Prognostic Accuracy of the HEART Score for Prediction of Major Adverse Cardiac Events in Patients Presenting With Chest Pain: A Systematic Review and Meta-analysis

Preclinical identification of acute coronary syndrome without high sensitivity troponin assays using machine learning algorithms

Preclinical management of patients with acute chest pain and their identification as candidates for urgent coronary revascularization without the use of high sensitivity troponin essays remains a critical challenge in emergency medicine. We enrolled 2760 patients (average age 70 years, 58.6% male) with chest pain and suspected ACS, who were admitted to the Emergency Department of the University Hospital Tübingen, Germany, between August 2016 and October 2020. Using 26 features, eight Machine learning models (non-deep learning models) were trained with data from the preclinical rescue protocol and compared to the "TropOut" score (a modified version of the "preHEART" score which consists of history, ECG, age and cardiac risk but without troponin analysis) to predict major adverse cardiac event (MACE) and acute coronary artery occlusion (ACAO). In our study population MACE occurred in 823 (29.8%) patients and ACAO occurred in 480 patients (17.4%). Interestingly, we found that all machine learning models outperformed the "TropOut" score. The VC and the LR models showed the highest area under the receiver operating characteristic (AUROC) for predicting MACE (AUROC = 0.78) and the VC showed the highest AUROC for predicting ACAO (AUROC = 0.81). A SHapley Additive exPlanations (SHAP) analyses based on the XGB model showed that presence of ST-elevations in the electrocardiogram (ECG) were the most important features to predict both endpoints.

Evaluating the performance of the HEART score in a Tanzanian emergency department

OBJECTIVE

The HEART score successfully risk stratifies emergency department (ED) patients with chest pain in high-income settings. However, this tool has not been validated in low-income countries.

METHODS

This is a secondary analysis of a prospective observational study that was conducted in a Tanzanian ED from January 2019 through January 2023. Adult patients with chest pain were consecutively enrolled, and their presenting symptoms and medical history were recorded. Electrocardiograms and point-of-care troponin assays were obtained for all participants. Thirty-day follow-up was conducted, assessing for major adverse cardiac events (MACEs), defined as death, myocardial infarction, or coronary revascularization (coronary artery bypass grafting or percutaneous coronary intervention). HEART scores were calculated for all participants. Likelihood ratios, sensitivity, specificity, and negative predictive values (NPVs) were calculated for each HEART cutoff score to predict 30-day MACEs, and area under the curve (AUC) was calculated from the receiver operating characteristic curve.

RESULTS

Of 927 participants with chest pain, the median (IQR) age was 61 (45.5-74.0) years. Of participants, 216 (23.3%) patients experienced 30-day MACEs, including 163 (17.6%) who died, 48 (5.2%) with myocardial infarction, and 23 (2.5%) with coronary revascularization. The positive likelihood ratio for each cutoff score ranged from 1.023 (95% CI 1.004-1.042; cutoff ≥ 1) to 3.556 (95% CI 1.929-6.555; cutoff ≥ 7). The recommended cutoff of ≥4 to identify patients at high risk of MACEs yielded a sensitivity of 59.4%, specificity of 52.8%, and NPV of 74.7%. The AUC was 0.61.

CONCLUSIONS

Among patients with chest pain in a Tanzanian ED, the HEART score did not perform as well as in high-income settings. Locally validated risk stratification tools are needed for ED patients with chest pain in low-income countries.

Pediatric Chest Pain: A Review of Diagnostic Tools in the Pediatric Emergency Department

Pediatric chest pain is a common chief complaint in the emergency department. Not surprisingly, children with chest pain are usually brought to the emergency department by their parents out of fear of heart disease. However, chest pain in the pediatric population is generally a benign disease. In this review, we have identified musculoskeletal pain as the most prevalent etiology of chest pain in the pediatric population, accounting for 38.7-86.3% of cases, followed by pulmonary (1.8-12.8%), gastrointestinal (0.3-9.3%), psychogenic (5.1-83.6%), and cardiac chest pain (0.3-8.0%). Various diagnostic procedures are commonly used in the emergency department for cardiac chest pain, including electrocardiogram (ECG), chest radiography, cardiac troponin examination, and echocardiography. However, these examinations demonstrate limited sensitivity in identifying cardiac etiologies, with sensitivities ranging from 0 to 17.8% for ECG and 11.0 to 17.2% for chest radiography. To avoid the overuse of these diagnostic tools, a well-designed standardized algorithm for pediatric chest pain could decrease unnecessary examination without missing severe diseases.

Predictors of major adverse cardiac events among patients with chest pain and low HEART score in the emergency department

AIM

For patients who present to the emergency departments (ED) with undifferentiated chest pain, the risk of major adverse cardiac events (MACE) may be underestimated in low-HEART score patients. We aimed to identify characteristics of patients who were classified as low risk by HEART score but subsequently developed MACE at 6 weeks.

