First, Do No Harm — Improving Patient Safety

reprinted from Issue 7, Fall 2008 of Frontiers of Medicine (PDF)

Robert Wachter, MD (center), chief of the Division of Hospital Medicine, reviews patient safety reports with
Andrew Auerbach, MD, and Arpana Vidyarthi, MD.
photo credit: Noah Berger


Until 10 years ago, we were taught in the health professions to believe that errors were manifestations of bad, careless people," says Robert Wachter, MD, chief of the Division of Hospital Medicine. "We now know that most errors are made by competent, well-trained, caring people trying to be careful, and the errors simply demonstrate they are human."

In 1999, an Institute of Medicine report entitled "To Err is Human: Building a Safer Health System" sent shockwaves through the medical community. It estimated that up to 98,000 people died each year due to medical errors.

Wachter has led UCSF's efforts to improve patient safety. In two bestselling books, Internal Bleeding: The Truth Behind America's Terrifying Epidemic of Medical Mistakes, and Understanding Patient Safety, Wachter champions a new perspective. Instead of blaming individuals, he argues for systemic changes to prevent mistakes from happening in the first place.

"The most common causes of medical mistakes are communication lapses — information didn't make it from place A to place B correctly, or from person A to person B correctly," says Wachter. Common mistakes include giving a patient the wrong drug or dose, performing surgery on the wrong patient or body part, or making the wrong diagnosis.

One example of how better communication has increased patient safety at UCSF is the way medical residents entrust their patients to other physicians at the end of their shifts — the handoff process.

Previously, this process was haphazard. Residents would spend lots of time hunting down key information — vital signs, lab reports and medication lists. Arpana Vidyarthi, MD, who arrived at UCSF as a hospital medicine fellow in 2002, says, "I distinctly remember seeing residents throw down a stack of index cards, two inches thick, with information for 50 patients hand-scribbled on them, and say to the next person, 'There's nothing to do.' Given how sick many of our patients are, this would cause me anxiety as I recognized the potential for harm during the cross-coverage."

In 2003, the Accreditation Council for Graduate Medical Education limited residents to an 80-hour workweek and maximum 30-hour shifts. Vidyarthi, who is now director of patient safety and quality programs for the Dean's Office of Graduate Medical Education, says, "The reductions in work hours were designed to reduce errors caused by fatigue." However, this has resulted in more frequent handoffs among residents, which also increases the risk of communication errors.

With the assistance of UCSF Medical Center, Vidyarthi and Jonathan Carter, MD (then a UCSF surgery resident) developed a computer-based system dubbed SynopSIS. It provides a snapshot of the most important information residents need to know during handoffs, including the patient's physical location and list of medications, why the patient was admitted, and anticipated problems.

Drawing on his or her understanding of each patient's case, the outgoing resident prepares a list of "if-then" statements: if the patient develops a fever, then test for infection and start certain antibiotics; if the patient becomes short of breath, then get an X-ray; if the patient is dying, call his daughter in Philadelphia at this phone number. Before SynopSIS, these vital pieces of information too often got lost in the handoff process.

Vidyarthi trains residents in best practices for these face-to-face meetings between outgoing and incoming residents. She recommends that they find a quiet place where they can review SynopSIS information. She also reminds them that tone of voice and facial expressions provide valuable information. Before the meeting ends, the incoming resident repeats back his or her understanding of the departing resident's recommendations.

"At the beginning of your internship, it takes a little longer, as you learn how to sign out effectively," Vidyarthi says. "By mid-year, using SynopSIS and verbally signing out is so ingrained in the culture that nobody thinks twice about it — it's like driving."

1 Picture = 1,000 Words

Dean Schillinger, MD
photo credit: Noah Berger

The opportunities to improve patient safety continue after discharge. "What goes on after the hospital, and in between visits?" asks Dean Schillinger, MD, director of the UCSF Center for Vulnerable Populations at San Francisco General Hospital (SFGH).

