The Checklist Manifesto

by Atul Gawande

  "David Lee Roth had a checklist!" I yelled at the radio.

  I ran my theory--about the necessity of checklists--by Jody Adams, the chef and own er of Rialto, one of my favorite restaurants in Boston. In the early 1990s, Food and Wine magazine named her one of America's ten best new chefs, and in 1997 she won a James Beard Foundation Best Chef award, which is the Oscar for food. Rialto is frequently mentioned on national best-restaurant lists, most recently Esquire magazine's. Her focus is on regional Italian cuisine, though with a distinctive take.

  Adams is self-taught. An anthropology major at Brown University, she never went to culinary school. "But I had a thing for food," as she puts it, and she went to work in restaurants, learning her way from chopping onions to creating her own style of cooking.

  The level of skill and craft she has achieved in her restaurant is daunting. Moreover, she has sustained it for many years now. I was interested in how she did it. I understood perfectly well how the Burger Kings and Taco Bells of the world operate. They are driven by tightly prescribed protocol. They provide Taylorized, assembly-line food. But in great restaurants the food is ever-evolving, refined, and individual. Nevertheless, they have to produce an extraordinary level of excellence day after day, year after year, for one to three hundred people per night. I had my theory of how such perfectionism is accomplished, but was it true? Adams invited me in to see.

  I spent one Friday evening perched on a stool in Rialto's long and narrow kitchen amid the bustle, the shouting, the grill flaming on one side, the deep fryer sizzling on another. Adams and her staff served 150 people in five hours. That night, they made a roasted tomato soap with sweated onions and garlic; squid ink ravioli filled with a salt cod brandade on a bed of squash blossoms and lobster sauce; grilled bluefish with corn relish, heirloom tomatoes, and pickled peppers; slow-roasted duck marinated in soy sauce, balsamic vinegar, mustard, rosemary, and garlic; and three dozen other mouthwatering dishes.

  Sitting there, I saw remarkable expertise. Half of Adams's staff had been to culinary school. Few had less than a decade of experience. They each had a kitchen specialty. There was a pastry chef, baker, grill chef, fry cook, dessert chef, sous chef, sommelier--you get the picture. Through the years, they had perfected their technique. I couldn't fathom the subtleties of most of what they did. Though I am a surgeon, they wouldn't let me anywhere near their knives. Jay, the pasta chef, showed me how to heat butter properly and tell by sight when gnocchi were perfectly boiled. Adams showed me how much a pinch of salt really was.

  People celebrate the technique and creativity of cooking. Chefs are personalities today, and their daring culinary exploits are what make the television cooking shows so popular. But as I saw at Rialto, it's discipline--uncelebrated and untelevised--that keeps the kitchen clicking. And sure enough, checklists were at the center of that discipline.

  First there was the recipe--the most basic checklist of all. Every dish had one. The recipes were typed out, put in clear plastic sleeves, and placed at each station. Adams was religious about her staff 's using them. Even for her, she said, "following the recipe is essential to making food of consistent quality over time."

  Tacked to a bulletin board beside the dessert station was what Adams called her Kitchen Notes--e-mails to the staff of her brief observations about the food. The most recent was from 12:50 the previous night. "Fritters--more herbs, more garlic ... more punch," it said. "Corn silk in corn! Creamed corn side on oval plates--not square! Mushrooms--more shallots, garlic, and marsala. USE THE RECIPES!"

  The staff didn't always love following the recipes. You make the creamed corn a few hundred times and you believe you have it down. But that's when things begin to slip, Adams said.

  The recipes themselves were not necessarily static. All the ones I saw had scribbled modifications in the margins--many of them improvements provided by staff. Sometimes there would be a wholesale revamp.