METHODS

We studied a multiethnic cohort of patients who presented with chest pain arousing suspicion of acute coronary syndrome to EDs in the Netherlands and Singapore. Patients were risk-stratified using HEART score and followed up for MACE at 6 weeks. Risk factors of developing MACE despite low HEART scores (scores 0-3) were identified using logistic and Cox regression models.

RESULTS

Among 1376 (39.8%) patients with low HEART scores, 63 (4.6%) developed MACE at 6 weeks. More males (53/806, 6.6%) than females (10/570, 2.8%) with low HEART score developed MACE. There was no difference in outcomes between ethnic groups. Among low-HEART score patients with 2 points for history, 21% developed MACE. Among low-HEART score patients with 1 point for troponin, 50% developed MACE, while 100% of those with 2 points for troponin developed MACE. After adjusting for HEART score and potential confounders, male sex was independently associated with increased odds (OR 4.12, 95%CI 2.14-8.78) and hazards (HR 3.93, 95%CI 1.98-7.79) of developing MACE despite low HEART score.

CONCLUSION

Male sex, highly suspicious history and elevated troponin were disproportionately associated with MACE. These characteristics should prompt clinicians to consider further investigation before discharge.

Early emergency department discharge for intermediate heart score patients presenting for chest pain

Study Objective

The use of the HEART score to risk stratify patients for short-term major adverse cardiac events in the emergency department (ED) setting is well established. Although discharge to home for low-risk HEART score patients is widely accepted as safe practice, there are limited outcomes data on moderate-risk HEART score patients discharged to home. We investigated the safety of discharging moderate-risk HEART score patients to home from the ED with established early cardiology follow-up.

Methods

We performed a retrospective cohort analysis of patients presenting to the ED with chest pain from April 2020 through December 2020. Patients were evaluated in the ED and underwent serial conventional troponin testing and electrocardiogram (ECG). Clinicians calculated a HEART score and employed shared decision-making with moderate-risk patients (score 4-6), offering hospital admission versus discharge home with a formalized process for rapid cardiology follow-up (within 2 business days). We assessed the frequency of acute myocardial infarction or death at 30 days and before cardiology follow-up.

Results

During our study period, 2939 patient encounters were screened for chest pain. Of these, 333 of 547 eligible moderate-risk HEART score patients were referred for rapid follow-up. The median time to follow-up appointment was 2.9 business days (interquartile range 1.3, 6.5), and 264 (79%) of patients kept their follow-up appointment. One patient (0.3%) suffered death within 30 days, before cardiology follow-up. There were no myocardial infarctions.

Conclusions

These results suggest that moderate-risk HEART score patients may be considered for discharge from the ED with rapid cardiology follow-up. Formalizing processes to facilitate these early evaluations may represent a viable alternative to hospital admission, without diminishing patient outcomes.

The potential of HEART score to detect the severity of coronary artery disease according to SYNTAX score

Clinical scoring systems such as the HEART score can predict major adverse cardiovascular events, but they cannot be used to demonstrate the degree and severity of coronary artery disease. We investigated the potential of HEART Score in detecting the existence and severity of coronary artery disease based on SYNTAX score. This multi-centric cross-sectional study investigated patients referred to the cardiac emergency departments of three hospitals between January 2018 and January 2020. Data including age, gender, risk factors, comorbidities, 12-lead ECG, blood pressure and echocardiogram were recorded for all the participants. Serum troponin I level was measured on admission and 6 h later. Coronary angiography was done via the femoral or radial route. HEART and SYNTAX scores were calculated for all patients and their association was assessed. 300 patients (65% female) with mean age of 58.42 ± 12.42 years were included. mean HEART Score was 5.76 ± 1.56 (min = 3, max = 9), and mean SYNTAX score was 14.82 ± 11.42 (min = 0, max = 44.5). Pearson correlation coefficient was 0.493 between HEART Score and SYNTAX score which was statistically significant (P < 0.001). We found that HEART Score of more than 6 is 52% sensitive and 74.7% specific to detect extensive coronary artery involvement (SNTAX score ≥ 23). The present study showed that the HEART score has a moderate and positive correlation with the SYNTAX score and HEART score with a cut-off value of 6 is a predictor for SYNTAX score of ≥ 23.

Utility of serial conventional troponin testing for emergency department patients stratified by HEART score and symptom timing

BACKGROUND

The HEART score for risk stratifying chest pain patients in the emergency department (ED) has been widely adopted in clinical practice, but is often employed with nonconformant serial troponin measurements.

OBJECTIVE

The primary objective of this study was to examine the utility of obtaining a second conventional 3-h troponin I (TnI) level in ED patients presenting with potential acute coronary syndrome (ACS), stratified by HEART score and duration of symptoms.

