Summary: More than 10 years ago, the Institute of Medicine released its landmark report, To Err Is Human: Building a Safer Health System, which estimated that as many as 98,000 people die in the U.S every year as a result of preventable medical errors. Since then, the Agency for Healthcare Research and Quality, the Institute for Healthcare Improvement, and World Health Organization, among other groups, have actively promoted patient safety. Yet many physicians remain unengaged. Quality Matters asked one of the country's leading experts on patient safety what's holding up progress.
By Sarah Klein and Douglas McCarthy
Pronovost, now the director of Johns Hopkins University's Quality and Safety Research Group, doesn't like referring to the program as a checklist (though he uses the term occasionally) because it understates the research that went into its development, as well as the effort necessary to implement it in hospitals. To get results, health care organizations not only have to rigorously measure and report infection rates, they must identify and remove the barriers that prevent clinicians from taking these steps and foster a culture in which nurses feel comfortable questioning physicians who don't follow the protocol. Without an investment in each of these three steps, the protocol is ineffective.
To make this point, Pronovost often asks the following question of hospital leaders who tell him they are using the checklist but seeing their infection rates remain high: "If a brand-new nurse in your hospital were to see the senior-most doctor placing a catheter and not complying with this checklist, would the nurse speak up and would the doctor listen?" The most common response he gets: "I am laughed at, truly laughed at. They say, 'Are you nuts? Of course that wouldn't happen,'" Pronovost says. "In what other industry would this happen? We have an indisputable standard. Failing to comply with it kills people. Yet we are not comfortable having one worker question another about it? These infections kill more than 30,000 people—the equivalent of a 747 crashing every few days. If the U.S public knew these were needless deaths, there would—and should—be a public outcry."
For Pronovost, conversations with hospital leaders about the nature of medical culture help to explain—at least in part—why the U.S. has not made more progress in achieving patient safety goals more than a decade after the Institute of Medicine published To Err Is Human: Building a Safer Health System, which suggested as many as 98,000 people die in the U.S. every year as a result of preventable medical errors. Quality Matters asked Pronovost what other barriers he's identified as he's traveled the country helping hospitals implement strategies for improving safety and what he thinks it will take to surmount them.
QM: What's your assessment of the country's patient safety efforts to date?
Pronovost: We're not getting very far. I think the reason is our efforts have been competitive rather than cooperative. They have been independent rather than interdependent. And they have been focused on efforts rather than results.
QM: How is a lack of cooperation impeding our efforts?
Pronovost: I like to use the example of polio. Polio was killing about 350,000 people a year in the late 1980s. It was only by working together—through a global effort that involved many organizations—that we lowered the rate to around 1,300 cases in 2007. Safety and quality improvement have not had a similar impact. Part of the problem is we don't have a focal point that is as big as polio. We need one. The topic itself is less important than learning how to work together as a U.S. health system to solve a problem. We need to learn how to work together and apply that learning to other areas.
QM: How is the lack of cooperation problematic in your work?
Pronovost: I see it in my work on health care–acquired infections. [Pronovost is working to expand the use of the protocol that was studied in Michigan to hospitals in 38 states.] I'm not alone in that. When the U.S. House of Representatives held hearings on health care–acquired infections last year they noted one of the problems was the federal agencies don't work together. It is striking. It almost takes a broker to get different agencies to work together. We've played that role to some extent when working with the Centers for Disease Control and Prevention, the Agency for Healthcare Research and Quality, and the Centers for Medicare and Medicaid Services, all of which are working together under the leadership of the Dept. of Health and Human Services. We said "OK, CDC you need to step up and provide a mechanism to monitor these infection rates and, yes, you have your surveillance system, but you know what? Hospitals say it is too burdensome so you have to make it simpler." "And CMS, yes, you have to align payment, yet you need measures that are valid—that clinicians believe." Within the states we ask that the quality improvement organizations, the state health departments, and the state hospital associations work together… In some states, this is the first time these groups have talked together, even though they are all working on health care–acquired infections.
QM: How else does collaboration help combat health care–acquired infections?
Pronovost: When we started this work there was this tension, if you will, about how to measure quality and safety at the national level. There was a difference of opinion about whether we could use different measures than clinicians use at the bedside. Some would say we can use discharge data to get a signal of infection rates. But one of the things we learned in Michigan is that it's not going to work. We have to have a good clinical measure that is meaningful at the bedside and then scale it up to provide reports at a national level.
