Bimodal distribution of atrial fibrillation burden in 3 distinct cohorts: What is 'paroxysmal' atrial fibrillation?

BACKGROUND: Burden of atrial fibrillation (AF), as a continuous measure, is an emerging alternative classification often assumed to increase linearly with progression of disease. Yet there are no descriptions of AF burden distributions across populations. METHODS: We examined patterns of AF burden (% time in AF) across 3 different cohorts: outpatients with AF undergoing Holter monitoring in a national registry (ORBIT-AF II), routine outpatients undergoing Holter monitoring in a tertiary healthcare system (UHealth), and patients >= 65 years with cardiac implantable electronic devices (Merlin.netTM linked to Medicare). RESULTS: We included 2,058 ORBIT-AF II patients, 4,537 UHealth patients, and 39,710 from Merlin.net. Mean age ranged from 56 to 77 years, sex ranged from 40% to 61% male, and mean CHA2DS2-VASc scores ranged from 2.2 to 4.9. Across all cohorts, AF burden demonstrated skewed frequency towards the extremes, with the vast majority of patients having either very low or very high AF burden. This bimodal distribution was consistent across cohorts, across clinically-documented AF types (paroxysmal v persistent), patients with or without a known AF diagnosis, and among patients with different types of cardiac implantable electronic devices. CONCLUSIONS: Across 3 broad, diverse cohorts with continuous monitoring, distribution of AF burden was consistently skewed towards the extremes without an even, linear distribution or progression. As AF burden is increasingly recognized as a descriptor and potential risk-stratifier, these findings have important implications for future research and patient care.

Utilizing Artificial Intelligence to Enhance Health Equity Among Patients with Heart Failure.

Patients with heart failure (HF) are heterogeneous with various intrapersonal and interpersonal characteristics contributing to clinical outcomes. Bias, structural racism, and social determinants of health have been implicated in unequal treatment of patients with HF. Through several methodologies, artificial intelligence (AI) can provide models in HF prediction, prognostication, and provision of care, which may help prevent unequal outcomes. This review highlights AI as a strategy to address racial inequalities in HF; discusses key AI definitions within a health equity context; describes the current uses of AI in HF, strengths and harms in using AI; and offers recommendations for future directions.

The Effects of Using a Clinical Prediction Rule to Prioritize Diagnostic Testing on Transmission and Hospital Burden: A Modeling Example of Early Severe Acute Respiratory Syndrome Coronavirus 2.

BACKGROUND: Prompt identification of infections is critical for slowing the spread of infectious diseases. However, diagnostic testing shortages are common in emerging diseases, low resource settings, and during outbreaks. This forces difficult decisions regarding who receives a test, often without knowing the implications of those decisions on population-level transmission dynamics. Clinical prediction rules (CPRs) are commonly used tools to guide clinical decisions. METHODS: Using early severe acute respiratory syndrome coronavirus disease 2 (SARS-CoV-2) as an example, we used data from electronic health records to develop a parsimonious 5-variable CPR to identify those who are most likely to test positive. To consider the implications of gains in daily case detection at the population level, we incorporated testing using the CPR into a compartmentalized model of SARS-CoV-2. RESULTS: We found that applying this CPR (area under the curve, 0.69; 95% confidence interval, .68-.70) to prioritize testing increased the proportion of those testing positive in settings of limited testing capacity. We found that prioritized testing led to a delayed and lowered infection peak (ie, "flattens the curve"), with the greatest impact at lower values of the effective reproductive number (such as with concurrent community mitigation efforts), and when higher proportions of infectious persons seek testing. In addition, prioritized testing resulted in reductions in overall infections as well as hospital and intensive care unit burden. CONCLUSION: We highlight the population-level benefits of evidence-based allocation of limited diagnostic capacity.SummaryWhen the demand for diagnostic tests exceeds capacity, the use of a clinical prediction rule to prioritize diagnostic testing can have meaningful impact on population-level outcomes, including delaying and lowering the infection peak, and reducing healthcare burden.

Prevalence of control groups in cardiovascular clinical trials: An analysis of ClinicalTrials.gov from 2009 through 2019.

Clinical trials provide the foundational evidence that guide many patient-facing decisions; however, the therapeutic effect and safety of an intervention is best evaluated when compared to a control group. We used ClinicalTrials.gov to describe the proportion of registered Phase III and IV cardiovascular clinical trials that contain a control group from 2009 through 2019. Of 1,677 registered Phase III and IV cardiovascular clinical trials, 81.2% contain a control group, and the annual prevalence remained unchanged between 2009 and 2019.