METHODS

This was a retrospective cohort study of consecutive adult ED patients with a complete HEART score. We assessed the utility of repeat TnI measurement by examining the positivity rate of ΔTnI = [Second TnI] - [Initial TnI] stratified by HEART score and time elapsed since onset or resolution of symptoms. Major adverse cardiac events (MACE) within 6 weeks of index visit were assessed.

RESULTS

A total of 944 patients were included with 433 (45.9%) assigned a low risk HEART score 0-3. Of the 268 (61.9%) low risk HEART score patients receiving a second TnI, only 3 (1.1%, [0.2-3.2%]) resulted in a positive ΔTnI, one of which occurred in the setting of an elevated initial TnI. Overall, patients presenting within 3 h of symptoms were more likely to experience positive ΔTnI, index MACE and MACE at 6 weeks compared to patients presenting ≥3 h since symptoms onset/resolution and patients with unknown timing of symptoms (15.9% vs 11.0% vs 10.3%, p < 0.001; 10.0% vs 5.3% vs 4.6%, p = 0.021; 12.7% vs 6.6% vs 6.4%, p = 0.047).

CONCLUSION

Our data suggest serial measurement of conventional troponin provides limited added benefit in low risk HEART score patients, regardless of duration and timing of symptoms. Conversely, serial troponin measurement may confer utility in moderate/high risk HEART score patients, particularly those presenting within 3 h of symptoms.

Computer-Generated ECG Interpretation Challenge

The predictive accuracy of 12-lead electrocardiogram (ECG) machines is often challenged across all clinical settings. Emergency clinicians must beware of computer-generated ECG reports specifically during the initial medical screening process. Blindly trusting computer-generated reports may delay care for patients with an acute cardiac disorder. Cardiology consultation is always advised, and there should be no hesitation when it comes to abnormal ECGs. However, cardiologists are often consulted on patients based on incorrect ECG interpretation, misdiagnosis, or overdiagnosis by computer-generated reports. The following 12-lead ECGs should encourage emergency providers to take caution and challenge computer-generated reports. The purpose of this exercise is to carefully review a set of 12-lead ECGs and determine whether the computer-generated interpretations are accurate.

A Closer Look at the HEART Score

The history, electrocardiogram, age, risk factors, and troponin (HEART) score is currently a widely used tool for acute chest pain risk stratification. Relatively soon after its inception in 2008, a number of validation studies on the HEART score showed it to be superior to Thrombolysis in Myocardial Infarction (TIMI) and Global Registry of Acute Coronary Events (GRACE) scores and at least as accurate to other existing scores for predicting short-term major adverse cardiovascular events (MACEs). However, partly due to its focus on simplicity, the HEART score has some limitations. In this article we review how the HEART score has evolved and taken on various modifications to circumvent some of its limitations. We also highlight the strength of the HEART score in comparison with other risk stratification tools and the current guidelines.

Can ED chest pain patients with intermediate HEART scores be managed as outpatients?

PURPOSE

Current guidelines recommend hospital admission for patients who present to the emergency department (ED) with chest pain and are scored as intermediate risk for adverse outcomes based on the HEART score. While hospital admission for these patients allows for timely investigation and treatment, it is a resource-intensive process. This study examines whether intermediate HEART score patients can be safely managed on an outpatient basis through rapid access chest pain clinics.

METHODS

This retrospective observational study included all ED chest pain patients referred to rapid access clinics from January 2018 to April 2020 in Regina and Saskatoon, Saskatchewan. ED physician HEART scores were used in lieu of reviewer HEART scores when available. The primary outcome was the rate of major adverse coronary events (MACE), a composite measure of death, acute coronary syndrome, stroke, coronary angiography, and revascularization at 6 weeks in intermediate-risk patients. Secondary outcomes were the type of MACE, rate of MACE before rapid access clinic appointment and the most predictive component of the HEART score.

RESULTS

There were 1989 ED referrals, of which 817 were for intermediate-risk patients. 9.3% of intermediate-risk patients had a MACE at 6 weeks. MACE occurred before rapid access clinic follow-up in 1.1% of intermediate-risk patients, with coronary angiography being the most common MACE. Excluding coronary angiography, the risk of MACE before rapid access clinic follow-up was 0.7% in intermediate-risk patients. Components of the HEART score most predictive of MACE were troponin (OR 11.0, 95% CI: 3.7-32.3) and history (5.3, 95% CI: 2.4-11.8).

CONCLUSION

This study demonstrates that rapid access clinics are likely a safe alternative to admission for intermediate-risk chest pain patients and could reduce costly inpatient admissions for chest pain. With angiography excluded, MACE rates were well below the American College of Emergency Physicians cited 2% threshold.