  One new dish they were serving was a split whole lobster in a cognac and fish broth reduction with littleneck clams and chorizo. The dish is Adams's take on a famous Julia Child recipe. Before putting a dish on the menu, however, she always has the kitchen staff make a few test runs, and some problems emerged. Her recipe called for splitting a lobster and then sauteing it in olive oil. But the results proved too variable. Too often the lobster meat was either overcooked or undercooked. The sauce was also made to order, but it took too long for the eight-to-ten-minute turnaround that customers expect.

  So she and two of her chefs reengineered the dish. They decided to make the sauce in advance and parboil the lobster ahead of time, as well. On repeated test runs, the lobster came out perfectly. The recipe was rewritten.

  There was also a checklist for every customer. When an order was placed up front, it was printed out on a slip back in the kitchen. The ticket specified the dishes ordered, the table number, the seat number, any preferences required by the customer or noted in a database from previous visits--food allergies, for instance, or how the steak should be cooked, or whether this was a special occasion like a birthday or a visit from a VIP whom Adams needed to go out and say hello to. The sous chef, who serves as a kind of field officer for operations, read the tickets off as they came in.

  "Fire mushrooms. Fire mozz. Lobo on hold. Steak very well done, no gluten, on hold."

  "Fire" meant cook it now. "On hold" meant it was a second course. "Lobo" was the lobster. The steak needed to be cooked all the way through and the customer had a gluten allergy. A read-back was expected to confirm that the line cooks had heard the order right.

  "Fire mushrooms. Fire mozz," said one.

  "Lobo on hold," said the seafood cook.

  "Steak very well done, no gluten, on hold," said the grill chef.

  As in the construction world, however, not everything could be anticipated and reduced to a recipe. And so Adams, too, had developed a communication checklist to ensure people recognized, and dealt with, unexpected problems as a team. At five o'clock, half an hour before opening, the staff holds what she calls the "pow wow." Everyone gathers in the kitchen for a quick check to discuss unanticipated issues and concerns--the unpredictable. The night I was there, they reviewed the reservation count, two menu changes, how to fill in for a sick staff member, and a sweet sixteen party with twenty girls who were delayed and going to arrive in the midst of the dinner rush. Everyone was given a chance to speak, and they made plans for what to do.

  Of course, this still couldn't guarantee everything would go right. There remained plenty of sources of uncertainty and imperfection: a soup might be plated too early and allowed to cool, a quail might have too little sauce, a striped bass might come off the grill too dry. So Adams had one final check in place. Every plate had to be reviewed by either her or the sous chef before it left the kitchen for the dining room. They made sure the food looked the way it should, checked it against the order ticket, gave it a sniff or, with a clean spoon, maybe even a taste.

  I counted the dishes as they went by. At least 5 percent were sent back. "This calamari has to be fried more," the sous chef told the fry cook. "We want more of a golden brown."

  Later, I got to try some of the results. I had the fried olives, the grilled clams, the summer succotash, and a local farm green salad. I also had the lobster. The food was incredible. I left at midnight with my stomach full and my brain racing. Even here, in one of our most particularized and craft-driven enterprises--in a way, Adams's cooking is more art than science--checklists were required. Everywhere I looked, the evidence seemed to point to the same conclusion. There seemed no field or profession where checklists might not help. And that might even include my own.

  5. THE FIRST TRY

  In late 2006, a woman with a British accent and a Geneva telephone number called me. She said that she was from the World Health Organization and she wanted to see whether I might help them organize a group of people to solve a small problem. Officials were picking up indications that the volume of surgery was increasing world
wide and that a significant portion of the care was so unsafe as to be a public danger. So they wanted to develop a global program to reduce avoidable deaths and harm from surgery.

  I believe my response was, "Um, how do you do that?"

  "We'll have a meeting," she said.

  I asked how much money they'd be devoting to the problem.

  "Oh, there's no real money," she said.

  I said no. I couldn't do it. I was busy.

  But she knew what she was about. She said something along the lines of, "Oh, sorry. I thought you were supposed to be some kind of expert on patient safety in surgery. My mistake."

  I agreed to help organize the meeting.