Adaptation of an NLP system to a new healthcare environment to identify social determinants of health.

Social determinants of health (SDoH) are increasingly important factors for population health, healthcare outcomes, and care delivery. However, many of these factors are not reliably captured within structured electronic health record (EHR) data. In this work, we evaluated and adapted a previously published NLP tool to include additional social risk factors for deployment at Vanderbilt University Medical Center in an Acute Myocardial Infarction cohort. We developed a transformation of the SDoH outputs of the tool into the OMOP common data model (CDM) for re-use across many potential use cases, yielding performance measures across 8 SDoH classes of precision 0.83 recall 0.74 and F-measure of 0.78.

Shedding Light on the Black Box: Explaining Deep Neural Network Prediction of Clinical Outcomes.

Deep neural network models are emerging as an important method in healthcare delivery, following the recent success in other domains such as image recognition. Due to the multiple non-linear inner transformations, deep neural networks are viewed by many as black boxes. For practical use, deep learning models require explanations that are intuitive to clinicians. In this study, we developed a deep neural network model to predict outcomes following major cardiovascular procedures, using temporal image representation of past medical history as input. We created a novel explanation for the prediction of the model by defining impact scores that associate clinical observations with the outcome. For comparison, a logistic regression model was fitted to the same dataset. We compared the impact scores and log odds ratios by calculating three types of correlations, which provided a partial validation of the impact scores. The deep neural network model achieved an area under the receiver operating characteristics curve (AUC) of 0.787, compared to 0.746 for the logistic regression model. Moderate correlations were found between the impact scores and the log odds ratios. Impact scores generated by the explanation algorithm has the potential to shed light on the "black box" deep neural network model and could facilitate its adoption by clinicians.

Systematic collection of patient-reported outcomes in atrial fibrillation: feasibility and initial results of the Utah mEVAL AF programme.

AIMS: Incorporating patient-reported outcomes (PROs) into routine care of atrial fibrillation (AF) enables direct integration of symptoms, function, and health-related quality of life (HRQoL) into practice. We report our initial experience with a system-wide PRO initiative among AF patients. METHODS AND RESULTS: All patients with AF in our practice undergo PRO assessment with the Toronto AF Severity Scale (AFSS), and generic PROs, prior to electrophysiology clinic visits. We describe the implementation, feasibility, and results of clinic-based, electronic AF PRO collection, and compare AF-specific and generic HRQoL assessments. From October 2016 to February 2019, 1586 unique AF patients initiated 2379 PRO assessments, 2145 of which had all PRO measures completed (90%). The median completion time for all PRO measures per visit was 7.3 min (1st, 3rd quartiles: 6, 10). Overall, 38% of patients were female (n = 589), mean age was 68 (SD 12) years, and mean CHA2DS2-VASc score was 3.8 (SD 2.0). The mean AFSS symptom score was 8.6 (SD 6.6, 1st, 3rd quartiles: 3, 13), and the full range of values was observed (0, 35). Generic PROs of physical function, general health, and depression were impacted at the most severe quartiles of AF symptom score (P < 0.0001 for each vs. AFSS quartile). CONCLUSION: Routine clinic-based, PRO collection for AF is feasible in clinical practice and patient time investment was acceptable. Disease-specific AF PROs add value to generic HRQoL instruments. Further research into the relationship between PROs, heart rhythm, and AF burden, as well as PRO-guided management, is necessary to optimize PRO utilization.

Hospital Variation in Adherence Rates to Secondary Prevention Medications and the Implications on Quality.