Qualitative Evaluation of an Artificial Intelligence–Based Clinical Decision Support System to Guide Rhythm Management of Atrial Fibrillation: Survey Study

Background

Despite the numerous studies evaluating various rhythm control strategies for atrial fibrillation (AF), determination of the optimal strategy in a single patient is often based on trial and error, with no one-size-fits-all approach based on international guidelines/recommendations. The decision, therefore, remains personal and lends itself well to help from a clinical decision support system, specifically one guided by artificial intelligence (AI). QRhythm utilizes a 2-stage machine learning (ML) model to identify the optimal rhythm management strategy in a given patient based on a set of clinical factors, in which the model first uses supervised learning to predict the actions of an expert clinician and identifies the best strategy through reinforcement learning to obtain the best clinical outcome—a composite of symptomatic recurrence, hospitalization, and stroke.

Objective

We qualitatively evaluated a novel, AI-based, clinical decision support system (CDSS) for AF rhythm management, called QRhythm, which uses both supervised and reinforcement learning to recommend either a rate control or one of 3 types of rhythm control strategies—external cardioversion, antiarrhythmic medication, or ablation—based on individual patient characteristics.

Methods

Thirty-three clinicians, including cardiology attendings and fellows and internal medicine attendings and residents, performed an assessment of QRhythm, followed by a survey to assess relative comfort with automated CDSS in rhythm management and to examine areas for future development.

Results

The 33 providers were surveyed with training levels ranging from resident to fellow to attending. Of the characteristics of the app surveyed, safety was most important to providers, with an average importance rating of 4.7 out of 5 (SD 0.72). This priority was followed by clinical integrity (a desire for the advice provided to make clinical sense; importance rating 4.5, SD 0.9), backward interpretability (transparency in the population used to create the algorithm; importance rating 4.3, SD 0.65), transparency of the algorithm (reasoning underlying the decisions made; importance rating 4.3, SD 0.88), and provider autonomy (the ability to challenge the decisions made by the model; importance rating 3.85, SD 0.83). Providers who used the app ranked the integrity of recommendations as their highest concern with ongoing clinical use of the model, followed by efficacy of the application and patient data security. Trust in the app varied; 1 (17%) provider responded that they somewhat disagreed with the statement, “I trust the recommendations provided by the QRhythm app,” 2 (33%) providers responded with neutrality to the statement, and 3 (50%) somewhat agreed with the statement.

Conclusions

Safety of ML applications was the highest priority of the providers surveyed, and trust of such models remains varied. Widespread clinical acceptance of ML in health care is dependent on how much providers trust the algorithms. Building this trust involves ensuring transparency and interpretability of the model.

Optical electrocardiogram monitor with a real-time analysis of an abnormal heart rhythm for home-based medical alerts

Sudden cardiac death (SCD) caused by cardiovascular disease is the greatest hidden danger to human life, accounting for about 25% of the total deaths in the world. Due to the early concealment of SCD and the heavy medical burden of long-term examination, telemedicine combined with home monitoring has become a potential medical alert method. Among all the existing human cardiac and electrophysiology monitoring methods, optics-based sensors attract the widest attention due to the advantages of low delay, real-time monitoring, and high signal-to-noise ratio. In this paper, we propose an optical sensor with the capabilities of long-term monitoring and real-time analysis. Combining an R-peak recognition algorithm, Lorenz plots (LP), and statistical analysis, we carried out the consistency analysis and result visualization of ECG sequences over 1 h. The results of 10 subjects show that the R-peak recognition accuracy of the optical ECG monitor is higher than 97.99%. The optical system can display abnormal heart rhythm in real-time through LP, and the readability is good, which makes the system suitable for self-monitoring at home. In addition, this paper provides a detailed long-term monitoring assessment method to effectively guide the practical clinical transformation of other optical wearable devices.

There is little association between prehospital delay, persistent symptoms, and post-discharge healthcare utilization in patients evaluated for acute coronary syndrome

AIMS

Test for an association between prehospital delay for symptoms suggestive of acute coronary syndrome (ACS), persistent symptoms, and healthcare utilization (HCU) 30-days and 6-months post hospital discharge.

BACKGROUND

Delayed treatment for ACS increases patient morbidity and mortality. Prehospital delay is the largest factor in delayed treatment for ACS.

METHODS

Secondary analysis of data collected from a multi-center prospective study. Included were 722 patients presenting to the Emergency Department (ED) with symptoms that triggered a cardiac evaluation. Symptoms and HCU were measured using the 13-item ACS Symptom Checklist and the Froelicher's Health Services Utilization Questionnaire-Revised instrument. Logistic regression models were used to examine hypothesized associations.

