The Checklist Manifesto

by Atul Gawande

  There was nothing particularly scientific or even consistent about the decision-making process. Operating on the wrong patient or the wrong side of the body is exceedingly rare too. But the checks to prevent such errors are relatively quick and already accepted in several countries, including the United States. Such mistakes also get a lot of attention. So those checks stayed in.

  In contrast, our checks to prevent communication breakdowns tackled a broad and widely recognized source of failure. But our approach--having people formally introduce themselves and briefly discuss critical aspects of a given case--was far from proven effective. Improving teamwork was so fundamental to making a difference, however, that we were willing to leave these measures in and give them a try.

  After our London meeting, we did more small-scale testing--just one case at a time. We had a team in London try the draft checklist and give us suggestions, then a team in Hong Kong. With each successive round, the checklist got better. After a certain point, it seemed we had done all we could. We had a checklist we were ready to circulate.

  The final WHO safe surgery checklist spelled out nineteen checks in all. Before anesthesia, there are seven checks. The team members confirm that the patient (or the patient's proxy) has personally verified his or her identity and also given consent for the procedure. They make sure that the surgical site is marked and that the pulse oximeter--which monitors oxygen levels--is on the patient and working. They check the patient's medication allergies. They review the risk of airway problems--the most dangerous aspect of general anesthesia--and that appropriate equipment and assistance for them are available. And lastly, if there is a possibility of losing more than half a liter of blood (or the equivalent for a child), they verify that necessary intravenous lines, blood, and fluids are ready.

  After anesthesia, but before incision, come seven more checks. The team members make sure they've been introduced by name and role. They confirm that everyone has the correct patient and procedure (including which side of the body--left versus right) in mind. They confirm that antibiotics were either given on time or were unnecessary. They check that any radiology images needed for the operation are displayed. And to make sure everyone is briefed as a team, they discuss the critical aspects of the case: the surgeon reviews how long the operation will take, the amount of blood loss the team should prepare for, and anything else people should be aware of; the anesthesia staff review their anesthetic plans and concerns; and the nursing staff review equipment availability, sterility, and their patient concerns.

  Finally, at the end of the operation, before the team wheels the patient from the room, come five final checks. The circulating nurse verbally reviews the recorded name of the completed procedure for accuracy, the labeling of any tissue specimens going to the pathologist, whether all needles, sponges, and instruments have been accounted for, and whether any equipment problems need to be addressed before the next case. Everyone on the team also reviews aloud their plans and concerns for the patient's recovery after surgery, to ensure information is complete and clearly transmitted.

  Operations require many more than nineteen steps, of course. But like builders, we tried to encompass the simple to the complex, with several narrowly specified checks to ensure stupid stuff isn't missed (antibiotics, allergies, the wrong patient) and a few communication checks to ensure people work as a team to recognize the many other potential traps and subtleties. At least that was the idea. But would it work and actually make a measurable difference in reducing harm to patients? That was the question.

  To find the answer, we decided to study the effect of the safe surgery checklist on patient care in eight hospitals around the world. This number was large enough to provide meaningful results while remaining manageable for my small research team and the modest bud get WHO agreed to furnish. We got dozens of applications from hospitals seeking to participate. We set a few criteria for selection. The hospital's leader had to speak English--we could translate the checklist for staff members but we didn't have the resources for daily communication with eight site leaders in multiple languages. The location had to be safe for travel. We received, for instance, an enthusiastic application from the chief of surgery in an Iraqi hospital, which would have been fascinating, but conducting a research trial in a war zone seemed unwise.

  I also wanted a wide diversity of participating hospitals--hospitals in rich countries, poor countries, and in between. This insistence caused a degree of consternation at WHO headquarters. As officials explained, WHO's first priority is, quite legitimately, to help the poorer parts of the world, and the substantial costs of paying for data collection in wealthier countries would divert resources from elsewhere. But I had seen surgery in places ranging from rural India to Harvard and seen failure across the span. I thought the checklist might make a difference anywhere. And if it worked in high-income countries, that success might help persuade poorer facilities to take it up. So we agreed to include wealthier sites if they agreed to support most, if not all, the research costs themselves.

  Lastly, the hospitals had to be willing to allow observers to measure their actual rates of complications, deaths, and systems failures in surgical care before and after adopting the checklist. Granting this permission was no small matter for hospitals. Most--even those in the highest income settings--have no idea of their current rates. Close observation was bound to embarrass some. Nonetheless, we got eight willing hospitals lined up from all over the globe.

  Four were in high-income countries and among the leading hospitals in the world: the University of Washington Medical Center in Seattle, Toronto General Hospital in Canada, St. Mary's Hospital in London, and Auckland City Hospital, New Zealand's largest. Four were intensely busy hospitals in low-or middle-income countries: Philippines General Hospital in Manila, which was twice the size of the wealthier hospitals we enrolled; Prince Hamza Hospital in Amman, Jordan, a new government facility built to accommodate Jordan's bursting refugee population; St. Stephen's Hospital in New Delhi, an urban charity hospital; and St. Francis Designated District Hospital in Ifakara, Tanzania, the lone hospital serving a rural population of nearly one million people.

