Machine learning for the diagnosis of acute coronary syndrome using a 12-lead ECG: a systematic review.

OBJECTIVES: Prompt diagnosis of acute coronary syndrome (ACS) using a 12-lead electrocardiogram (ECG) is a critical task for emergency physicians. While computerized algorithms for ECG interpretation are limited in their accuracy, machine learning (ML) models have shown promise in several areas of clinical medicine. We performed a systematic review to compare the performance of ML-based ECG analysis to clinician or non-ML computerized ECG interpretation in the diagnosis of ACS for emergency department (ED) or prehospital patients. METHODS: We searched Medline, Embase, Cochrane Central, and CINAHL databases from inception to May 18, 2022. We included studies that compared ML algorithms to either clinicians or non-ML based software in their ability to diagnose ACS using only a 12-lead ECG, in adult patients experiencing chest pain or symptoms concerning for ACS in the ED or prehospital setting. We used QUADAS-2 for risk of bias assessment. Prospero registration CRD42021264765. RESULTS: Our search yielded 1062 abstracts. 10 studies met inclusion criteria. Five model types were tested, including neural networks, random forest, and gradient boosting. In five studies with complete performance data, ML models were more sensitive but less specific (sensitivity range 0.59-0.98, specificity range 0.44-0.95) than clinicians (sensitivity range 0.22-0.93, specificity range 0.63-0.98) in diagnosing ACS. In four studies that reported it, ML models had better discrimination (area under ROC curve range 0.79-0.98) than clinicians (area under ROC curve 0.67-0.78). Heterogeneity in both methodology and reporting methods precluded a meta-analysis. Several studies had high risk of bias due to patient selection, lack of external validation, and unreliable reference standards for ACS diagnosis. CONCLUSIONS: ML models have overall higher discrimination and sensitivity but lower specificity than clinicians and non-ML software in ECG interpretation for the diagnosis of ACS. ML-based ECG interpretation could potentially serve a role as a "safety net", alerting emergency care providers to a missed acute MI when it has not been diagnosed. More rigorous primary research is needed to definitively demonstrate the ability of ML to outperform clinicians at ECG interpretation.

RéSUMé: OBJECTIFS: Le diagnostic rapide du syndrome coronarien aigu (SCA) à l'aide d'un électrocardiogramme à 12 dérivations (ECG) est une tâche essentielle pour les urgentologues. Bien que la précision des algorithmes informatisés pour l'interprétation de l'ECG soit limitée, les modèles d'apprentissage automatique (ML) se sont révélés prometteurs dans plusieurs domaines de la médecine clinique. Nous avons effectué une revue systématique pour comparer la performance de l'analyse ECG basée sur le ML à l'interprétation ECG informatisée clinicienne ou non-ML dans le diagnostic du SCA pour les urgences (ED) ou les patients préhospitaliers. MéTHODES: Nous avons effectué des recherches dans les bases de données Medline, Embase, Cochrane Central et CINAHL de la création au 18 mai 2022. Nous avons inclus des études qui comparaient les algorithmes de ML à des cliniciens ou à des logiciels non basés sur ML dans leur capacité à diagnostiquer le SCA en utilisant uniquement un ECG à 12 dérivations, chez des patients adultes présentant des douleurs thoraciques ou des symptômes concernant le SCA dans le cadre de l'urgence ou préhospitalier. Nous avons utilisé QUADAS-2 pour l'évaluation du risque de biais. Prospero registration CRD42021264765. RéSULTATS: Notre recherche a donné 1062 résumés. 10 études satisfaisaient aux critères d'inclusion. Cinq types de modèles ont été testés, dont les réseaux neuronaux, la forêt aléatoire et le gradient boosting. Dans cinq études avec des données de performance complètes, les modèles de ML étaient plus sensibles mais moins spécifiques (plage de sensibilité 0,59-0,98, plage de spécificité 0,44-0,95) que les cliniciens (plage de sensibilité 0,22-0,93, plage de spécificité 0,63-0,98) dans le diagnostic du SCA. Dans quatre études qui l'ont rapporté, les modèles de ML avaient une meilleure discrimination (zone sous la courbe ROC plage 0,79-0,98) que les cliniciens (zone sous la courbe ROC 0,67-0,78). L'hétérogénéité de la méthodologie et des méthodes de déclaration a empêché une méta-analyse. Plusieurs études présentaient un risque élevé de biais en raison de la sélection des patients, du manque de validation externe et de normes de référence peu fiables pour le diagnostic du SCA. CONCLUSIONS: Les modèles de ML ont globalement une discrimination et une sensibilité plus élevées mais une spécificité plus faible que les cliniciens et les logiciels non-ML dans l'interprétation de l'ECG pour le diagnostic du SCA. L'interprétation de l'ECG basée sur le ML pourrait servir de « filet de sécurité ¼, alertant les fournisseurs de soins d'urgence d'une IM aiguë manquée lorsqu'elle n'a pas été diagnostiquée. Des recherches primaires plus rigoureuses sont nécessaires pour démontrer définitivement la capacité du ML à surpasser les cliniciens lors de l'interprétation de l'ECG.