RESULTS

For patients with ACS (n = 325), longer prehospital delay was associated with fewer MD/NP visits (OR, 0.986) at 30 days. Longer prehospital delay was associated with higher odds of calling 911 for any reason (OR, 1.015), and calling 911 for chest related symptoms (OR, 1.016) 6 months following discharge. For non-ACS patients (n = 397), longer prehospital delay was associated with higher odds of experiencing chest pressure (OR, 1.009) and chest discomfort (OR, 1.008) at 30 days. At 6 months, longer prehospital delay was associated with higher odds of upper back pain (OR, 1.013), palpitations (OR 1.014), indigestion (OR, 1.010), and calls to the MD/NP for chest symptoms (OR, 1.014).

CONCLUSIONS

There were few associations between prehospital delay and HCU for patients evaluated for ACS in the ED. Associations between prolonged delay and persistent symptoms may lead to increased HCU for those without ACS.

Composite Outcomes in Clinical Prediction Modeling: Are We Trying to Predict Apples and Oranges?

Composite outcomes are widely used in clinical research. Existing literature has considered the pros and cons of composite outcomes in clinical trials, but their extensive use in clinical prediction has received much less attention. Clinical prediction assists decision-making by directing patients with higher risks of adverse outcomes toward interventions that provide the greatest benefits to those at the greatest risk. In this article, we summarize our existing understanding of the advantages and disadvantages of composite outcomes, consider how these relate to clinical prediction, and highlight the problem of key predictors having markedly different associations with individual components of the composite outcome. We suggest that a "composite outcome fallacy" may occur when a clinical prediction model is based on strong associations between key predictors and one component of a composite outcome (such as mortality) and used to direct patients toward intervention when these predictors actually have an inverse association with a more relevant component of the composite outcome (such as the use of a lifesaving intervention). We propose that clinical prediction scores using composite outcomes should report their accuracy for key components of the composite outcome and examine for inconsistencies among predictor variables.

Disparities in cardiovascular outcomes among emergency department patients with mental illness

BACKGROUND

Patients with mental illness have been shown to receive lower quality of care and experience worse cardiovascular (CV) outcomes compared to those without mental illness. This present study examined mental health-related disparities in CV outcomes after an Emergency Department (ED) visit for chest pain.

METHODS

This retrospective cohort included adult Medicaid beneficiaries in Washington state discharged from the ED with a primary diagnosis of unspecified chest pain in 2010-2017. Outcomes for patients with any mental illness (any mental health diagnosis or mental-health specific service use within 1 year of an index ED visit) and serious mental illness (at least two claims (on different dates of service) within 1 year of an index ED visit with a diagnosis of schizophrenia, other psychotic disorder, or major mood disorder) were compared to those of patients without mental illness. Our outcomes of interest were the incidence of major adverse cardiac events (MACE) within 30 days and 6 months of discharge of their ED visit, defined as a composite of death, acute myocardial infarction (AMI), CV rehospitalization, or revascularization. Secondary outcomes included cardiovascular diagnostic testing (diagnostic angiography, stress testing, echocardiography, and coronary computed tomography (CT) angiography) rates within 30 days of ED discharge. Only treat-and-release visits were included for outcomes assessment. Hierarchical logistic random effects regression models assessed the association between mental illness and the outcomes of interest, controlling for age, gender, race, ethnicity, Elixhauser comorbidities, and health care use in the past year, as well as fixed year effects.

RESULTS

There were 98,812 treat-and-release ED visits in our dataset. At 30 days, enrollees with any mental illness had no differences in rates of MACE (AOR 0.96; 95% CI, 0.72-1.27) or any of the individual components. At 6 months, enrollees with any mental illness (AOR 1.86; 95% CI, 1.11-3.09) and serious mental illness (AOR 2.60; 95% CI 1.33-5.13) were significantly more likely to be hospitalized for a CV condition compared to those without mental illness. Individuals with any mental illness had higher rates of testing at 30 days (AOR 1.16; 95% CI 1.07-1.27).

CONCLUSION

Patients with mental illness have similar rates of MACE, but higher rates of certain CV outcomes, such as CV hospitalization and diagnostic testing, after an ED visit for chest pain.

The reduction of race and gender bias in clinical treatment recommendations using clinician peer networks in an experimental setting

Bias in clinical practice, in particular in relation to race and gender, is a persistent cause of healthcare disparities. We investigated the potential of a peer-network approach to reduce bias in medical treatment decisions within an experimental setting. We created "egalitarian" information exchange networks among practicing clinicians who provided recommendations for the clinical management of patient scenarios, presented via standardized patient videos of actors portraying patients with cardiac chest pain. The videos, which were standardized for relevant clinical factors, presented either a white male actor or Black female actor of similar age, wearing the same attire and in the same clinical setting, portraying a patient with clinically significant chest pain symptoms. We found significant disparities in the treatment recommendations given to the white male patient-actor and Black female patient-actor, which when translated into real clinical scenarios would result in the Black female patient being significantly more likely to receive unsafe undertreatment, rather than the guideline-recommended treatment. In the experimental control group, clinicians who were asked to independently reflect on the standardized patient videos did not show any significant reduction in bias. However, clinicians who exchanged real-time information in structured peer networks significantly improved their clinical accuracy and showed no bias in their final recommendations. The findings indicate that clinician network interventions might be used in healthcare settings to reduce significant disparities in patient treatment.

