Infectious Disease
Biomedical Informatics
Oral Health Policy
Radiology/Biomedical Engineering
Research Briefs
Around & About
Point of View

A bleeding male patient is brought to the emergency department at 3 a.m. on a Sunday. At the same time the emergency room doctor must stabilize the patient, he or she must quickly run down a mental checklist: Has the man been in a fight? Is the injury drug-related? Does he have psychiatric problems?

The physician notices that the man has a needle mark on his arm. Does this mean he might have recently received a transfusion or that he is an intravenous drug user? The doctor determines the patient has a fever, chills, and an abnormal heart beat. He may have an infection at the site of the needle mark.

Situations such as these are rife for one of the unwanted outcomes of medical care: medical errors. Although most patients receive good treatment in critical care settings, a high-pressure environment, in which action must be taken quickly and medicines often given without adequate information about the patient, can lead to mistakes.

Now, Vimla Patel, Ph.D., professor of biomedical informatics and psychiatry, and her colleagues at Columbia University Medical Center are studying how decisions are made in critical care settings with the goal of understanding and minimizing the number of medical errors. Her research collaborators include Desmond Jordan, M.D., associate professor of clinical anesthesiology, Carlos Almeida, M.D., director of psychiatric emergency services, Robert Green, M.D., M.P.H., assistant professor of clinical medicine and associate director of emergency services, and Edward Shortliffe, M.D., Ph.D., professor and chairman of biomedical informatics. This study is being conducted in collaboration with Jiajie Zhang, Ph.D., and his colleagues at the University of Texas School of Health Information Sciences in Houston.

"I don't think medical errors can be eliminated entirely because there is always some human error, but the number and severity of mistakes can be reduced," Dr. Patel says.

A 1999 Institute of Medicine report on the prevalence of medical errors has generated considerable interest and awareness, both among the medical profession and the public. But it's not a new area for Dr. Patel, who's been researching the topic, along with how medical students reason and learn, for two decades. Dr. Patel came to Columbia in August 2000 from McGill University in Montreal, where she was director of the school's research center for medical education.

Dr. Patel and her colleagues want to develop a "cognitive framework" of medical errors in critical care environments, associating each category of medical error with a specific cognitive mechanism.

"Our philosophy is to take small groups of subjects, such as health personnel in the emergency department or the intensive care unit, and to study in great detail communication, policy, workflow, training, competence, and the role of technology in everyday functioning," Dr. Patel says. "We study the whole picture or process, which is messy but real. Then we do a deep analysis of the underlying reasoning used in making decisions about patient care."

The research is an expansion of a pilot study in which Dr. Patel analyzed activity in one 24-hour period in the CUMC intensive care unit last year. She and her colleagues studied the actions of a resident, a faculty doctor, and a student. They found that although the expert – the faculty doctor – made errors, he corrected them more quickly than the resident or the student. The expert also corrected very significant, life-threatening errors and let the less important ones go because they could be fixed later. The resident corrected a mix of serious and minor errors. The student corrected the least number of errors.

The pilot study showed that experts make as many errors as nonexperts but the errors differ in nature. "The difference is that a good expert tends to manage errors and fix them quickly, before serious problems can develop," Dr. Patel says. "Others don't know they've made a mistake. By the time they recognize an action as a mistake, it's too late."

Dr. Patel and her colleagues want to determine which factors – stress, understaffing, lack of knowledge, or overtaxing a provider's memory, among others – push students, inexperienced health care providers, and even experts to practice outside the boundaries of error-free medicine into the realm of making mistakes. With the stressors identified, the critical care environment could be better managed, physicians could be trained differently, or suitable technology could be introduced to reduce errors.

The researchers expect that their results will help doctors, nurses and hospital staff develop better ways of working together so that they are more likely to anticipate an error and prevent it, immediately correct it, or at least minimize it before it can adversely affect a patient. Such information could also prove useful in helping medical device companies design equipment to prevent a health care worker from giving the wrong medication or the wrong dose to a patient.

"You can't simply identify errors and try to reduce them," Dr. Patel says. "That approach won't change anything significantly because you don't necessarily understand the root causes of the mistakes. Mistakes are inevitable, but you have to learn to manage them. Once you know exactly why errors happen, you can train health care personnel to adapt better to the many different situations with which they are confronted, especially in critical care settings."

Dr. Patel's research is funded in part by the National Library of Medicine.

—Matthew Dougherty