To Gauge Hospital Quality, Patients Deserve More Outcome Measures

Patients, providers and the public have much to celebrate. Recently, the Centers for Medicare and Medicaid Services’ Hospital Compare website added central line-associated bloodstream infections in intensive care units to its list of publicly reported quality of care measures for individual hospitals.

Why is this so important? There is universal support for the idea that the U.S. health care system should pay for value rather than volume, for the results we achieve rather than efforts we make. Health care needs outcome measures for the thousands of procedures and diagnoses that patients encounter. Yet we have few such measures and instead must gauge quality by looking to other public data, such as process of care measures (whether patients received therapies shown to improve outcomes) and results of patient surveys rating their hospital experiences.

Unfortunately, we lack a national approach to producing the large number of valid, reliable outcome measures that patients deserve. This is no easy task. Developing these measures is challenging and requires investments that haven’t yet been made.

The addition of bloodstream infections data is a huge step forward. These potentially lethal complications, measured using Centers for Disease Control and Prevention’s methods, are among the most accurately measured outcomes. In addition, the science of how to significantly reduce these infections is mature, and hundreds of hospitals of all types and sizes have nearly eliminated them. A program to reduce these infections that started at Johns Hopkins Hospital was spread throughout Michigan, and is now being implemented throughout the U.S., demonstrating substantial reductions.

Achieving these results isn’t easy. It requires hospital CEOs to commit to a goal of zero infections, as well as unit and infection prevention leaders to work together to meet that goal. ICU clinicians need to take ownership of this campaign, recognizing that bloodstream infection measures are a canary in the coal mine—an indicator of how well a quality improvement system is functioning.

Prior to adding bloodstream infections to Hospital Compare, the public had patchy information regarding hospitals’ infection rates. A handful of states made the data public and in those that did, we noticed a concerning trend. While most hospitals achieved reductions, a small number had high rates and failed to improve.

This is largely from lack of leadership focus, not because their patients are sicker and more susceptible to pathogens. In our national work focusing on bloodstream infections, we have seen that all types of hospitals and ICUs can dramatically reduce rates (though burn ICUs have higher rates than others).

Now that the public has data on bloodstream infection rates, let’s hope they use it, and let’s hope we finally create a mechanism to develop the many other measures that patients deserve. The key is to ensure that the measures are valid and that they tell the whole story. Our past research has revealed that “the more you look, the more you find” for certain complications, such as harmful blood clots. We don’t want hospitals to look worse than others in the public eye simply because they worked harder to identify when their patients were being harmed.

Public reporting has risks, and it can do more harm than good when the wrong measures are used. Though bloodstream infection rates are not a perfect measure, they perform well. Given that these infections kill approximately 30,000 people each year in the U.S. (slightly fewer than the number of people who die from breast or prostate cancer) and they are almost entirely preventable, they have not received the attention they deserve. The Hospital Compare website can help shine a light on those hospitals that need to redouble their efforts.

What should patients do with this data? First, if they think they may end up in an ICU, they should investigate the infection rates of their local hospitals. If a hospital has a high rate, they should discuss it with their physician and weigh the risks and benefits of seeking care elsewhere. Even if you are not seeking care and your local hospital has a high rate, write to the hospital’s CEO and board of trustees to ask what they are doing to reduce these infections, to honor their commitment to serve the community.

Bravo CMS, let’s hope this trend continues.

Note: Due to a technical glitch, bloodstream infection data for The Johns Hopkins Hospital was not immediately available on Hospital Compare, but it will be in the future. In the meantime, bloodstream infection data is available via the Maryland Health Care Commission at this link.

Director of the Johns Hopkins Medicine Armstrong Institute for Patient Safety and Quality, Peter Pronovost, MD, PhD is a practicing anesthesiologist and critical care physician who is dedicated to making hospitals and health care safer for patients. Pronovost has chronicled his work in his book, Safe Patients, Smart Hospitals: How One Doctor’s Checklist Can Help Us Change Health Care from the Inside Out. His posts will appear occasionally on THCB and on his own blog, Points from Pronovost .