The "NUTS" statistic: Applying an EBM disease model to defensive medicine

Physicians believe that malpractice concerns result in unnecessary testing, and many emergency physicians state that avoiding malpractice is a contributing factor to ordering medically unnecessary tests. Unfortunately, defensive medicine does not come without possible harm to patients who may be subject to non-beneficial, downstream testing, procedures, and hospitalizations. We submit a novel statistic, "NUTS" or "Number of Unnecessary Tests to avoid one Suit. " We calculated a NUTS of 4737 for troponin testing in ED patients with suspected myocardial infarction, meaning a clinician will need to order 4737 medically unnecessary troponin tests to avoid one missed myocardial infarction lawsuit. The NUTS framework offers us an evidence-based lens to examine defensive medicine less superstitiously and more based on currently available data.

Applications of machine learning to undifferentiated chest pain in the emergency department: A systematic review

Background

Chest pain is amongst the most common reason for presentation to the emergency department (ED). There are many causes of chest pain, and it is important for the emergency physician to quickly and accurately diagnose life threatening causes such as acute myocardial infarction (AMI). Multiple clinical decision tools have been developed to assist clinicians in risk stratifying patients with chest. There is growing recognition that machine learning (ML) will have a significant impact on the practice of medicine in the near future and may assist with diagnosis and risk stratification. This systematic review aims to evaluate how ML has been applied to adults presenting to the ED with undifferentiated chest pain and assess if ML models show improved performance when compared to physicians or current risk stratification techniques.

Methods and findings

We conducted a systematic review of journal articles that applied a ML technique to an adult patient presenting to an emergency department with undifferentiated chest pain. Multiple databases were searched from inception through to November 2020. In total, 3361 articles were screened, and 23 articles were included. We did not conduct a metanalysis due to a high level of heterogeneity between studies in both their methods, and reporting. The most common primary outcomes assessed were diagnosis of acute myocardial infarction (AMI) (12 studies), and prognosis of major adverse cardiovascular event (MACE) (6 studies). There were 14 retrospective studies and 5 prospective studies. Four studies reported the development of a machine learning model retrospectively then tested it prospectively. The most common machine learning methods used were artificial neural networks (14 studies), random forest (6 studies), support vector machine (5 studies), and gradient boosting (2 studies). Multiple studies achieved high accuracy in both the diagnosis of AMI in the ED setting, and in predicting mortality and composite outcomes over various timeframes. ML outperformed existing risk stratification scores in all cases, and physicians in three out of four cases. The majority of studies were single centre, retrospective, and without prospective or external validation. There were only 3 studies that were considered low risk of bias and had low applicability concerns. Two studies reported integrating the ML model into clinical practice.

Conclusions

Research on applications of ML for undifferentiated chest pain in the ED has been ongoing for decades. ML has been reported to outperform emergency physicians and current risk stratification tools to diagnose AMI and predict MACE but has rarely been integrated into practice. Many studies assessing the use of ML in undifferentiated chest pain in the ED have a high risk of bias. It is important that future studies make use of recently developed standardised ML reporting guidelines, register their protocols, and share their datasets and code. Future work is required to assess the impact of ML model implementation on clinical decision making, patient orientated outcomes, and patient and physician acceptability.

Trial registration

International Prospective Register of Systematic Reviews registration number: CRD42020184977.

A Prospective Evaluation of Clinical HEART Score Agreement, Accuracy, and Adherence in Emergency Department Chest Pain Patients

STUDY OBJECTIVE

The HEART score is a risk stratification aid that may safely reduce chest pain admissions for emergency department patients. However, differences in interpretation of subjective components potentially alters the performance of the score. We compared agreement between HEART scores determined during clinical practice with research-generated scores and estimated their accuracy in predicting 30-day major adverse cardiac events.

METHODS

We prospectively enrolled adult ED patients with symptoms concerning for acute coronary syndrome at a single tertiary center. ED clinicians submitted their clinical HEART scores during the patient encounter. Researchers then independently interviewed patients to generate a research HEART score. Patients were followed by phone and chart review for major adverse cardiac events. Weighted kappa; unweighted Cohen's kappa; prevalence-adjusted, bias-adjusted kappa (PABAK); and test probabilities were calculated.