QM: What's the problem with using discharge data?
Pronovost: I think the biggest problem is that many clinicians don't believe it. And they are probably right not to believe it because it is pretty noisy. There's a second issue and that is whether discharge data is making it back to the clinicians who have to do the work to improve. So we are spending a lot of resources to collect information that is of questionable value. Through the collaboration we came to the conclusion that using central-line associated bloodstream infections as a measure is a better approach. Everyone is going to use it.
QM: Are there other ways collaboration has advanced your work?
Pronovost: Yes. There was also tension about whether to take a centralized (regulatory) or a free market approach to solving these problems. It's the same tension that we [as a country] are facing with financial reform or education reform. And one of the "aha moments" we had was that is a false choice. It is not an either-or. What we ought to do is centralize those pieces of this project that are most efficient and effective to do in a centralized way and leave to the market—in other words the individual hospitals and clinicians—to do what they do best, which is innovate. So we said we will centralize the way we measure infections (the CDC did that) and the evidence on how to reduce these infections, but we will encourage local innovation based on local culture. That approach has been remarkably well received.
QM: And how do you encourage interdependence among health care providers, insurers, and others when it is appropriate?
Pronovost: We're working with the Blue Cross and Blue Shield Association to create common incentives for reducing infections. We also want to establish the use of common measures for these infections. What's unique about the program is that would create incentives not only for hospitals to reduce infections, but also for the CEOs and medical directors of the health plans and hospital association executives. Our goal is to align everyone with a common measure they believe in and some incentives to work together. The incentive plan would most likely be modeled on what we did in Michigan, where hospitals were first rewarded for submitting data and then rewarded for demonstrating improvement. Hawaii was the first to pilot test this approach and they loved it. Their CEOs and medical directors are financially incentivized for the infection rates in the hospitals in their states. We also have interdependence within and among hospitals. Clinicians share best practices and implement what works in their area.
QM: What about the third piece? You said you're also concerned that the country has been focused on efforts rather than results. Can you give us some examples of this?
Pronovost: When the U.S. House of Representatives held its hearings they looked at hospital efforts to combat catheter-related bloodstream infections state by state. All 50 states reported that hospitals were using the checklist, but only 11 states reported the hospitals were formally measuring infection rates and none were anywhere near as low as the rates of infection achieved in Michigan. I think it shows, in my own belief, a complete accountability failure. The public does not care whether checklists are used or not; they care about not getting infected.
QM: What else is slowing progress on patient safety?
Pronovost: The lack of a scientific approach to quality improvement. In most quality improvement projects, they average about 60 to 80 percent missing data. You don't need to be a statistician to say about the only thing I can conclude when I have that little data is that I have a heck of a lot of missing data.
QM: How do you address this in your improvement work?
Pronovost: In Michigan, we sent the hospital CEOs a letters that said, in effect, "If you don't get missing data down, you will be kicked out of the collaborative. Don't waste your or our time." This was heretical in quality improvement because the common mantra was everybody just does his or her own thing with data. We said no, the public deserves better than this. You need to be accountable for doing this. In the end, nobody dropped out. As reported in our British Medical Journal article, we had only 5 percent missing data. It was really remarkable. Prior to this experience, I think people underestimated the importance of getting complete data. One of the reasons we had support for the protocol is because people believed the data. They may not like their results, but at least they didn't challenge us to say it is not good data.
QM: So what's the next step in improving patient safety?
Pronovost: I think the public needs to tell us what is of most concern to them. For that we will need research. I wish we had a forum where we could do focus groups around the country, or the ability to get public opinion data or some broader consumer input. But right now, I think the best we can do is to focus on where the preventable harm is occurring. We know how to measure and we have good models for ventilator-associated pneumonia. There are probably 30,000 people that die needlessly from that annually. Catheter-related bloodstream infections and ventilator-associated pneumonia together account for the vast majority of preventable deaths in the hospital.
QM: And after that?
