Scarier Than Ebola: Human Error

Texas Health Presbyterian Hospital in Dallas on Sept. 30 Photograph by LM Otero/AP Photo

The Dallas hospital treating the first Ebola case diagnosed in the U.S. sent the patient, Thomas Duncan, home the first time he showed up because the doctors who saw him never learned that he’d just come from West Africa. The hospital has blamed a flaw in its electronic health records for keeping information collected by a nurse, including Duncan’s travel history, from being presented to the treating physician, who mistook Duncan’s symptoms for a low-level infection, on Sept. 25. (Update: The hospital has since changed its explanation.)

The apparent mistake meant Duncan was not admitted and isolated until Sept. 28. That increased the risk of infection for those he came in contact with while he was sick, including his family, who are now quarantined in their Dallas apartment. It also widened the circle of contacts that public health officials must trace and monitor for symptoms.

America’s risk of an Ebola epidemic remains vanishingly small. The country has the public health resources and hospital capacity to stop the spread of the infection, which is only transmitted through direct contact with bodily fluids after a patient exhibits symptoms. The misstep at Texas Health Presbyterian Hospital Dallas, though, indicates something patients should be spooked about: the very real chance that errors, oversights, or deviations from established procedures could kill them.

It’s hard to say precisely how often this happens. A 2013 review of studies in the Journal of Patient Safety suggested medical errors cause somewhere between 210,000 and 400,000 deaths each year in the U.S. In a landmark report (PDF) 15 years ago, the Institute of Medicine put the number between 44,000 and 98,000. Even the lower estimate would mean medical errors kill more Americans than car accidents do. The moment when one clinician turns over care of a patient to another is particularly hazardous, says Marty Makary, a Johns Hopkins surgeon who has written on hospital safety. “The most dangerous procedure in American emergency rooms is a patient handoff,” Makary says. Breakdowns in communication during patient handoffs “are endemic in American health care,” he says.

Electronic health records have sometimes been hailed as a tool to help standardize care. Many doctors complain that they’re a distraction, collecting too much information without prioritizing the most important facts. In this case, the design of the software apparently stopped doctors from seeing crucial information that might have made a difference in their initial diagnosis. According to Texas Health Dallas, nurses and the doctor who initially saw Duncan followed proper procedures when he arrived with a fever, abdominal pain, and a headache.

The intake nurse took a travel history, along with an array of other information, and recorded in an electronic medical record that he had been in Africa within the past four weeks. That important fact never made it to the doctor who saw Duncan, though, because the software has separate workflows for nurses and doctors. “As designed, the travel history would not automatically appear in the physician’s standard workflow,” according to an e-mailed statement from Texas Health Dallas. That’s now been corrected. Travel history, including specific references to regions with Ebola outbreaks, has been made more visible “to alert all providers,” the hospital says.

It’s impossible to say whether the same mistake would have happened had the hospital been using paper charts. “Even in traditional paper and verbal communication, the lack of a headline is one of the greatest problems in relaying information that results in patient harm,” Makary says.

Texas Health, to its credit, released details of what went wrong “in the interest of transparency, and because we want other U.S. hospitals and providers to learn from our experience.” Let’s hope they do.

    Before it's here, it's on the Bloomberg Terminal.