From bottom to top, a culture of safety
It was a full decade ago that Katherine Galagan, MD, visited a Toyota Motor Corp. manufacturing plant in Japan. But her memories of it are so vivid that they could have happened yesterday: “I remember standing up on this catwalk, watching doors flying around overhead with the cars moving on a conveyor belt down below, and these carts going in and out for each worker,” she says. “It was like watching a ballet.”
If Dr. Galagan, director of clinical laboratories at Virginia Mason Medical Center, Seattle, sounds just a little star-struck, she’s not the only one. Cathie Furman, RN, Virginia Mason’s senior vice president of quality and compliance, was on that trip also, and she too came away from it utterly wowed by the Toyota manufacturing process. The Toyota production process she witnessed was not only speedy and efficient but also safe, thanks to a philosophy that encouraged any worker at any time to “stop the line,” that is, halt production until a problem was corrected.
“We were so impressed with the [Toyota] culture—the empowerment of high-school-trained assembly-line workers who felt completely comfortable stopping a multi-million-dollar line rather than sending a defective product to their teammate,” Furman says. “That was so different than what we experienced in health care, which has historically been a very blaming, hierarchical culture.”
Along with many other Virginia Mason leaders, Furman and Dr. Galagan had made the trip to Japan to study the famously innovative and effective Toyota Production System, and to determine how it could be adapted to a hospital setting. “We recognized very clearly that the Toyota workers felt completely accountable for fixing mistakes in the moment rather than just pushing them on,” says Furman. “And we said, ‘If they can do that with cars, we should be able to do that with our patients.’”
So they did. Ten years later, the Virginia Mason Production System is in full swing, having saved the hospital more than $2 million in inventory costs, $11 million in planned capital investments, and $500,000 in overtime and temporary labor expenses (the latter in just one year). Professional liability insurance premiums are down by 56 percent (for 2006 to 2011), and the turnaround times of many laboratory tests are down too.
How, exactly, has Virginia Mason achieved all this? In part by adopting a just-in-time philosophy of supply management; in part through introducing activities such as rapid process improvement workshops, during which staff members of all levels and departments meet to analyze and improve processes; and in part through a patient safety alert (PSA) system in which, just as at Toyota, every employee at every level has the power to halt a process when something unsafe occurs or is likely to occur.
To “stop the line,” an employee calls the patient safety department’s 24-hour hotline or uses an online reporting system. A patient safety specialist then designates the alert red, orange, or yellow, depending on its severity. Red alerts require resolution within 24 hours, while orange alerts must be resolved within 72 hours, and yellow alerts within one week. Each alert is transmitted to a senior executive who, in tandem with the relevant supervisor, is responsible for investigating the event, and who has the authority to stop a process or pull a piece of equipment or staff member “offline” until the problem is solved. The executive must also communicate with the employee who reported the PSA to let him or her know it has been resolved. The first month the PSA system was implemented, three alerts were issued; that number has now risen to an average of more than 550 per month.
“This replaced our old quality improvement system,” explains Dr. Galagan, “when you’d fill out a QI report and it would get sent to the other departments, where it often would get lost in the wash. What the PSA system allows us to do is pull in the pertinent parties and get things solved. It works pretty darn well.”
“For example, patient identification has been a big focus in our labs,” she continues. “We’re able to aggregate the PSA data in ways that allow us to then go work with the high-offender groups, look at their processes, and bring down their error rates. We’ve seen very good success with that.”
In one instance, signs were not being posted on the doors of patients under contact precautions, so that “our phlebotomists would go in there without realizing they were supposed to be gowning up,” she says. “It was such a dissatisfier. So they reported these as PSAs. The nurse administrator in charge convened a group, and they came up with a solution, and we kept having issues. It took two or three iterations, but we now have a solution that works. For the phlebotomists to be able to report that and actually see it improve was huge. It made them real believers. So now when I say to them, ‘Why don’t you put in a PSA?,’ they respond, ‘Okay, good idea.’”
The main challenge? Creating a culture in which employees really, truly knew that it was safe to report a PSA—that no one would be yelled at, ignored, or embarrassed for speaking up. To do that, it was crucial to get leadership on board, Furman says: “We knew from the very beginning that the executive team had to be on board and support this, because our [previous] culture was such that if a nurse or housekeeper or patient care technician tried to stop the line, that was not going to be felt to be safe. So that’s why the executive team agreed we would be the ones to assess the situation, to assess whether the line had to be stopped.”
In addition, since “health care providers learn by case study,” Furman says, “we would present case studies of different patient safety alerts at all of our usual meetings, like monthly managers’ meetings. We chose case studies that directly signaled our culture was going to change.” For example, when a physician exhibited disrespectful behavior toward a nurse, and as part of the same incident was found to have not followed a clinical standard of care, he was placed on administrative leave until he apologized, enrolled in anger management counseling, reviewed the literature on the relevant standard, and agreed to a plan to document his behavior improvement. “So we used that as a case study, to share that ‘We’ve got your back, and this type of behavior is no longer acceptable.’”
