not receive the right care for sepsis. To be clear, no hospital in the country has hit 100%. As the site points out, the top 10% of all hospitals achieved a rate of 80%, and so St. Vincent, at 73% probably feels pretty good about where it stands.

Your job as a reporter, though, is to give your audiences the full picture. Averages and benchmarks are important. But identifying steps that hospitals could take to make health care safer and better for everyone is even more important.