13 replies »

  1. Solution: Deal with the sham certification processes of the Joint Commission and health departments. They get many bonified complaints and do nothing, NOTHING!!! Oh, sorry, they give the illusion to the public that hospitals are safe, especially if it is a monopoly hospital that has the economic force to control all public and political proceedings in a region.

    Do ya think that a patient with sepsis is going to look up the infection rate in an ICU before going to the ER, Professor?

  2. Hey, southern doc, so if every measurement if flawed, then what’s your solution? Just let us go on assuring everyone our care is the best? Let’s hear solutions instead of the incessant quibbling with the ‘data’. Health care is the only industry, short of the software industry(EMR, anyone?), where black magic is still allowed to rule the day with the public. That has got to change, for both.

  3. “Mean scores of restaurants experiencing foodborne disease outbreaks did not differ from restaurants with no reported outbreaks”

    CDC Emerging Infectious Diseases 2004

    Data’s a bitch.

  4. No correlation between flunking the health department survey and food borne illness, Southern Doc says. Really? Data showing that? My guess: could be because no one goes into a “C” restaurant and the “D’/”E” ones are shut down.

    Or is this a made-up statistic? Like the physicians who assure me that “everyone knows” that almost all those victims of medical error were going to die soon, anyway. (Easy to believe that: just doesn’t happen to be true.)

  5. There’s no correlation between a restaurant’s health score and the rate of foodborne illness among their patrons (or the quality of their food).

    Beware of false end-points.

  6. Good post. I agree, quality, not quanitity should be measured. It’s the responsibility of each member of the healthcare team to make sure we are providing quality care. I agree also that restaurants are supposed to report their scores, so why shouldn’t hospitals. The public has a right to know, and chose were to get medical care based on that measure, the same way I won’t eat at a restaurant with a low score. Medicare is now watching hospitals closely, and if there is an infection acquired while in the hospital, they will not reimburse the hospital for the cost of treatment. Just another reason to pay closer attention to central line infections.

  7. “Quality” has gotten a bit moldy (and I’m a Senior member of ASQ), the new cliche buzz word is “Value.” Begging similar concerns, because, popularly,

    “Value = Quality / Cost”

    e.g., if my “quality’ (i.e. “outcome) = “1” and the expenditure is 1.00, then my “value” is “1.” But double the expenditure for the same outcome, now my “value” is 0.5. If Grandma’s hip job cost $43k and her outcome was… well, you get it.

    And, while “costs” are indeed “ratio-level” quantitative data (to the extent you can nail them down honestly per pt / episode of care, population. etc), you still have numerator problems. Mix & match your data types trying to look all analytically “progressive,” and, well, you get what you get.

    Which is more fog, really, good intentions notwithstanding.

  8. Well put, Peter. It is important to continue to highlight the failures (and successes) of leadership by individual hospitals, all of which have grown accustomed to camouflaging their true actions with careful releases of selected indicators on their website. And, of course, the “public” must be represented by employers, health plans, journalists and politicians keeping up the pressure.

  9. Absolutely agree with this post here. The lack of performance quality standards for hospitals is alarming and inefficiency reigns supreme. Another issue raised in this discussion is the definition of “quality.” As seen in the link below, quality is a highly subjective term that is defined differently throughout the healthcare system. One of the greatest difficulties we are faced with is to engage all parties in a discussion surrounding what quality means and how that can be measured.


  10. “they should investigate the infection rates of their local hospitals”

    Restaurants are required to visibly post their health inspection grades, why shouldn’t we get the same from hospitals on infection control.

  11. Thanks for focusing our attention on this important issue of oucomes and quality of care, Dr. Pronovost. As hospital settings are where the bulk of our health care dollars are spent, consumers and patients should have access to this data so that hospitals can be judged on how well they perform.