That message is still sinking in more slowly than Furman would like. “In our last AHRQ [Agency for Healthcare Research and Quality] Survey on Patient Safety Culture, we had over an 80 percent participation rate, which is more than double the average of all hospitals that use the AHRQ,” she says. “So we get really good participation, but we still have about 20 percent of our staff who are fearful to report things. We’re much better than most other hospitals. But we still have 20 percent. And I’ve been working on this for 10 years. I guess the message is: Culture is very slow to change.”
Still, Virginia Mason’s laboratory shows every sign of having embraced the PSA system. Take the case of the complex spine surgeries. “This is
a very complicated surgery, and the patients require a lot of coagulation testing support and sometimes blood product support,” Dr. Galagan says. “There was one case that could have gone better, and a PSA was filed because there was concern about not having been able to get the blood products and laboratory results fast enough.”
A group of anesthesiologists, blood bankers, laboratorians, and surgeons reviewed the case and came up with two solutions. “One was that we developed a massive transfusion protocol so that we could order all the things we needed simply from the blood bank. It also put the right level of urgency on the order,” she says. “We also looked at our fibrinogen assay and realized there was a big delay in getting a result, particularly if it was low, when they need it most. The test typically is run at a one-to-20 dilution. If it’s really low, the analyzer won’t be able to measure it. So we validated a laboratory-developed test and extended the linearity down to a lower range; we now start with a one-to-four dilution as well as the regular specimen that will get the one-to-20 dilution, and then if it is measuring too low, we have the one-in-four already in hand. That way we can give them a really fast turnaround on fibrinogen levels.”
In another example, a 2009 PSA was sparked when laboratory staff realized that a patient who had entered the ER with a CNS bleed had not received fresh frozen plasma as quickly as would have been ideal. (Because Seattle has a centralized blood bank system, the plasma must be thawed and then transported from that central bank to the hospital.) “So we looked at that in a [process improvement group] of neurologists, ED doctors, laboratorians, and blood bankers, to visit this in-depth, and what we realized was that we could use thawed plasma,” she says. “Even though the coagulation factors are a little bit lower, they’re still good for most purposes. So we switched the whole hospital to thawed plasma, which made for much better turnaround times for everyone, and also saved the hospital a lot of money—because we couldn’t return the fresh frozen plasma. It got wasted, whereas with thawed plasma, if we returned it within a certain time frame, we could get a credit.”
And then there’s the Creutzfeldt-Jakob question. “There was a case where a laboratorian handled a cerebrospinal fluid specimen, and there was a small spill, and then later it was found out that the patient tested positive for Creutzfeldt-Jakob disease,” Dr. Galagan says. “So the laboratorians put in a PSA, wanting some better signaling and processing methods for handling those cases. This involved laboratorians, neurologists, infection control, infectious disease, and employee health, as well as our information system partners.” (The spill caused no harm to laboratory staff.)
One element essential to creating a culture of continuous improvement, Furman says, is accepting that mistakes happen not necessarily because people are careless, but because processes themselves are flawed. “Nursing schools are still teaching nursing students that if they just try really hard and check things three times, they’re not going to make mistakes,” she says. “But they will, because they’re human. So we’re learning where our mistakes are, and then we’re creating what’s known as poka-yoke, or mistake-proof, processes, so that these defects don’t happen.”
In 2005, for example, the hospital instituted computerized physician order entry. Before that, “we had defect [test or medication ordering] rates as high as 85 percent,” Furman says. “When we implemented our CPOE, in one day we went down to a defect rate of 0.1 percent, and that was only because we didn’t implement that tool in our anesthesia and preop area.”
Also critical: Seeking the input of people outside your specialty. “They can look at processes we have and say, ‘Why are you doing that?’” Dr. Galagan says. At a recent improvement workshop she led on critical results reporting, for example, she came into it assuming that “some complicated IS thing” would need to be implemented. Instead, a young surgical resident said, “Why aren’t you using MPages?” “And we said, ‘MPages? What’s that?’” It turned out to be a Cerner program that allows users to extract data from the EMR into one page. “And this particular page is set up so it shows you the current resident, attending, and intern, with a pager button beside it, so all you have to do is go there, hit the button, and automatically page the person. Real cool—and we had no idea.”
Among Virginia Mason’s next steps will be expanding the patient safety concept to encompass staff safety as well. “People have to feel safe in order for us to keep our patients safe,” Furman points out. “So we’re going to be expanding our program to ensure that staff can speak up about personal safety, as well as focus on respect for people. How important that is—that every single staff member, every single day, can say, ‘I feel supported, I feel appreciated, and I have the resources I need to do my job.’”
Anne Ford is a writer in Evanston, Ill.