BACKGROUND: Medication adherence is important to improve the long-term outcomes after acute myocardial infarction (MI). We hypothesized that there is significant variation among US hospitals in terms of medication adherence after MI, and that patients treated at hospitals with higher medication adherence after MI will have better long-term cardiovascular outcomes. METHODS: We identified 19 704 Medicare patients discharged after acute MI from 347 US hospitals participating in the ACTION Registry-GWTG (Acute Coronary Treatment and Intervention Outcomes Network Registry-Get With the Guidelines) from January 2, 2007, to October 1, 2010. Using linked Medicare Part D prescription filling data, medication adherence was defined as proportion of days covered >80% within 90 days after discharge. Cox proportional hazards modeling was used to compare 2-year major adverse cardiovascular events among hospitals with high, moderate, and low 90-day medication adherence. RESULTS: By 90 days after MI, overall rates of adherence to medications prescribed at discharge were 68% for ß-blockers, 63% for statins, 64% for angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, and 72% for thienopyridines. Adherence to these medications up to 90 days varied significantly among hospitals: ß-blockers (proportion of days covered >80%; 59% to 75%), statins (55% to 69%), thienopyridines (64% to 77%), and angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (57% to 69%). Compared with hospitals in the lowest quartile of 90-day composite medication adherence, hospitals with the highest adherence had lower unadjusted and adjusted 2-year major adverse cardiovascular event risk (27.5% versus 35.3%; adjusted hazard ratio, 0.88; 95% confidence interval, 0.80-0.96). High-adherence hospitals also had lower adjusted rates of death or readmission (hazard ratio, 0.90; 95% confidence interval, 0.85-0.96), whereas there was no difference in mortality after adjustment. CONCLUSIONS: Use of secondary prevention medications after discharge varies significantly among US hospitals and is inversely associated with 2-year outcomes. Hospitals may improve medication adherence after discharge and patient outcomes through better coordination of care between inpatient and outpatient settings.

The association of left ventricular ejection fraction with clinical outcomes after myocardial infarction: Findings from the Acute Coronary Treatment and Intervention Outcomes Network (ACTION) Registry-Get With the Guidelines (GWTG) Medicare-linked database.

BACKGROUND: Little is known about the relationship between ejection fraction (EF) and clinical outcomes among older patients with myocardial infarction in contemporary clinical practice. METHODS: Data on 82,558 patients 65 years or older with ST-elevation myocardial infarction or non-ST-elevation myocardial infarction who survived to hospital discharge in the ACTION Registry-GWTG (2007-2011) were linked to Medicare data. Multivariable Cox proportional hazard modeling was used to assess the association between EF reported during hospitalization and 1-year mortality, using EF as a categorical variable (≤35%, >35% and ≤45%, >45% and <55%, and ≥55%) and as a continuous variable. Secondary outcomes of interest were 1-year all-cause, cardiovascular, and heart failure readmissions. RESULTS: The risk of 1-year mortality was 29.0% in patients with EF ≤ 35%, compared with 13.0% in patients in the reference group, EF ≥ 55% (adjusted hazard ratio [HR] 1.58, 95% CI 1.51-1.66). Relative to patients with EF ≥ 55%, patients with EF ≤ 35% had an increased risk of 1-year all-cause readmission (adjusted HR 1.20, 95% CI 1.17-1.24), cardiovascular readmission (adjusted HR 1.36, 95% CI 1.31-1.41), and heart failure readmission (adjusted HR 2.43, 95% CI 2.28-2.60). For patients with EF ≤ 40%, the hazard of mortality increased by 26% for every 5% decrease in EF, a finding that remained after risk adjustment (adjusted HR 1.11, 95% CI 1.09-1.12). CONCLUSIONS: Low EF after MI remains an important risk factor for postdischarge mortality and hospital readmission, even after adjustment for patient and hospital characteristics.

Changes in Oral Anticoagulant Treatment Rates in Atrial Fibrillation before and after the Introduction of Direct Oral Anticoagulants.

BACKGROUND: Direct oral anticoagulants (DOACs) have the potential to improve stroke prevention among atrial fibrillation (AF) patients. We sought to determine if oral anticoagulation (OAC) treatment rates have increased since the approval of DOACs. METHODS: We identified 6,688 patients with AF at an academic medical center from January 2008 to June 2015. We examined OAC prescription rates over time and according to CHA2DS2VASc score using multivariable Poisson regression models, with an interaction term between risk score and year of AF diagnosis. RESULTS: Among 6,688 AF patients, 78% had CHA2DS2VASc scores ≥2, 51.6% of whom received an OAC prescription within 90 days of diagnosis. The OAC prescription rate was 47.8% in the pre-DOAC era and peaked at 56.4% in 2014. Relative to the pre-DOAC era, prescription rates increased in 2012 and leveled off thereafter. The prescription rate for the highest risk group was 58.5%, compared with 45.0% in patients with a CHA2DS2VASc score of 2 (p < 0.01). In the adjusted analysis, prescription rates were higher for the higher risk group (adjusted relative risk 1.24 for CHA2DS2VASc score 7-9 vs. 2, 95% CI 1.09-1.40). CONCLUSIONS: OAC treatment rates have increased since DOAC introduction, but substantial treatment gaps remain, specifically among the higher risk patients.

Cost-effectiveness of genotype-guided and dual antiplatelet therapies in acute coronary syndrome.