  This was an almost ridiculous range of hospitals to study. Annual health care spending in the high-income countries reached thousands of dollars per person, while in India, the Philippines, and East Africa, it did not rise beyond the double digits. So, for example, the bud get of the University of Washington Medical Center--over one billion dollars per year--was more than twice that of the entire country of Tanzania. Surgery therefore differed starkly in our eight hospitals. On one end of the spectrum were those with state-of-the-art capabilities allowing them to do everything from robotic prostatectomies to liver transplants, along with factory loads of planned, low-risk, often day-surgery procedures like hernia repairs, breast biopsies, and ear-tube placements for drainage of chronic ear infections in children. On the other end were hospitals forced by lack of staff and resources to prioritize urgent operations--emergency cesarean sections for mothers dying in childbirth, for example, or procedures for repair of severe traumatic injuries. Even when the hospitals did the same operations--an appendectomy, a mastectomy, the placement of a rod in a broken femur--the conditions were so disparate that the procedures were the same only in name. In the poorer hospitals, the equipment was meager, the teams' training was more limited, and the patients usually arrived sicker--the appendix having ruptured, the breast cancer having grown twice as large, the femur proving not only broken but infected.

  Nonetheless, we went ahead with our eight institutions. The goal, after all, was not to compare one hospital with another but to determine where, if anywhere, the checklist could improve care. We hired local research coordinators for the hospitals and trained them to collect accurate information on deaths and complications. We were conservative about what counted. The complications had to be significant--a pneumonia, a heart attack, bleeding requiring a return to the operating room or more than four units of blood, a
documented wound infection, or the like. And the occurrence had to actually be witnessed in the hospital, not reported from elsewhere.

  We collected data on the surgical care in up to four operating rooms at each facility for about three months before the checklist went into effect. It was a kind of biopsy of the care received by patients across the range of hospitals in the world. And the results were sobering.

  Of the close to four thousand adult surgical patients we followed, more than four hundred developed major complications resulting from surgery. Fifty-six of them died. About half the complications involved infections. Another quarter involved technical failures that required a return trip to the operating room to stop bleeding or repair a problem. The overall complication rates ranged from 6 percent to 21 percent. It's important to note that the operating rooms we were studying tended to handle the hospital's more complex procedures. More straightforward operations have lower injury rates. Nonetheless, the pattern confirmed what we'd understood: surgery is risky and dangerous wherever it is done.

  We also found, as we suspected we would, signs of substantial opportunity for improvement everywhere. None of the hospitals, for example, had a routine approach to ensure that teams had identified, and prepared for, cases with high blood-loss risk, or conducted any kind of preoperative team briefing about patients. We tracked performance of six specific safety steps: the timely delivery of antibiotics, the use of a working pulse oximeter, the completion of a formal risk assessment for placing an airway tube, the verbal confirmation of the patient's identity and procedure, the appropriate placement of intravenous lines for patients who develop severe bleeding, and finally a complete accounting of sponges at the end of the procedure. These are basics, the surgical equivalent of unlocking the elevator controls before airplane takeoff. Nevertheless, we found gaps everywhere. The very best missed at least one of these minimum steps 6 percent of the time--once in every sixteen patients. And on average, the hospitals missed one of them in a startling two-thirds of patients, whether in rich countries or poor. That is how flawed and inconsistent surgical care routinely is around the world.

  Then, starting in spring 2008, the pilot hospitals began implementing our two-minute, nineteen-step surgery checklist. We knew better than to think that just dumping a pile of copies in their operating rooms was going to change anything. The hospital leaders committed to introducing the concept systematically. They made presentations not only to their surgeons but also to their anesthetists, nurses, and other surgical personnel. We supplied the hospitals with their failure data so the staff could see what they were trying to address. We gave them some PowerPoint slides and a couple of YouTube videos, one demonstrating "How to Use the Safe Surgery Checklist" and one--a bit more entertaining--entitled "How Not to Use the Safe Surgery Checklist," showing how easy it is to screw everything up.

  We also asked the hospital leaders to introduce the checklist in just one operating room at first, ideally in procedures the chief surgeon was doing himself, with senior anesthesia and nursing staff taking part. There would surely be bugs to work out. Each hospital would have to adjust the order and wording of the checklist to suit its particular practices and terminology. Several were using translations. A few had already indicated they wanted to add extra checks. For some hospitals, the checklist would also compel systemic changes--for example, stocking more antibiotic supplies in the operating rooms. We needed the first groups using the checklist to have the seniority and patience to make the necessary modifications and not dismiss the whole enterprise.

  Using the checklist involved a major cultural change, as well--a shift in authority, responsibility, and expectations about care--and the hospitals needed to recognize that. We gambled that their staff would be far more likely to adopt the checklist if they saw their leadership accepting it from the outset.

