Computerbased decision aids

Acute cardiac ischemia time-insensitive predictive instrument

The acute cardiac ischemia time-insensitive predictive instrument (ACI-TIPI) [26] computes a 0% to 100% probability that a given patient has ACS (either acute MI or unstable angina pec-toris) (see Tables 1 and 2). Applicable to any patient in the ED presenting with any symptom suggestive of ACS, it is based on a logistic regression equation that uses presenting symptoms and ECG variables. Originally in hand-held calculator form, it is now incorporated into conventional electrocardiographs so that the patient's ACI-TIPI probability is printed with the standard ECG header text. In large controlled interven-tional trials in a wide range of hospitals, its use by ED physicians has been shown to reduce unnecessary admissions of patients who do not have ACSs and patients who have stable angina, while not reducing appropriate hospitalization for patients who have ACSs. It has also been shown to help the triage speed and accuracy of less-trained and less-supervised residents. Although this decision aid is a widely available software option offered by all the major ECG machines manufacturers in the United States, it has not been widely used in clinical practice. The greater dissemination and use of ACI-TIPI could result in significant positive impact on the triage of patients who have ACSs in the ED.

Thrombolytic predictive instrument

Correctly diagnosing STEMI for prompt use of coronary reperfusion therapy can be lifesaving. In EDs this can be difficult, especially for less obvious candidates [121-123]. Efforts by physician leaders, the NHAAP, health care organizations and monitoring entities, and pharmaceutical companies [123-132], have helped increase use and promptness of reperfusion therapy [133]. Further improvement is needed [133,134], however, particularly for other than anterior STEMI, the category for which thrombolytic therapy first was recognized as effective [121,122] and for women who have received less coronary reperfusion therapy than men [134,135]. Also, a need remains for ways to support prompt and accurate coronary reperfusion therapy decisions in hospitals and pre-hospital EMS settings where consultation with off-site physicians is required.

Selker and colleagues [136] developed the thrombolytic predictive instrument (TPI), a collection of five component predictive instruments designed to accurately assess the patient-specific likely benefits and risks from the use of thrombo-lytic therapy for STEMI. The TPI helps clinicians identify patients for coronary reperfusion therapy based on their probabilities of benefits and complications and facilitates earliest possible use of reperfusion therapy [123,136]. In conventional computerized ECGs when significant ST segment elevation of STEMI is detected, the TPI predictions are automatically computed and printed on the ECG header: probabilities for acute (30-day) mortality either treated with thrombolytic therapy or untreated; 1-year mortality rates treated with thrombolytic therapy; cardiac arrest treated with thrombolytic therapy and untreated; thrombolytic therapy-related stroke and major bleeding requiring transfusion.

To test whether the ECG-based TPI improves ED selection of patients for coronary reperfusion therapy and promptness of treatment, the authors ran a 22-month randomized controlled clinical effectiveness trial of its impact on the use of thrombolytic therapy and overall coronary reperfusion therapy. The trial ran in EDs at 28 urban, suburban, and rural hospitals across the United States, from major cardiac centers to small community hospitals. Study endpoints were percentages of patients receiving (a) thrombolytic therapy; (b) thrombolytic therapy within 1 hour of ED presentation; and (c) all coronary reperfusion therapy, either by thrombolytic therapy or PTCA.

At participating hospitals, software generating TPI predictions was installed on their conventional computerized ECGs. When a significant ST elevation characteristic of STEMI was automatically detected, the ECG randomly assigned the patient to the control or intervention group. If assigned to the intervention group, the ECG automatically prompted the user to enter information needed to compute the TPI predictions: age, sex, history of hypertension or of diabetes, blood pressure, and time since ischemic symptom onset. The remaining variables, based on ECG waveform measurements, were automatically acquired by the ECG. Then the ECG was printed with TPI predictions on its header.

