Predictive instrument aids triage in ED

Real-time usage indicates likelihood of ACI

A predictive instrument developed by researchers at the New England Medical Center (NEMC) in Boston provides emergency department (ED) physicians with additional information to more accurately identify and triage patients who may be at risk for acute cardiac ischemia (ACI). In so doing, it helps reduce medical errors and avoid unnecessary admissions to the critical care unit (CCU). Called the ACI-TIPI (Acute Cardiac Ischemia Time-Insensitive Predictive Instrument), it generates a patient’s 0% to 100% probability of having ACI [acute myocardial infarction (MI) or unstable angina], to help ED physicians gauge the patient’s need for hospitalization. The ACI-TIPI is available for most ED electrocardiographs in the United States.

"Knowing precisely where to triage ED patients is not always a clear decision," points out Joni R. Beshansky, RN, MPH, assistant professor of medicine and associate director of the Center for Cardiovascular Health Services Research, Division of Clinical Care Research, at NEMC. "We thought if physicians could actually have some numerical probability of ACI, it would improve their recognition of whether or not patients actually had ACI and improve their ED triage decision making."

Predictive instruments, she explains, typically are developed to examine a process or treatment that is expensive, uses a lot of resources, or can involve a lot of people. "A large percentage of the people who present with chest pain in the ED will not have an MI or unstable angina." The software is based on a validated mathematical formula that predicts a patient’s probability of truly having ACI.

Developed through federally funded research, ACI-TIPI has evolved steadily, with initial research that began in 1979 and the first clinical trials taking place in 1984 with nearly 6,000 patients. The results were published in the New England Journal of Medi-cine. "It reduced unnecessary admissions to the CCU by 30%," Beshansky notes.

The problem was that the technology at the time had not caught up to the concept. "You had to use a calculator to determine the probability of ACI, but no one would want to walk around with a calculator," she says. Eventually, the percentages were noted in a chart hung on the ED wall, but they often suffered the fate of other wall charts — i.e., staff would tack things up on them, and they could be damaged easily.

A breakthrough occurred in 1993, when the instrument was incorporated into a computerized electrocardiograph by Phillips Technology and GE Medical Systems and tested in a clinical trial of more than 10,000 patients. Now, when a cardiogram is done, it prints directly from the TIPI device. In addition to the customary waveform printout, the physician receives an ACI-TIPI score in the header text illustrating the likelihood of ACI.

In addition to these probabilities, the printout includes patient demographic information, notes Denise H. Daudelin, RN, MPH, project director of the Medical Error Prevention Project, Clinical Care Research, at NEMC. "The ultimate goal of the information system is to take those probabilities and combine them with patient-specific information to create a powerful database and reporting system for medical error prevention, clinical practice evaluation, and quality improvement."

Beshansky notes that the instrument can be especially valuable to less-experienced physicians or unsupervised residents. The software is commercially available, at a cost of approximately $1,000, from Phillips or GE. "In Boston, it’s being used here at NEMC and at Brigham & Women’s Hospital in their ED," she says.

More recently, the TIPI information system has evolved with new technology and is being expanded to include the ability to conduct retrospective analyses. "On a computer in your office, you will have access via an intranet to an information system that stores data about all the patients who came into the ED and had an ECG, including probabilities and demographics," Daudelin says. "If you have lab results, those will be in the database, too."

Since there is no manual data entry or collection involved, she explains, there will be no need to add a full-time equivalent. "Reports can be run off that show average probabilities, or groups of patients by physician," she adds, "so doctors can look at their performance, the probabilities associated with those patients they admitted or sent home, and compare it to the average." This information must be accessed through an intranet. "We have great concern about the Health Insurance Portability and Accountability Act, so everything is stored within the firewall of the hospital," she says. "Anything that’s sent to us for benchmarking is encrypted."

With this retrospective capability, physicians can design specific indicator reports — i.e., patients above a certain probability who are sent home, or below a certain probability who are admitted to the intensive care unit (ICU). "You could determine, for example, whether some physicians send low-risk patients to the ICU too often, using more resources than necessary," Daudelin offers.

The system has any number of QI applications, she adds. "Quality managers can define the range of indicators they want to review. It also has an important error-reduction component. When patients are sent home meeting certain criteria, like a high probability of ACI, the system alerts the department that the patients were sent home, so the hospital can call and bring them back in."

The system also helps the physician complete his follow-up and shows when and if that follow-up was completed. "It also gives the physician a certain perspective they don’t have now on their patients," says Daudelin.

Beshansky also notes that the most recent Joint Commission on Accreditation of Healthcare Organizations standards on patient safety call for proactive identification of high-risk areas, "and this falls very nicely into that area."

Finally, she advises, it’s not enough for the quality manager to believe in the system. "You really have to have buy-in from the top down and from the bottom up — the quality manager, the ED director, and other parts of the organization as well," she says. "In essence, the whole hospital needs to buy in."

Need more information?

For more information on the Acute Cardiac Ischemia Time-Insensitive Predictive Instrument, contact:

Joni R. Beshansky, RN, MPH, Associate Director, The Center for Cardiovascular Health Services Research, Clinical Care Research, New England Medical Center, 750 Washington St., Box 63, Boston, MA 02111. E-mail:

Denise H. Daudelin, RN, MPH, Project Director, Medical Error Prevention Clinical Care Research, New England Medical Center, 750 Washington St., Box 63, Boston 02111. E-mail:

Clinical Care Systems Inc. Telephone: (541) 745-9023.