Sorry, you need to enable JavaScript to visit this website.
Skip to main content

In Conversation With… Paul H. O'Neill, MPA

January 1, 2017 

Editor's note: Mr. O'Neill served as the United States Secretary of the Treasury under President George W. Bush and, prior to that, chairman and CEO of Alcoa. During his tenure at Alcoa, he led the company to become the safest workplace in the world. He also has a strong interest in patient safety. We spoke with him about workplace safety and its relationship to patient safety and organizational excellence.

Dr. Robert M. Wachter: One of the defining elements of your career has been a passion for the safety of the workforce. Where did that come from?

Mr. Paul H. O'Neill: It's a long and complicated story, but the essence of it is that I've been a student for all of my adult life of how organizations work or don't. One of the things that really sharpened my belief about it came when I was first the vice president, and ultimately the president, of International Paper Company. I set out to understand why people who appear to have the same resources—including their human, material, equipment, and technology—can have such markedly different results. I concluded that truly great organizations worked every day to get better at everything that they did—not just a few things.

One thing they worked really hard at was engaging and helping the people in their organization to contribute everything they were capable of contributing. I realized that, in light of the 20% discretionary energy that human beings have, they only give if they have a commitment to the goals of an organization. From all of that, I developed a logic train that says organizations cannot be habitually excellent at everything they do unless they have leadership that engages the people around aspirational goals—that is goals that cannot be legitimately denied—beginning with a goal that cannot be disputed: people should never be hurt at work. I set out to understand that better by examining organizations that were substantially better at workplace safety than others, and I discovered early on that one of the companies that excelled at workplace safety was DuPont.

I went and studied DuPont and found that, after several decades of experience of making gunpowder, which was their original business, they had decided that it was better for people not to have explosions or be hurt at work than to have to be continually rebuilding the place because somebody didn't follow the protocol on how to treat the dangerous gunpowder they were manufacturing. DuPont set out to get the organization to understand that, in spite of all the so-called occupational hazards, it was possible with the right attention to detail and the right engagement of the workforce, even in a so-called dangerous industry, people didn't need to be hurt at work. It went along with learning and observing earlier in my career that places that concentrated on financial things were generally not very good at other things. If the leadership and the people in the organization concentrated on what produced excellent results, the financial results will take care of themselves.

The day I arrived at Alcoa, even though they were already in the top 20% or so of avoiding injuries to their workforce, I announced we should be an injury-free workplace. And over the 13 years I was there we got close to achieving zero injuries. Along the way, we got really good at everything we did. In fact, I adopted a motto that said our goal was to be the best in the world of everything we did, beginning with workplace safety. And it was possible with that kind of an idea to engage everyone in the pursuit of excellence if the leader took away all the excuses.

In all the organizations I've ever worked in, the people had the same excuses: "But we cannot afford it." Or in the case of workplace safety: "Accidents will happen." It caused me to realize we needed to change the language. We shouldn't have any accidents anymore; we should have incidents. Because that gives our brain a license to figure out what contributed to the incident and take actions so that the circumstances are never presented again. It's not something that produces immediate results, but after people get the rhythm of learning—ideally in real time, by identifying what went wrong then making resources available to understand the root cause of what contributed to this incident, then changing the contributing factors—you can actually move forward every day.

Another thing that was really important in the case of Alcoa was we had 143,000 people in 43 countries, so if we could learn something about the contributing causes of an accident in Davenport, Iowa, if we had similar circumstances in plants in Russia, China, Hungary, or any place in the world, as long as we were working in real time, with the advent of computer technology we were able to translate improvement policies around the world within a week. The goal was identify, analyze, share, and implement within 4 days of an incident occurring anywhere in the world, to make sure it never happened again any place in the world.

It created enormous leverage for us. People would ask if we focused on money. And I would say no. But here's the money story: The so-called market value of Alcoa was $4 billion when I started and when I left it was $36 billion. We went from a company that was good at workplace safety to a company that was almost without peer in worker safety and had a 9-fold improvement in the market value of the company at the same time.

RW: Once you started getting interested in health care, what were your thoughts about patient safety as viewed through this lens of workforce safety?

PO: I have to go back in my career a little bit. I have a long-standing engagement with health and medical care going back to the 1960s. Particularly at Alcoa, I was involved and interested in what was going on in the health and medical care community and began to look at the data. Ken Shine, who was a leader in writing the 1999 To Err Is Human report, was a board member with me at the Rand Corporation where I'd been involved from the early 1960s. It was interesting for me to see that our health and medical care delivery system in Pittsburgh looked like the national average, which meant on average 2% of the people in acute care hospitals with a central line had a central line infection, and that was considered normal and inevitable.