"Ninety-nine percent of the care is going on at their homes," Schillinger says. "The number of medications, the severity of people's illnesses, and the expectations we have for patients to self-manage their conditions have increased. The potential for patient safety issues to arise in the outpatient setting has worsened."

For example, Schillinger and Edward Machtinger, MD, found that nearly 50% of patients on blood thinners were unaware that they were taking their medication improperly. "These are very high-risk populations taking high-risk medications," says Schillinger. "I call it the 'Goldilocks medicine': you have to take it just right. You can't take too much, or you may bleed and die; you can't take too little, or you may have a stroke and die." Often, patients need to take different amounts of the drug on different days, further increasing the likelihood of incorrect dosage.

Visual medication
schedule (VMS)

Schillinger and Machtinger developed a visual medication schedule (VMS), a computer-generated weekly calendar showing the type and amount of medication to be taken each day, with written instructions in the patient's native language (see image at right). They also had patients "teach back" the dosage instructions to their doctors, so doctors could confirm that patients understood correctly. Their study showed that patients who received the VMS plus the "teach back" opportunity reached the target safe level for their anticoagulant almost twice as fast as patients who did not use this method. This tool was especially effective among Spanish-speaking patients.

"Do the Right Thing" by Default

Much of the Department of Medicine's safety and quality research is focused on how to get a life-saving treatment to its ultimate destination: the patient.

"How do you get physicians to adopt it?" asks Andrew Auerbach, MD, MPH, associate clinical professor of medicine and director of research for the Division of Hospital Medicine. "How do you get systems to deliver it regularly? Moreover, how do you measure that implementation process? In business, there are whole areas of management theory around how to manage change in complex systems. But in health care, that's a very underdeveloped field."

Auerbach recently redesigned UCSF's physician order forms to, in his words, "make it easier to do the right thing, and harder to do the not-right thing." For example, deep vein thrombosis, or clotting of the blood in a vein such as the leg, can occur shortly after surgery. It can be fatal if a clot travels to the lungs and obstructs blood flow, causing what is called pulmonary embolism.

Fortunately, this is completely preventable — if a patient receives the appropriate blood thinner. Until recently, however, only half of UCSF surgery patients received blood thinners. This was partly because there was no systematic way for surgeons to prescribe them. Three years ago, Auerbach developed an easy-to-use order form and trained surgeons on the importance of prescribing blood thinners.

"There are three legs to a quality improvement stool," says Auerbach. "Education — explain why this is the right thing to do; change the system; and then audit and feedback — we pull charts at random, and if patients did not get the right drug, we send a report to the physicians involved." Today, 95% of eligible UCSF surgery patients receive the proper blood thinners.

Auerbach is also developing the Hospital Medicine Reporting Network. This Network shares patient quality and safety data to provide benchmarking information. This allows hospitals to see where they need to improve and what they can learn from other institutions. "It's entirely possible that there's some innovation out there that would be easily disseminated to others," says Auerbach. "I think of that as the 'gene discovery' of quality. We have the ability now with large databases to start sifting around for those gems."

A Little Fear is Healthy

Robert Wachter, MD
photo credit: Noah Berger

"If there was a way to have gotten this job done without scaring people, that would have been better," says Wachter. "Systems are so recalcitrant to change that unless people did have some anxiety about the current state of affairs, we wouldn't have changed a thing.

"I got a call from a reporter from a state in the Midwest that had just begun requiring hospitals to report serious errors," says Wachter. "One hospital had 15 reports, and another hospital had zero. I said, 'You wouldn't catch me dead going to the hospital with zero — because they either have a culture in which nobody talks about these things, or they're lying.' The state of medicine is such that you can't have a hospital that does not periodically harm or kill somebody through errors.

"What I want to see is that hospitals are open and honest, using each error as an opportunity to make themselves better," says Wachter. "I'm very proud of our organization, because I think that's what we're doing. This is where I get my health care, and I know what's in the sausage factory."

Frontiers of Medicine Main Page