  My team and I hit the road, fanning out to visit the pilot sites as the checklist effort got under way. I had never seen surgery performed in so many different kinds of settings. The contrasts were even starker than I had anticipated and the range of problems was infinitely wider.

  In Tanzania, the hospital was two hundred miles of sometimes one-lane dirt roads from the capital, Dar es Salaam, and flooding during the rainy season cut off supplies--such as medications and anesthetic gases--often for weeks at a time. There were thousands of surgery patients, but just five surgeons and four anesthesia staff. None of the anesthetists had a medical degree. The patients' families supplied most of the blood for the blood bank, and when that wasn't enough, staff members rolled up their sleeves. They conserved anesthetic supplies by administering mainly spinal anesthesia--injections of numbing medication directly into the spinal canal. They could do operations under spinal that I never conceived of. They saved and resterilized their surgical gloves, using them over and over until holes appeared. They even made their own surgical gauze, the nurses and anesthesia staff sitting around an old wood table at teatime each afternoon cutting bolts of white cotton cloth to size for the next day's cases.

  In Delhi, the charity hospital was not as badly off as the Tanzanian site or hospitals I'd been to in rural India. There were more supplies. The staff members were better trained. But the volume of patients they were asked to care for in this city of thirteen million was beyond comprehension. The hospital had seven fully trained anesthetists, for instance, but they had to perform twenty thousand operations a year. To provide a sense of how ludicrous this is, our New Zealand pilot hospital employed ninety-two anesthetists to manage a similar magnitude of surgery. Yet, for all the equipment shortages, power outages, waiting lists, fourteen-hour days, I heard less unhappiness and complaining from the surgical staff in Delhi than in many American hospitals I've been to.

  The differences were not just between rich and poor settings, either. Each site was distinctive. St. Mary's Hospital, for example, our London site, was a compound of red brick and white stone buildings more than century and a half old, sprawling over a city block in Paddington. Alexander Fleming discovered penicillin here in 1928. More recently, under its chairman of surgery, Lord Darzi of Denham, the hospital has become an international pioneer in the development of minimally invasive surgery and surgical simulation. St. Mary's is modern, well equipped, and a draw for London's powerful and well-to-do--Prince William and Prince Harry were born here, for example, and Conservative Party leader David Cameron's severely disabled son was cared for here, as well. But it is hardly posh. It remains a government hospital in the National Health Ser vice, serving any Briton without charge or distinction.

  Walking through St. Mary's sixteen operating rooms, I found they looked much the same as the ones where I work in Boston--high-tech, up-to-date. But surgical procedures seemed different at every stage. The patients were put to sleep outside the operating theater, instead of inside, and then wheeled in, which meant that the first part of the checklist would have to be changed. The anesthetists and circulating nurses didn't wear masks, which seemed like sacrilege to me, although I had to admit their necessity is unproven for staff who do not work near the patient's incision. Almost every term the surgical teams used was unfamiliar. We all supposedly spoke English, but I was often unsure what they were talking about.

  In Jordan, the working environment was also at once recognizable and alien, but in a different way. The operating rooms in Amman had zero frills--this was a still-developing country and the equipment was older and heavily used--but they had most of the things I am used to as a surgeon, and the level of care seemed very good. One of the surgeons I met was Iraqi. He'd trained in Baghdad and practiced there until the chaos following the American invasion in 2003 forced him to flee with his family, abandoning their home, their savings, and his work. Before Saddam Hussein, in the last years of his rule, gutted the Iraqi medical system, Baghdad had provided some of the best medical care in the Middle East. But, the surgeon said, Jordan now seemed positioned to take that role and he felt fortunate to be there. I learned that more than 200,000 foreign pati
ents travel to Jordan for their health care each year, generating as much as one billion dollars in revenues for the country.

  What I couldn't work out, though, was how the country's strict gender divide was negotiated in its operating rooms. I remember sitting outside a restaurant the day I arrived, studying the people passing by. Men and women were virtually always separated. Most women covered their hair. I got to know one of the surgery residents, a young man in his late twenties who was my guide for the visit. We even went out to see a movie together. When I learned he had a girlfriend of two years, a graduate student, I asked him how long it was before he got to see her hair.

  "I never have," he said.

  "C'mon. Never?"

  "Never." He'd seen a few strands. He knew she had dark brown hair. But even in the more modern dating relationship of a partly Westernized, highly educated couple, that was it.

  In the operating rooms, all the surgeons were men. Most of the nurses were women. The anesthetists split half and half. Given the hierarchies, I wondered whether the kind of team-work envisioned by the checklist was even possible. The women wore their head scarves in the operating rooms. Most avoided eye contact with men. I slowly learned, however, that not all was what it seemed. The staff didn't hesitate to discard the formalities when necessary. I saw a gallbladder operation in which the surgeon inadvertently contaminated his glove while adjusting the operating lights. He hadn't noticed. But the nurse had.

  "You have to change your glove," the nurse told him in Arabic. (Someone translated for me.)