I started talking to people in the health and medical care community about things they could learn from what we were doing in workplace safety, namely to identify as quickly as possible anything gone wrong—including medication errors, pressure sores, and all the other things that afflict patients—and apply the tools of identification, real-time learning, and systematic widespread sharing without regard to hierarchy, so that organizations could learn and people could follow an aspirational goal of zero harm to patients associated with zero harm to workers.

RW: I'm sure you got the same pushback: impossible, not enough time, too expensive?

PO: When I was at Alcoa, I created the Alcoa University and offered to teach people in the community these ideas. An early student at the university was Dr. Rick Shannon, who was then at the Allegheny General Hospital in charge of three intensive care units (ICU). He decided these ideas were right. When he looked at his data, he found that in the baseline study year they had 1789 patients. Of those 1789 ICU patients, there were 39 infections and 19 people died. So Rick engaged all of his people, literally everybody—the people who cleaned the rooms, clinicians, nurses, technicians, food service people—in first setting a goal of eliminating these infections. To do it, we must have more agreement on how we do things. They had so much variation on how they did things; it was hard to figure out the probable cause of an infection. The nurses themselves discovered that, because they had all gone to different nursing schools and thought they had learned the correct way to do everything, there were unbelievable variations in how to prep a site for a central line and how to tend to a central line and all the rest. After some deliberation, with Rick's engagement but not his domination, they agreed that they would adopt a standard practice of how to prep and tend lines, and after 18 months or so, we had another baseline period. There was one infection among 1800 patients, and no one died.

RW: Wow.

PO: Which was a demonstration in a real place, in real time, with real people that with the right kind of thought process and the right kind of shared learning and engagement, it's possible to do things in health care that to this day we have not fully adopted across the whole of health and medical care—which is one of the things that gets me up every morning.

RW: Did you find any fundamental differences in health care that shaped your thinking?

PO: Again, several things. I need to be careful here not to overgeneralize. Health care as an industry tends to be quite hierarchical in the sense that people with the most education and the biggest titles are deemed to be always right even if they're not.

RW: More so than in business?

PO: I think there are more businesses that have learned the ideas of continuous learning and continuous improvement. As a consequence, the better ones have become less hierarchical. For example, if you go to a Toyota plant, you see all the people are truly engaged and they're all expected to contribute to everyday improvement in the smallest details of everything. In an awful lot of health care, it's like an isolation booth—you find a variety of practices even in the same environment. The University of Pittsburgh Medical Center has about 22 hospitals. They're markedly different, and the people are not interchangeable. That is what's different about health care. We wrote a report about what we did at Allegheny General. I thought people will now see this, they will look at their own place, and they will adopt these ideas because they obviously work. It didn't happen. You know, it's a curious thing that I found really strange.

In the same way, I demonstrated things in my own work environments when I was Secretary of the Treasury that I had first learned and demonstrated at Alcoa—that it's possible in a great organization for the financial department to close the books, fully audited, every month in 2 and a half days. You don't need a white coat to know whether that's good or not. That is a really good thing. There are very, very few health care organizations that can close their books in 6 months. It's just astounding to me that they don't seem to understand the idea of wall-to-wall excellence in everything that they do.

For me, workplace safety is a leading indicator of whether an organization is on the trail to habitual excellence or not. A place that I've worked with that actually is pretty far down this trail in most aspects is Cincinnati Children's Hospital. They were not quick to pick up on the idea of workplace safety, but they did about 8 or 9 years ago now. And the results are significantly better than other health care institutions around the country.

I was recently asked to come in by an outside board member to a very famous, very large medical care institution. Their OSHA recordable number says 7 people out of every 100 of their staff have an injury every year. In a great organization, that number is 0.3. And so next to their fantastic rankings by US News & World Report is the telltale data that tells you this workplace is not habitually excellent. I believe organizations that are not good at assuring the everyday safety of their workforce are probably no better than average at all the other things they do, like central line infections and pressure sores.

RW: Because the approaches and the culture are essentially the same.

PO: Exactly. When I first started talking about the right aspirational goal for all of these things we measure is zero, people said, "Well you don't understand, we are not making Toyotas—we cannot do it in health care." It's true that every human being is different, every diagnosis needs to be individualized, and treatments need to be crafted to the individual requirements. But it's not true that you cannot systematically eliminate things gone wrong if you're constantly improving what you do by learning from things gone wrong. It's amazing if you look at the number of medication errors in the country—starting with the wrong meds, but traveling through a miscommunication between the order giver and the pharmacy and the prep process and the pharmacy and the delivery process to the ward and the mistakes that are made in the administration and all the rest of that. The last time I looked at the US data, if you take all the errors including unreadable orders, incorrect orders, misplaced decimal points, and all the rest of that, the number is something like 300 million a year. Which is one medication error for each of us. That is just unforgivable.

