skip navigation nih record
Vol. LVIII, No. 24
December 1, 2006

previous story

next story

To Improve Care, Talk About Mistakes
Medical Error Both Systemic and Individual, Says Weingart

On the front page...

In March 1984, Libby Zion, an 18-year-old with a history of depression, was admitted to New York Hospital. Eight hours later, she died.

Zion’s case resulted in a lawsuit in which a jury found that three medical residents had contributed to her death. In a series of errors, the residents had ordered Demerol even though, prior to admission, Zion had taken Nardil, a powerful anti-depressant that cross-reacts with it. The jury also found that the first-year resident should have heeded a nurse’s call to the patient’s bedside and should have requested


  Dr. Saul Weingart, director of the Center for Patient Safety at Dana-Farber Cancer Institute, speaks at NIH.  
  Dr. Saul Weingart, director of the Center for Patient Safety at Dana-Farber Cancer Institute, speaks at NIH.  

The Libby Zion case prompted several major investigations and ultimately changed the training of medical residents as New York became the first state to restrict their work hours and increase their supervision. By 2003, the Accreditation Council for Graduate Medical Education (ACGME) implemented similar changes for the 7,800 medical residency programs it oversees.

Nevertheless, recent estimates show that medical error causes between 40,000 to 90,000 preventable medical deaths per year and injures more than a million people.

“We are faulty engines,” says Dr. Saul Weingart, vice president for patient safety and director of the Center for Patient Safety at Dana-Farber Cancer Institute. In a recent Clinical Grand Rounds lecture, he called for “systems built and designed to make errors transparent, rather than blaming and shaming,” and invited his audience into the process.

“Hands up,” he said. “How many of you have made a mistake?” Lipsett Amphitheater showed a sea of hands. “How many of you have made a mistake as a clinician? Now—you don’t have to put your hand up this time—how many of you made a mistake that injured a patient?”

Weingart recalled the case of Boston Globe health care columnist Betsy Lehman, who died suddenly of an overdose of chemotherapy.

“I knew the intern on that case,” he said, “An ambiguous chemotherapy order led to Betsy Lehman’s death.”

In response to such seminal cases, he said, a 1999 study by the National Academy of Science’s Institute of Medicine revealed that more Americans died each year of medication errors than from workplace injuries and that annual U.S. costs ran as high as $29 billion. A recent follow-up study counts 1.5 million medication-related injuries per year.

Weingart outlined how the patient safety movement has grown and now affects policy change on several fronts: the quality interagency coordination task force and congressional hearings; the Agency for Healthcare Research and Quality; the Joint Commission on Accreditation of Healthcare Organizations (JCAHO); ACGME; National Quality Forum; and various advocacy groups.

Medical Errors: Definitions
  • Adverse event: an injury that results from medical care, not necessarily an error. Example: An allergic response to a first-time administration of penicillin.
  • Adverse drug event (ADE): an injury that results from a drug.
  • Preventable ADE stems from an error. Example: An allergic response to administration of penicillin, even though caregivers already have information about the allergy.
  • Potential ADE is a near miss or close call; some are intercepted events.

“There’s been a lot of action,” he said, citing JCAHO’s ’07 National Patient Safety Goals. Nonetheless, he noted “a little bit of push-back from physicians. Although error is ubiquitous, in a single year a physician may not be aware that he or she made any harmful error.”

Perhaps that’s because “patients most likely to be harmed are at the extremes of age”—pediatric patients and seniors—“and the sickest”—including those admitted for high-risk surgeries.

“Some researchers argue that the number of error-related deaths is overestimated, since many of these patients would have died anyway of their disease. My own view is that we underestimate the amount of harm and that even a low-ball number is too many.”

In routine medical practice, where a patient’s average face-time with a physician is 12 minutes and the average time to first interruption is 18 seconds, Weingart reported that 75 percent of patients have unanswered questions. Doctors fail to act; patients fail to inform; that can hurt when a patient is most vulnerable—say, right after hospital discharge. The average full-time primary care provider reviews 1,000 lab results a week—no wonder we need “safety scientists” to help design care improvements, he suggested.

Weingart showed cognitive models for how we err as individuals. “But errors are also attributes of systems,” he explained. “Sometimes the hazards line up like holes in Swiss cheese, as interconnected series of events. There are design flaws in health care and if we’re not free to talk about such errors we’ll never learn from them.”

He showed a slew of such flaws, including one handwritten order so illegible that Weingart called its signer “Dr. Zorro, who is privileged at many institutions.” Another order included the scrawled comment—or plea—“me forget.” As for electronic orders with pre-programmed alerts, in one study 90 percent of alerts were overridden.

“‘We can’t improve the human condition,’” said Weingart, quoting psychology professor James Reason, who studies human error, “‘but we can improve the conditions under which humans labor.’ How do we build a system so this never happens again?”

NCI’s Dr. John Cole (l) and Robin Brown are OECs—occupant
Weingart shows how seminal cases of patient injury sparked a safety movement.

He called for teamwork training, which includes assertiveness, good communication and big-picture awareness. He also cautioned clinicians to beware of drugs with high adverse-drug-event rates such as heparin, insulin, chemotherapeutics, narcotics and antibiotics.

Other suggestions seemed pointed at institutional policies on workloads: “Do not rely on memory or vigilance,” he said dryly, “and expect excellent performance with fatigue.”

Weingart’s own current projects include studies of the patient’s role in identifying errors and adverse events; patient safety leadership; and creating medical error curricula for clinicians.

The reaction of the audience—chock-full of clinicians—included the rueful laughter of recognition.

“I’ve given this lecture to audiences who were afraid to laugh,” Weingart said, “and that can be dangerous. It’s a sign that a place is very hierarchical, that people are afraid of looking dumb.”

A doc who’d survived a big-city hospital fellowship seemed frustrated: “From the point of view of the house officer,” he said, “there is a Kafkaesque sense of things. What’s the point of complaining? What do you do if you are over capacity?” Research and education are crucial, Weingart said. “Teach patients and families to advocate for themselves,” he urged. “Bring a list of medications along with you. Challenge the doctor, but find a way that works; don’t make the doc feel defensive.” He also noted that the CC’s high nurse-to-patient ratio is a huge plus.

Asked about trends in countries with a national health service, he responded, “We found the same rate and type of error in the U.K. and Canada.”

What’s amazing, he said, is seeing the spread of best practices. Johns Hopkins’ rate of central line infections is down to zero; and while average hand-hygiene compliance runs from 30 to 40 percent, Dana-Farber’s is now as high as 95 percent. Improvement can be dramatic, he said. NIHRecord Icon

back to top of page