  One of the benefits of joining up to work with WHO was gaining access to the health system reports and data from the organization's 193 member countries. And compiling the available numbers in surgery, my research team and I found that the WHO officials' impression was correct: the global volume of surgery had exploded. By 2004, surgeons were performing some 230 million major operations annually--one for every twenty-five human beings on the planet--and the numbers have likely continued to increase since then. The volume of surgery had grown so swiftly that, without anyone's quite realizing, it has come to exceed global totals for childbirth--only with a death rate ten to one hundred times higher. Although most of the time a given procedure goes just fine, often it doesn't: estimates of complication rates for hospital surgery range from 3 to 17 percent. While incisions have gotten smaller and recoveries have gotten quicker, the risks remain serious. Worldwide, at least seven million people a year are left disabled and at least one million dead--a level of harm that approaches that of malaria, tuberculosis, and other traditional public health concerns.

  Peering at the numbers, I understood why WHO--an organization devoted to solving large-scale public health problems--should suddenly have taken an interest in something as seemingly specific and high-tech as surgical care. Improvement in global economic conditions in recent decades had produced greater longevity and therefore a greater need for essential surgical services--for people with cancers, broken bones and other traumatic injuries, complications during child delivery, major birth defects, disabling kidney stones and gallstones and hernias. Although there remained some two billion people, especially in rural areas, without access to a surgeon, health systems in all countries were now massively increasing the number of surgical procedures performed. As a result, the safety and quality of that care had become a major issue everywhere.

  But what could be done about it? Remedying surgery as a public health matter is not like remedying, say, polio. I'd traveled with WHO physicians overseeing the campaign to eradicate polio globally and seen how hard just providing vaccines to a population could be. Surgery was drastically more complex. Finding ways to reduce its harm in a single hospital seemed difficult enough. Finding a way that could reach every operating room in the world seemed absurd. With more than twenty-five hundred different surgical procedures, ranging from brain biopsies to toe amputations, pacemaker insertions to spleen extractions, appendectomies to kidney transplants, you don't even know where to start. Perhaps, I thought, I could work with WHO to focus on reducing the harm of just one procedure--much like with central lines--but how much of a dent would that make in a problem of this scale?

  In January 2007, at WHO headquarters in Geneva, we convened a two-day meeting of surgeons, anesthesiologists, nurses, safety experts, even patients from around the world to puzzle through what could be done. We had clinicians from top facilities in Europe, Canada, and the United States. We had the chief surgeon for the International Committee of the Red Cross, who had sent teams to treat sick and wounded refugees everywhere from Mogadishu to Indonesia. We had a father from Zambia whose daughter inadvertently suffocated from lack of oxygen during treatment. As the group told stories of their findings and experiences with surgery around the world, I became only more skeptical. How could we possibly attempt to address so many different issues in so many different places?

  A medical officer in his forties from western Ghana, where cocoa growing and gold mining had brought a measure of prosperity, told of the conditions in his district hospital. No surgeon was willing to stay, he said. Ghana was suffering from a brain drain, losing many of its highest skilled citizens to better opportunities abroad. He told us his entire hospital had just three medical officers--general physicians with no surgical training. Nevertheless, when a patient arrives critically ill and bleeding after two days in labor, or sick and feverish from appendicitis, or with a collapsed lung after a motorbike crash, the untutored doctors do what they have to do. They operate.

  "You must understand," he said. "I manage everything. I am the pediatrician, obstetrician, surgeon, everything." He had textbooks and a manual of basic surgical techniques. He had an untrained assistant who had learned how to give basic anesthesia. His hospital's equipment was rudimentary. The standards were poor. Things sometimes went wrong. But he was convinced doing something was better than doing nothing at all.

  A Russian bioengineer spoke. He'd spent much of his career overseeing the supply and ser vice of medical equipment to hospitals in different parts of the world, and he described dangerous problems in both high-and low-income settings: inadequately maintained surgical devices that have set fire to patients or electrocuted them; new technologies used incorrectly because teams had not received proper training; critical, lifesaving equipment that was locked away in a cabinet or missing when people needed it.