RW: I think there were people who felt that a focus on workplace safety was misplaced, in that it shouldn't be about us. It's about the patient, and such a focus might be a distraction from patient safety. Obviously you feel that the habitual practices that get both accomplished are very much harmonized. How do you address that concern?

PO: One thing that helped drive me on this issue in the forest product and in the aluminum industry was to observe that when people are hurt, they cannot work. One thing I've learned in health and medical care is they have this wonderful dodge; if you hurt your back lifting a 300-pound patient because you didn't use the mechanical lift or there wasn't one available, they put you at the nurses' desk for a few days so your back can recover. So you're no longer taking care of four patients; you're sitting on your butt waiting for your back to get better.

Another thing in health care: the reporting is terrible. When I first asked my friends at the University of Pittsburgh Medical Center what their lost workday rate was and their OSHA recordable rate, they couldn't tell me. After about 3 weeks, they gave me some data, but I didn't believe it. Because in a really great system, people know up and down the organization this is not a report to the leadership. There is a real-time safety data system that people use as a screen saver, and you can see if your friend on ward C had an injury yesterday. Because it's there for everybody to see; it doesn't have to go through a hierarchy. One thing you have to do is beat back the legal staff who don't want you to tell anybody that there was an injury.

I'll tell you another data point from a famous place that I worked with. In a 2-month period, they had 68 cases of the caregivers getting splashed by blood and body fluids. They found that in 98% of the cases, the people were not wearing their protective equipment. In my factories, we had a rule about safety equipment. For example, especially at Alcoa, there was a danger of people dropping things on their feet, so we said, "Do not come to work without safety shoes." Then we said, "By the way we are going to buy them. We are going to buy you steel-toed shoes, and you must not come on the worksite without steel-toed shoes. You no longer have an excuse. We are going to protect you." If we were really at the top of our game in health care, we would say to people "Do not put yourself at risk when you're dealing with the situation that could produce a blood or a body fluid splash. Even if it's a remote possibility, get yourself geared up. And it's not okay for you to get splashed ever again." Either your people are important or they're not.

RW: Clearly if people get injured on the job, they cannot do their work effectively. Dealing with that argument you hear periodically—which is it's not about us, it's really all about the patients—did you feel like you had to have other responses to that?

PO: I think if it's all about the patient, then you need to be in good order. Many times when people are injured, there is collateral damage. If you're trying to lift someone out of bed without assistance and they're heavy and you hurt your back, and they fall and break a leg then they're immobilized and they get pneumonia—that's collateral damage, especially if they die. So these things go together. It's not correct to say that those of us in the caregiving business should be prepared to give our life in the name of taking care of our patients. That's just idiocy.

RW: From what you have seen in health care for the last decade or so, are you seeing progress, and how do you see the future playing out?

PO: I wish I could say we are making great progress. But I cannot honestly say that. I'm a member at the Deming Center at Columbia University, and we give annual prizes to people. We gave one prize to a very famous person in health and medical care. When we were sitting at dinner before the prize was awarded, I asked him, "Do you have any idea what the adherence to the hand-sanitizing policy is in your place?" And he replied, "I do. The answer is 78%." And so I said to him, "That means 22% of the time people are at risk." Is that okay? I don't think so. Moreover, when I see a place that says that their adherence to the hand sanitizing policy was 95%, I say that I don't believe your observations.

RW: Well that's our number and I've had the same response, but I've come to believe that it's real.

PO: That's great. There ought to be confirming data with the workplace safety data. I haven't looked at it today, but last week the year-to-date lost workday safety rate at Alcoa was 0.07.

RW: Hmm, it's impressive that you still check it.

PO: I look at it every day.

RW: Anything else that you'd like to mention?

PO: It would be such a blessing if we could raise our game in both worker safety and patient safety in this country, because it's not about spending more money. It's about creating learning systems where people are enabled to make improvements every day in how the work is done for their own benefit and for the benefit of patients. If we could up our game that would be fantastic.

This project was funded under contract number 75Q80119C00004 from the Agency for Healthcare Research and Quality (AHRQ), U.S. Department of Health and Human Services. The authors are solely responsible for this report’s contents, findings, and conclusions, which do not necessarily represent the views of AHRQ. Readers should not interpret any statement in this report as an official position of AHRQ or of the U.S. Department of Health and Human Services. None of the authors has any affiliation or financial involvement that conflicts with the material presented in this report. View AHRQ Disclaimers
Related Resources