RESULTS

From November 2016 to June 2019, 336 patients were enrolled, 261 (77.7%) were admitted, and 30 (8.9%) had major adverse cardiac events. Dichotomized HEART score agreement was 78% (kappa 0.48, 95% confidence interval [CI] 0.37 to 0.58; PABAK 0.57, 95% CI 0.48 to 0.65) with the lowest agreement in the history (72%; WK 0.14, 95% CI 0.06 to 0.22) and electrocardiogram (85%; WK 0.4, 95% CI 0.3 to 0.49) components. Compared with researchers, clinicians had 100% sensitivity (95% CI 88.4% to 100%) (versus 86.7%, 95% CI 69.3% to 96.2%) and 27.8% specificity (95% CI 22.8% to 33.2%) (versus 34.6%, 95% CI 29.3% to 40.3%) for major adverse cardiac events. Four participants with low research HEART scores had major adverse cardiac events.

CONCLUSION

ED clinicians had only moderate agreement with research HEART scores. Combined with uncertainties regarding accuracy in predicting major adverse cardiac events, we urge caution in the widespread use of the HEART score as the sole determinant of ED disposition.

A Methodological Appraisal of the HEART Score and Its Variants

We performed a methodological appraisal of the history, electrocardiogram, age, risk factors, and troponin (HEART) score and its variants in the context of Annals of Emergency Medicine's methodological standards for clinical decision rules. We note that this chest pain risk stratification tool was not formally derived, omits sex and other known predictors, has weak interrater reliability, and its 0, 1, and 2 score weightings do not align with their known predictivities. Its summary performance (pooled sensitivities of 96% to 97% with lower confidence interval bounds of 93% to 94%) is below that which emergency physicians state a willingness to accept, below the 98% sensitivity exhibited by baseline practice without the score, and below the 1% to 2% acceptable miss threshold specified by the American College of Emergency Physicians chest pain policy. Two variants (HEART Pathway, HEART-2) have the same inherent structural limitations and demonstrate slightly better but still suboptimal sensitivity. Although a simple prediction tool for chest pain outcomes is appealing, we believe that the widespread use of the HEART score and its variants should be reconsidered.

The Role of Using HEART Score to Risk Stratify Chest Pain Among Emergency Department High Utilizers

The HEART score is used to effectively risk stratify undifferentiated chest pain patients in the Emergency Department (ED). It is unclear whether such risk stratification can be applied among ED high utilizers. We aim to determine the efficacy and safety of using the HEART score to predict 30-day short-term major adverse cardiac events (MACE) in ED high utilizers. We conducted a retrospective, observational study in which ED high utilizers were defined as patients who had four or more ED visits within the past 12 months. ED high utilizers presenting at the study ED with chest pain were enrolled. Patients in which the HEART score was utilized were placed in the HEART group and patients with no HEART scores documented were placed to the usual care group. Hospital admissions and cardiac stress tests performed during the index hospitalizations, and 30-day MACE rates were analyzed and compared between the HEART and usual care groups. From January 1, 2017 to December 31, 2019, a total of 8,315 patient visits from ED high utilizers were enrolled. In the HEART group, 49% of ED visits were admitted with 20% receiving stress tests. A 30-day MACE outcome occurred among 1.4% of visits. In the usual care group, 44% of ED visits were admitted, with only 9% receiving index stress tests and a 1.5% of 30-day MACE occurrence (p=0.727). The study showed that similar short-term MACE outcomes occurred between patients using HEART scores and usual care to risk stratify chest pain among ED high utilizers.

HEART Score of Four for Age and Risk Factors: A Case Series

Chest pain is a frequent chief complaint in the ED. Identifying acute coronary syndrome (ACS) and establishing proper disposition for further risk assessment for major adverse cardiac events are paramount. The HEART Score is a key decision-making tool used to determine patient risk and disposition. One scenario with a potential drawback of the HEART Score is found in patients with a score of four based solely on age and risk factors. The HEART Score categorizes a score of three or less as low risk, and patients with scores above this threshold are typically admitted. We present six cases of chest pain presenting to a military emergency department with a score of four based solely on age and risk factors. They represent every such case found in a previously created database used to validate the HEART Score. We followed each case forward one year in electronic medical records to identify major adverse cardiac events. With the exception of one case that was placed on hospice for non-cardiac reasons and subsequently lost to follow up, there were no adverse events. There is a rising concern for increasing hospital admission rates, overuse of resources, and cost. We highlight that this subset of HEART Score patients requires a more nuanced risk stratification in the ED. It may be worth the time and effort to risk stratify this subset with coronary computed tomography angiography. This additional effort may help reduce admission at such a patient's current and future presentations to the ED for chest pain.

