Computerized tools for diagnosing heart attack could be inaccurate in some people

May 1 2005

A computerized tool to help emergency room physicians determine whether a patient is having a heart attack may not work as well among some racial and ethnic groups, according to research of almost 12,000 patients at nine medical centers.

"It's notorious that women and elderly patients have markedly different heart attack symptoms from the younger male patient," said Chadwick D. Miller, M.D., from Wake Forest University School of Medicine. "This study shows us that race and ethnicity also play a role in symptoms."

Results from the research, conducted at Wake Forest and eight other medical centers, are reported in the May issue of Academic Emergency Medicine. The researchers studied a computerized risk stratification tool, called the Acute Coronary Ischemia-Time Insensitive Predictive Instrument (ACI-TIPI), which is designed to predict whether a patient is having a heart attack. Although ACI-TIPI itself is not widely used clinically, its elements form the basis of many other risk assessment tools.

There is no single, definitive test to diagnose heart attacks, making it difficult to evaluate chest pain patients. Risk assessment tools have become popular because they allow doctors to make "evidence-based" decisions based on age, gender, health history, questions about chest pain and an electrocardiogram.

"These tools have mostly been tested in an American, mixed-race population of patients. For example, a typical study population may be comprised of 60 percent Caucasian, 30 percent African-American and 10 percent Hispanic patients," said Miller, an instructor in emergency medicine at Wake Forest's School of Medicine, which is part of Wake Forest University Baptist Medical Center. "This design does not detect subtle differences that may exist among the groups."

It has been demonstrated that race and ethnicity influence both the perception of chest pain and the time it takes people to seek treatment. Miller said these differences may make the risk assessment tools inaccurate if they are applied to other population groups.

The study compared how well the tool performed in a mixed-race population in the United States versus an Asian population in Singapore.

"What we found was that in Singapore, patients were less likely to exhibit the typical symptom of heart attack: chest pain." Miller said. "Age and male gender also had little predictive power in evaluating whether these patients were having a heart attack."

Miller said the results suggest that doctors should consciously consider the effects of racial or ethnic differences when they use the tools. In addition, they point to the importance of taking ethnic differences into consideration when designing new tools.

"Given the previously demonstrated differences in ethnic groups, combined with our findings, one must question the utility of population-based cardiac risk assessment," said Miller. "We know that typical American patient presents with crushing chest pain. But, this approach doesn't take into the account the different ethnicities that might present differently."

For the study, the researchers analyzed data from a registry of patients with cardiac symptoms who came to the emergency departments of eight U.S. medical centers and one medical center in Singapore. Any patient who came in with symptoms that might be cardiac – chest pain, shortness of breath, etc. – was included in the study. The researchers looked at the accuracy of ACI-TIPI in predicting acute coronary syndrome (ACS).

ACS is an umbrella term used to cover any group of clinical symptoms compatible with chest pain due to insufficient blood supply to the heart muscle resulting from heart vessel disease. Out of 11,991 patients in the study, 1,120 were diagnosed with ACS.

Other researchers were V. Anantharaman, M.D., from Singapore General Hospital; Christopher Lindsell, Ph.D., and W. Brian Gibler, M.D., from the University of Cincinnati, Charles Pollack, M.S., M.D., and Judd Hollander, M.D., from the University of Pennsylvania, Brian Tiffany, M.D., Ph.D., from Chandler Emergency Medical Group,and James Hoekstra, M.D., and Julie Greenway, B.S., from Wake Forest Baptist.