Pronovost: My sense is that robust programs to prevent obesity—especially childhood obesity—and smoking should be huge priorities. Some may say these are not patient safety programs, but I am a pragmatist. I care more about preventable harm (whatever its causes, whether inadequate education, lack of healthy food choices, or a failure to provide smoking cessation counseling and resources) than a taxonomy of safety. More than 50 percent of our health outcomes—and health care costs—are determined by lifestyle choices. This needs serious investment and focused attention. The approach we used to reduce catheter-related bloodstream infections can be applied to those problems.
Pronovost: It starts with national measures of obesity and smoking. We then need to summarize the evidence about what works, and we need to incentivize innovation in implementing the evidence to reduce smoking and obesity. Where there are knowledge gaps, we need to support research to fill them. We also need to set a measurable common goal, just as John F. Kennedy did when he challenged the country to put a man on the moon and bring him back in 10 years.
QM: Do you have any concerns about how these next problems are approached?
Pronovost: Yes. The fear I have is that upcoming programs will not rely heavily enough on good measure development, an understanding of what interventions work, and learning how to implement those interventions based on local context. Our Michigan program worked because there was an awful lot of underlying science. Other initiatives won't work if the science isn't good. In the U.S. there has been woefully little investment in what I would call the basic science of quality and safety.
QM: How would you solve this problem?
Pronovost: I think there ought to be an agency that develops outcomes measures and makes them public. The state-of-the-art of outcome measures is woefully inadequate; most of them are misguided and they are going to misinform the public. That is not a limit of science but a limit of our investment. We have put no money into measurement development. The National Quality Forum (NQF) validates measures but they have little money to develop new measures. The National Committee for Quality Assurance (NCQA) also has measures, and both organizations have helped improved quality and advance measurement. Yet I think we need an organization similar to the U.S. Securities and Exchange Commission that is solely responsible for measuring outcomes. The private sector sets the rules, has accountability, publicly reports outcomes, and encourages private sector analysis. I think NCQA and/or NQF can be that private sector entity, yet I think it needs to be housed in a federal agency with accountability.
QM: How much money would be required?
Pronovost: I would like to see a rough number of $50 million to $100 million in a pool right off the bat to coordinate the development of these measures. I can see the process of developing measures could be standardized. Obviously the content will vary. I think there is a way to finance it as part of the $19 billion in federal stimulus money committed to the implementation of electronic health records.
QM: What is your opinion of the regulatory agencies, especially at the state level? Are they effective and how does their role need to change?
Pronovost: Regulation is needed in our society because there are some health care organizations that go awry. We need wise regulation to ensure markets compete on truth. Because regulation focuses on finding the bad apple, it is only going to assure the minimum level of quality. Alone it will never produce the high level of quality that the U.S. public wants and deserves. . We need regulation, yet we need more.
QM: What about public reporting?
Pronovost: My fear is that the politics of demanding transparency has exceeded our investment in the science of measurement. We are putting data out there now that isn't very informative. Clinicians are, I believe, justifiably pushing back because they are concerned about the accuracy of the data. For many measures we do not know how accurately they can be calibrated or to what extent they measure what can be prevented. Meanwhile, the insurers and the public are pushing back saying, "We want information." What I don't see is leadership that can bridge this divide by saying, in a sense, "You are both right." The consumer should have information—I think any thoughtful clinician or health care organization agrees with that. But consumers also should have truthful information. How are we going to get accurate information except through an investment in the science? I haven't seen a plan to accomplish this in health reform. I believe an agency, operating in much the same way the U.S. Securities and Exchange Commission does, can make it happen.
 R. M. Wachter, Patient Safety at Ten: Unmistakable Progress, Troubling Gaps, Health Affairs, January 2010 29(1):165–73.
 L.A. Mermel, Prevention of Intravascular Catheter-Related Infections, Annals of Internal Medicine, March 2000 132(5):391–402; N. P. O'Grady, M. Alexander, E. Patchen Dellinger et al., Guidelines for the Prevention of Intravascular Catheter-Related Infections, Morbidity and Mortality Weekly Report, August 2002 51(RR10):1–26.The five evidence-based recommendations include hand washing, use of full barrier precautions, cleaning the skin with chlorhexidine, and avoiding the femoral site when possible. The fifth recommendation is to remove unnecessary catheters.
 P. J. Pronovost, G. A. Goeschel, E. Colantuoni et al., Sustaining Reductions in Catheter Related Bloodstream Infections in Michigan Intensive Care Units: Observational Study, British Medical Journal, February 4, 2010.