  The chief of surgery for the largest hospital in Mongolia described shortages of pain control medications, and others from Asia, Africa, and the Middle East recounted the same. A New Zealand researcher spoke of terrifying death rates in poor countries from unsafe anesthesia, noting that although some places in Africa had fewer than one in five thousand patients die from general anesthesia, others had rates more than ten times worse, with one study in Togo showing one in 150 died. An anesthesiologist from India chimed in, tracing problems with anesthesia to the low respect most surgeons accord anesthetists. In her country, she said, they shout anesthetists down and disregard the safety issues that her colleagues raise. Medical students see this and decide not to go into anesthesiology. As a result, the most risky part of surgery--anesthesia--is done by untrained people far more often than the surgery itself. A nurse from Ireland joined the clamor. Nurses work under even worse conditions, she said. They are often ignored as members of the team, condescended to, or fired for raising concerns. She'd seen it in her home country, and from her colleagues abroad she knew it to be the experience of nurses internationally.

  On the one hand, everyone firmly agreed: surgery is enormously valuable to people's lives everywhere and should be made more broadly available. Even under the grimmest conditions, it is frequently lifesaving. And in much of the world, the serious complication rates seem acceptably low--in the 5 to 15 percent range for hospital operations.

  On the other hand, the idea that such rates are "acceptable" was hard to swallow. Each percentage point, after all, represented millions left disabled or dead. Studies in the United States alone had found that at least half of surgical complications were preventable. But the causes and contributors were of every possible variety. We needed to do something. What, though, wasn't clear.

  Some suggested more training programs. The idea withered almost upon utterance. If these failures were problems in every country--indeed, very likely, in every hospital--no training program could be deployed widely enough to make a difference. There was neither the money nor the capacity.

  We discussed incentive approaches, such as the pay-for-performance schemes recently initiated on a trial basis in the United States. In these programs, clinicians receive financial rewards for being more consistent about giving, say, heart attack patients the proper care or incur penalties for not doing so. The strategy has shown results, but the gains have been modest--the country's largest pay-for-performance trial, for example, registered just 2 to 4 perce
nt improvement. Furthermore, the measurements required for incentive payments are not easy to obtain. They rely on clinicians' self-reported results, which are not always accurate. The results are also strongly affected by how sick patients are to begin with. One might be tempted, for example, to pay surgeons with higher complication rates less, but some might simply have sicker patients. The incentive programs have thus far been expensive, incremental, and of limited benefit. Taking them global was unimaginable.

  The most straightforward thing for the group to do would have been to formulate and publish under the WHO name a set of official standards for safe surgical care. It is the approach expert panels commonly take. Such guidelines could cover everything from measures to prevent infection in surgery to expectations for training and cooperation in operating rooms. This would be our Geneva Convention on Safe Surgery, our Helsinki Accord to Stop Operating Room Mayhem.

  But one had only to take a walk through the dim concrete basement hallways of the otherwise soaring WHO headquarters to start doubting that plan. Down in the basement, while taking a shortcut between buildings, I saw pallet after pallet of two-hundred-page guideline books from other groups that had been summoned to make their expert pronouncements. There were guidelines stacked waist-high on malaria prevention, HIV/AIDS treatment, and influenza management, all shrink-wrapped against the gathering dust. The standards had been carefully written and were, I am sure, wise and well considered. Some undoubtedly raised the bar of possibility for achievable global standards. But in most cases, they had at best trickled out into the world. At the bedsides of patients in Bangkok and Brazzaville, Boston and Brisbane, little had changed.

  I asked a WHO official whether the organization had a guidebook on how to carry out successful global public health programs. She regarded me with a look that a parent might give a toddler searching the dog's mouth for the thing that makes the barking noise. It's a cute idea but idiotic.