Provision of medical assistance in dying: a scoping review.

OBJECTIVES: The purpose of this study is to map the characteristics of the existing medical literature describing the medications, settings, participants and outcomes of medical assistance in dying (MAID) in order to identify knowledge gaps and areas for future research. DESIGN: Scoping review. SEARCH STRATEGY: We searched electronic databases (MEDLINE, EMBASE, PsychINFO, CINAHL and CENTRAL), clinical trial registries, conference abstracts and professional guidelines from jurisdictions where MAID is legal, up to February 2020. Eligible report types included technical summaries, institutional policies, practice surveys, practice guidelines and clinical studies that describe MAID provision in adults who have provided informed consent for MAID. RESULTS: 163 articles published between 1989 and 2020 met eligibility criteria. 75 studies described details for MAID administered by intravenous medications and 50 studies provided data on oral medications. In intravenous protocols, MAID was most commonly administered using a barbiturate (34/163) or propofol (22/163) followed by a neuromuscular blocker. Oral protocols most often used barbiturates alone (37/163) or in conjunction with an opioid medication (7/163) and often recommended using a prokinetic agent prior to lethal drug ingestion. Complications included prolonged duration of the dying process, difficulty in obtaining intravenous access and difficulty in swallowing oral agents. Most commonly, the role of physicians was prescribing (83/163) and administering medications (75/163). Nurses' roles included administering medications (17/163) and supporting the patient (16/163) or family (13/163). The role of families involved providing support to the patient (17/163) and bringing medications from the pharmacy for self-administration (4/163). CONCLUSIONS: We identified several trends in MAID provision including common medications and doses for oral and parenteral administration, roles of healthcare professionals and families, and complications that may cause patient, family and provider distress. Future research should aim to identify the medications, dosages, and administration techniques and procedures that produce the most predictable outcomes and mitigate distress for those involved.

Methods for the Selection of Covariates in Nutritional Epidemiology Studies: A Meta-Epidemiological Review.

BACKGROUND: Observational studies provide important information about the effects of exposures that cannot be easily studied in clinical trials, such as nutritional exposures, but are subject to confounding. Investigators adjust for confounders by entering them as covariates in analytic models. OBJECTIVE: The aim of this study was to evaluate the reporting and credibility of methods for selection of covariates in nutritional epidemiology studies. METHODS: We sampled 150 nutritional epidemiology studies published in 2007/2008 and 2017/2018 from the top 5 high-impact nutrition and medical journals and extracted information on methods for selection of covariates. RESULTS: Most studies did not report selecting covariates a priori (94.0%) or criteria for selection of covariates (63.3%). There was general inconsistency in choice of covariates, even among studies investigating similar questions. One-third of studies did not acknowledge potential for residual confounding in their discussion. CONCLUSION: Studies often do not report methods for selection of covariates, follow available guidance for selection of covariates, nor discuss potential for residual confounding.

Red and Processed Meat Consumption and Risk for All-Cause Mortality and Cardiometabolic Outcomes: A Systematic Review and Meta-analysis of Cohort Studies.

This article has been corrected. The original version (PDF) is appended to this article as a Supplement. Background: Dietary guidelines generally recommend limiting intake of red and processed meat. However, the quality of evidence implicating red and processed meat in adverse health outcomes remains unclear. Purpose: To evaluate the association between red and processed meat consumption and all-cause mortality, cardiometabolic outcomes, quality of life, and satisfaction with diet among adults. Data Sources: EMBASE (Elsevier), Cochrane Central Register of Controlled Trials (Wiley), Web of Science (Clarivate Analytics), CINAHL (EBSCO), and ProQuest from inception until July 2018 and MEDLINE from inception until April 2019, without language restrictions, as well as bibliographies of relevant articles. Study Selection: Cohort studies with at least 1000 participants that reported an association between unprocessed red or processed meat intake and outcomes of interest. Data Extraction: Teams of 2 reviewers independently extracted data and assessed risk of bias. One investigator assessed certainty of evidence, and the senior investigator confirmed the assessments. Data Synthesis: Of 61 articles reporting on 55 cohorts with more than 4 million participants, none addressed quality of life or satisfaction with diet. Low-certainty evidence was found that a reduction in unprocessed red meat intake of 3 servings per week is associated with a very small reduction in risk for cardiovascular mortality, stroke, myocardial infarction (MI), and type 2 diabetes. Likewise, low-certainty evidence was found that a reduction in processed meat intake of 3 servings per week is associated with a very small decrease in risk for all-cause mortality, cardiovascular mortality, stroke, MI, and type 2 diabetes. Limitation: Inadequate adjustment for known confounders, residual confounding due to observational design, and recall bias associated with dietary measurement. Conclusion: The magnitude of association between red and processed meat consumption and all-cause mortality and adverse cardiometabolic outcomes is very small, and the evidence is of low certainty. Primary Funding Source: None. (PROSPERO: CRD42017074074).