Among the 1197 patients who developed STEMI, the trial showed that the TPI increased use of thrombolytic therapy, use of thrombolytic therapy within 1 hour, and use of overall coronary reperfusion therapy by 11% to 12% for patients who have developed inferior STEMI, 18% to 22% for women, and 30% to 34% for patients who have an off-site physician [137]. Although the TPI's effect was minimal on patients who had high baseline coronary reperfusion therapy rates, such as men who present with anterior STEMIs, for the groups needing the most improvement in their rates of recognition for coronary reperfusion, women and those with less obvious STEMIs, and where involved physicians were off-site, the TPI increased recognition of STEMI and use and timeliness of coronary reperfusion therapy. It is hoped that as TPI-capable ECGs become more widely available in ED and EMS settings, its use will facilitate the accuracy and speed of recognition and treatment of these patients.

Goldman Chest Pain Protocol

The Goldman Chest Pain Protocol [7] is based on a computer-derived model using recursive-partitioning analysis to predict MI in patients who have chest pain (see Tables 1 and 2). It has good sensitivity (about 90%) for AMI, but was not developed to also detect UAP. In a clinical impact study of ''low-intensity, non-intrusive intervention'' performed in the ED of a teaching hospital [16], no differences in hospitalization rate, length of stay, or estimated costs were demonstrated between the experimental group that used the protocol and the control group. Goldman and colleagues [35] eventually switched to predicting the need for intensive or other levels of care. They found that clinicians who had higher levels of training had a higher sensitivity for detecting AMI, but at the expense of decreased specificity [138]. Reilly and colleagues [139] developed a consensus to adapt the Goldman prediction rule and found a favorable impact on physicians' hospital triage decisions solely by different triage decisions for very low-risk patients. Unfortunately, any algorithm that incorporates only clinical elements and the ECG findings at presentation is likely to be suboptimal because of the substantial proportion of patients who present with atypical symptoms or with no or minimal ECG abnormalities.

Other computer-based decision aids

Several investigators have reported various computer-based decision aids to diagnose AMI (see Tables 1 and 2). The artificial neural network by Baxt and Skora [140] had high sensitivity and high specificity for AMI in a prospective study, but the clinical impact has not been demonstrated. A predictive model with automatic ECG interpretation has been shown to increase the use of fibri-nolytic therapy for acute ST-segment elevation MI, especially in historically undertreated patients, such as those who have inferior AMI and women who have AMI [141].


It is becoming apparent that one of the major contributing factors for the misdiagnosis of ische-mic heart disease in women is pathophysiologic differences that appear to exist between men and women who have these syndromes and between the various ischemic syndromes themselves. Obstructive CAD continues to be a major public health problem in women and has represented the leading cause of death and disability for more than a decade [1,142,143]. In the United States alone, more than a quarter of a million women die of coronary heart disease (CHD) each year, translating into 1 death every 2 minutes [1]. Increased awareness of these statistics and a recent focus on women's health issues in general have resulted in renewed interest in trying to understand important gender differences in patients who have ACSs. Although women have a lower prevalence of obstructive CAD compared with men who have similar symptoms [144-146], women have a higher frequency of angina/chest pain than men partly because of the higher prevalence of the less common causes of ischemia, such as vaso-spastic angina and microvascular endothelial dysfunction; transient left ventricular apical ballooning syndrome (takotsubo cardiomyopathy) [147]; and syndromes of nonischemic chest pain, such as mitral valve prolapse [148]. Add to this the fact that young women who have obstructive CAD experience a significantly worse outcome with regard to prognosis after MI compared with men [149], and that older women who have obstructive CAD often have greater comorbidities that influence their outcome adversely after AMI or myocardial revascularization than do men [150-153]. Furthermore, women presenting with ACSs are less likely to receive effective acute diagnostic and treatment strategies than men [3,154,155]. When women develop obstructive CAD, they have a greater functional expression of their disease and disability compared with men. Most women who do not have obstructive CAD at angiography continue to have symptom-related disability and consume considerable health care resources [156-158], partly because the patho-physiology of ischemia in women is incompletely understood and gender-specific diagnostic and treatment strategies are underdeveloped [159]. Gender differences with regard to ischemic heart disease appear to exist in several areas, including established and novel risk factors; the metabolic syndrome; the physiology of endogenous reproductive hormones; the role of endothelial dysfunction in producing obstructive macrovascular CAD, myocardial ischemia, chronic chest pain syndromes, and ACS; genetic factors; proteomics; the menstrual cycle and reproductive status; pain threshold/perception; neurohumoral control; and behavioral/psychosocial factors.