Chest Pain Evaluation in the Emergency Department: Risk Scores and High-Sensitivity Cardiac Troponin

PURPOSE OF REVIEW

As many as 10 million patients present annually to the emergency department in the USA with symptoms concerning for acute myocardial infarction. The use of risk scores for patients with chest pain or equivalent without ST-segment elevation on the electrocardiogram. The adaptation in the USA of high sensitivity troponin assays requires rethinking of how to best optimize troponin testing within a risk score.

RECENT FINDINGS

Patients are risk stratified using a combination of validated risk scores, biomarkers, and both noninvasive and invasive testing. The advent of high-sensitivity troponins has served to augment existing risk scores in the identification of low-risk patients for early discharge, as well as led to the introduction of new rapid rule-out protocols by which acute myocardial infarction can be excluded by biomarker evaluation more quickly. The emergence of machine learning algorithms may further enhance provider's ability to quickly diagnose or exclude myocardial infarction in the emergency department. The addition of high sensitive troponin assays to established emergency department risk scores is providing new opportunities to improve the timeliness and accuracy of the evaluation of patients presenting with a possible myocardial infarction. Utilizing the time between troponin measures as a variable combined with clinical risk factors with new algorithms may further serve to improve diagnostic accuracy.

Heart rate n-variability (HRnV) and its application to risk stratification of chest pain patients in the emergency department

Background

Chest pain is one of the most common complaints among patients presenting to the emergency department (ED). Causes of chest pain can be benign or life threatening, making accurate risk stratification a critical issue in the ED. In addition to the use of established clinical scores, prior studies have attempted to create predictive models with heart rate variability (HRV). In this study, we proposed heart rate n-variability (HRnV), an alternative representation of beat-to-beat variation in electrocardiogram (ECG), and investigated its association with major adverse cardiac events (MACE) in ED patients with chest pain.

Methods

We conducted a retrospective analysis of data collected from the ED of a tertiary hospital in Singapore between September 2010 and July 2015. Patients > 20 years old who presented to the ED with chief complaint of chest pain were conveniently recruited. Five to six-minute single-lead ECGs, demographics, medical history, troponin, and other required variables were collected. We developed the HRnV-Calc software to calculate HRnV parameters. The primary outcome was 30-day MACE, which included all-cause death, acute myocardial infarction, and revascularization. Univariable and multivariable logistic regression analyses were conducted to investigate the association between individual risk factors and the outcome. Receiver operating characteristic (ROC) analysis was performed to compare the HRnV model (based on leave-one-out cross-validation) against other clinical scores in predicting 30-day MACE.

Results

A total of 795 patients were included in the analysis, of which 247 (31%) had MACE within 30 days. The MACE group was older, with a higher proportion being male patients. Twenty-one conventional HRV and 115 HRnV parameters were calculated. In univariable analysis, eleven HRV and 48 HRnV parameters were significantly associated with 30-day MACE. The multivariable stepwise logistic regression identified 16 predictors that were strongly associated with MACE outcome; these predictors consisted of one HRV, seven HRnV parameters, troponin, ST segment changes, and several other factors. The HRnV model outperformed several clinical scores in the ROC analysis.

Conclusions

The novel HRnV representation demonstrated its value of augmenting HRV and traditional risk factors in designing a robust risk stratification tool for patients with chest pain in the ED.

Hospital Medicine Update: High-Impact Literature from March 2018 to April 2019

To assist busy hospital medicine clinicians, we summarized 10 impactful articles from last year. The authors reviewed articles published between March 2018-April 2019 for the Hospital Medicine Updates at the Society of Hospital Medicine and the Society of General Internal Medicine Annual Meetings. The authors voted to select 10 of 30 presented articles based on quality and clinical impact for this summary. The key findings include: (1) Vancomycin or fidaxomicin are the first-line treatment for initial Clostridioides difficile infection; (2) Unnecessary supplemental oxygen is linked to increased mortality; aim for a target oxygen saturation of 90%-94% in most hospitalized patients; (3) Stigmatizing language in medical records impacts physician trainees' attitudes and pain management practices; (4) Consider ablation for atrial fibrillation in patients with heart failure; (5) Patients with opioid use disorder should be offered buprenorphine or methadone therapy; (6) Apixaban is safe and may be preferable over warfarin in patients with atrial fibrillation and end-stage kidney disease; (7) It is probably safe to discontinue antimethicillin-resistant Staphylococcus aureus (MRSA) coverage in patients with hospital-acquired pneumonia who are improving and have negative cultures; (8) Selected patients with left-sided endocarditis (excluding MRSA) may switch from intravenous (IV) to oral antibiotics if they are clinically stable after 10 days; (9) Oral antibiotics may be equivalent to IV antibiotics in patients with joint and soft tissue infections; (10) A history-electrocardiogram-age-risk factors-troponin (HEART) score ≥4 is a reliable threshold for determining the patients who are at risk for short-term major adverse cardiac events and may warrant further evaluation.