Reduction of Red and Processed Meat Intake and Cancer Mortality and Incidence: A Systematic Review and Meta-analysis of Cohort Studies.

This article has been corrected. The original version (PDF) is appended to this article as a Supplement. Background: Cancer incidence has continuously increased over the past few centuries and represents a major health burden worldwide. Purpose: To evaluate the possible causal relationship between intake of red and processed meat and cancer mortality and incidence. Data Sources: Embase, Cochrane Central Register of Controlled Trials, Web of Science, CINAHL, and ProQuest from inception until July 2018 and MEDLINE from inception until April 2019 without language restrictions. Study Selection: Cohort studies that included more than 1000 adults and reported the association between consumption of unprocessed red and processed meat and cancer mortality and incidence. Data Extraction: Teams of 2 reviewers independently extracted data and assessed risk of bias; 1 reviewer evaluated the certainty of evidence, which was confirmed or revised by the senior reviewer. Data Synthesis: Of 118 articles (56 cohorts) with more than 6 million participants, 73 articles were eligible for the dose-response meta-analyses, 30 addressed cancer mortality, and 80 reported cancer incidence. Low-certainty evidence suggested that an intake reduction of 3 servings of unprocessed meat per week was associated with a very small reduction in overall cancer mortality over a lifetime. Evidence of low to very low certainty suggested that each intake reduction of 3 servings of processed meat per week was associated with very small decreases in overall cancer mortality over a lifetime; prostate cancer mortality; and incidence of esophageal, colorectal, and breast cancer. Limitation: Limited causal inferences due to residual confounding in observational studies, risk of bias due to limitations in diet assessment and adjustment for confounders, recall bias in dietary assessment, and insufficient data for planned subgroup analyses. Conclusion: The possible absolute effects of red and processed meat consumption on cancer mortality and incidence are very small, and the certainty of evidence is low to very low. Primary Funding Source: None. (PROSPERO: CRD42017074074).

Patterns of Red and Processed Meat Consumption and Risk for Cardiometabolic and Cancer Outcomes: A Systematic Review and Meta-analysis of Cohort Studies.

This article has been corrected. The original version (PDF) is appended to this article as a Supplement. Background: Studying dietary patterns may provide insights into the potential effects of red and processed meat on health outcomes. Purpose: To evaluate the effect of dietary patterns, including different amounts of red or processed meat, on all-cause mortality, cardiometabolic outcomes, and cancer incidence and mortality. Data Sources: Systematic search of MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials, CINAHL, Web of Science, and ProQuest Dissertations & Theses Global from inception to April 2019 with no restrictions on year or language. Study Selection: Teams of 2 reviewers independently screened search results and included prospective cohort studies with 1000 or more participants that reported on the association between dietary patterns and health outcomes. Data Extraction: Two reviewers independently extracted data, assessed risk of bias, and evaluated the certainty of evidence using GRADE (Grading of Recommendations Assessment, Development and Evaluation) criteria. Data Synthesis: Eligible studies that followed patients for 2 to 34 years revealed low- to very-low-certainty evidence that dietary patterns lower in red and processed meat intake result in very small or possibly small decreases in all-cause mortality, cancer mortality and incidence, cardiovascular mortality, nonfatal coronary heart disease, fatal and nonfatal myocardial infarction, and type 2 diabetes. For all-cause, cancer, and cardiovascular mortality and incidence of some types of cancer, the total sample included more than 400 000 patients; for other outcomes, total samples included 4000 to more than 300 000 patients. Limitation: Observational studies are prone to residual confounding, and these studies provide low- or very-low-certainty evidence according to the GRADE criteria. Conclusion: Low- or very-low-certainty evidence suggests that dietary patterns with less red and processed meat intake may result in very small reductions in adverse cardiometabolic and cancer outcomes. Primary Funding Source: None. (PROSPERO: CRD42017074074).