The diagnostic evaluation of women who have suspected ACSs continues to be a major challenge. Knowing whether gender influences the likelihood that a given patient in the ED is having ACSs, and whether any specific presenting clinical features are differentially associated with ACSs in women compared with men, can aid clinicians in the accurate diagnosis of ACSs. The incidence of AMI in the general population has been shown to be higher in men than women [160-163], but until recently it has not been clear whether this gender difference holds among symptomatic patients who come to the ED. In addition, several studies have looked at gender differences in the presentation of patients who have AMI [157,164-167]. In a retrospective analysis of patients who have confirmed AMI, women had higher rates of atypical presentations, such as abdominal pain, paroxysmal dyspnea, or congestive heart failure [46,160,168-170]. In a group of patients in the ED who had typical presentations such as chest pain, the prevalence of AMI was lower in women [34,171]. However, in another study of patients in the ED who had chest pain, when adjustments were made for other presenting clinical features (specifically ECG), the gender difference was no longer significant [166]. From these results it is difficult to assess whether the gender-specific differences in AMI prevalence among symptomatic patients in the ED were the result of gender-specific biology or limitations in a particular study's patient selection.

One reason for such challenges is that chest pain in women is neither sensitive nor specific in predicting the presence of underlying CHD. The highest sensitivity is found in women presenting with symptoms of typical angina pectoris, whereas the highest specificity is found in women presenting with nonspecific symptoms. In fact, although women who have ACSs may present with symptom patterns that differ from men (ie, atypical symptoms), for most women, typical symptoms are the strongest symptom predictors of ACSs [172-174]. A further challenge is gender differences in reporting pain, including but not limited to chest pain. Gender differences in endogenous pain-modulating systems may contribute to differences in pain perception. Regarding biomarkers, there is a suggestion that women who have obstructive CHD may have a different pattern of presenting biomarkers compared with men. In a TACTICS-TIMI 18 subanalysis examining patients who have UA/NSTEMI, there was a different pattern of presenting biomarkers between the sexes: men were more likely to have elevated CK-MB and troponins, whereas women were more likely to have elevated C-reactive protein and brain natriuretic peptide, suggesting that a multimarker approach to diagnosis may improve triage in the ED. Lastly, the Women's Ischemic Syndrome Evaluation Study Group and others have suggested that those women who have chest pain without flow-limiting lesions by angiography may have associated microvascular endothelial dysfunction (cardiac syndrome X) [175] and impaired coronary flow reserve [176,177]. Preliminary data suggest that coronary microvascular dysfunction is associated with an increased rate of hospitalization for chest pain, poor quality of life, and ongoing health care costs. Newer technologies, such as MR spectroscopy and gadolinium cardiac MRI, may allow for the identification of abnormalities in vascular function and structure not identifiable by coronary angiogram that can cause or contribute to development of myocardial ischemia, and may aid in the initial risk assessment of UA/NSTEMI, especially in women [178].

Finally, numerous studies have found that women have poorer outcomes than men after a diagnosis of ACSs. Explanations have included gender differences in pathophysiology and response to treatment; prehospital delays in symptom recognition and action; and gender differences regarding evaluation and treatment in emergency medical services. Recent reports of similar or better outcomes in women who have ACSs compared with men suggest that pathophysiologic differences can be overcome with early, aggressive therapy [153,179]. The diagnosis of ischemia influences prognosis differently in various clinical ischemic syndromes. In unstable angina the detection of ischemia indicates the likelihood of persistence or recurrence of plaque instability and hence carries much more severe prognostic implications than in chronic stable angina, whereas although angina in patients who have normal coronary angiograms is not associated with increased short-term risk of infarction or sudden death, these patients may be crippled by pain. Inconsistent response to nitrates and antianginal drugs indicates the need for research on the various potential causes of coronary vascular dysfunction that burden women disproportionately, to facilitate development of rational forms of therapy.

Your Heart and Nutrition

Your Heart and Nutrition

Prevention is better than a cure. Learn how to cherish your heart by taking the necessary means to keep it pumping healthily and steadily through your life.

Get My Free Ebook

Post a comment