BACKGROUND: The choice of antiplatelet therapy after acute coronary syndrome (ACS) is complicated: Ticagrelor and prasugrel are novel alternatives to clopidogrel, patients with some genotypes may not respond to clopidogrel, and low-cost generic formulations of clopidogrel are available. OBJECTIVE: To determine the most cost-effective strategy for dual antiplatelet therapy after percutaneous coronary intervention for ACS. DESIGN: Decision-analytic model. DATA SOURCES: Published literature, Medicare claims, and life tables. TARGET POPULATION: Patients having percutaneous coronary intervention for ACS. TIME HORIZON: Lifetime. PERSPECTIVE: Societal. INTERVENTION: Five strategies were examined: generic clopidogrel, prasugrel, ticagrelor, and genotyping for polymorphisms of CYP2C19 with carriers of loss-of-function alleles receiving either ticagrelor (genotyping with ticagrelor) or prasugrel (genotyping with prasugrel) and noncarriers receiving clopidogrel. OUTCOME MEASURES: Direct medical costs, quality-adjusted life years(QALYs), and incremental cost-effectiveness ratios (ICERs). RESULTS OF BASE-CASE ANALYSIS: The clopidogrel strategy produced$179 301 in costs and 9.428 QALYs. Genotyping with prasugrel was superior to prasugrel alone, with an ICER of $35 800 per QALY relative to clopidogrel. Genotyping with ticagrelor was more effective than genotyping with prasugrel ($30 200 per QALY relative to clopidogrel). Ticagrelor was the most effective strategy($52 600 per QALY relative to genotyping with ticagrelor). RESULTS OF SENSITIVITY ANALYSIS: Stronger associations between genotype and thrombotic outcomes rendered ticagrelor substantially less cost-effective ($104 800 per QALY). Genotyping with prasugrel was the preferred therapy among patients who could not tolerate ticagrelor. LIMITATION: No randomized trials have directly compared genotyping strategies or prasugrel with ticagrelor. CONCLUSION: Genotype-guided personalization may improve the cost-effectiveness of prasugrel and ticagrelor after percutaneous coronary intervention for ACS, but ticagrelor for all patients may bean economically reasonable alternative in some settings.

Clinical Trial Participation and COVID-19: a Descriptive Analysis from the American Heart Association's Get With The Guidelines Registry.

As COVID-19 cases begin to decrease in the USA, learning from the pandemic experience will provide insights regarding disparities of care delivery. We sought to determine if specific populations hospitalized with COVID-19 are equally likely to be enrolled in clinical trials. We examined patients hospitalized with COVID-19 at centers participating in the American Heart Association's COVID-19 CVD Registry. The primary outcome was odds of enrollment in a clinical trial, according to sex, race, and ethnicity. Among 14,397 adults hospitalized with COVID-19, 9.5% (n = 1,377) were enrolled in a clinical trial. The proportion of enrolled patients was the lowest for Black patients (8%); in multivariable analysis, female and Black patients were less likely to be enrolled in a clinical trial related to COVID-19 compared to men and other racial groups, respectively. Determination of specific reasons for the disparities in trial participation related to COVID-19 in these populations should be further investigated.

Relationships between 2018 UNOS heart policy and transplant outcomes in metropolitan, micropolitan, and rural settings.

BACKGROUND: In 2018, United Network for Organ Sharing (UNOS) extended the radius for which a heart transplant candidate can match with a donor, and outcomes across population densities are unknown. We sought to determine whether the policy change was associated with differences in heart transplant waitlist time or death post-transplant for patients from rural, micropolitan, and metropolitan settings. METHODS: Using the Scientific Registry of Transplant Recipients, we evaluated U.S. adult patients listed for heart transplant from Janurary 2017 to September 2019 with follow-up through March 2020. Patients were stratified by home zip-codes to either metropolitan, micropolitan, or rural settings. Fine and Gray and Cox models were respectively used to estimate Sub-distribution hazard ratios (SHR) of heart transplant with death or removal from transplant list as a competing event, and HR of death post-transplant within population densities after versus before the UNOS policy change date, October 18, 2018. Analyses were adjusted for demographics, comorbidities, and labs. RESULTS: Among 8,747 patients listed for heart transplant, 84.7% were from metropolitan, 8.6% micropolitan, and 6.6% rural settings. The 2018 UNOS policy was associated with earlier receipt of heart transplant for metropolitan [SHR 1.56 (95% CI: 1.46-1.66)] and micropolitan [SHR 1.48 (95% CI: 1.21-1.82)] populations, but not significantly for rural [SHR 1.20 (95% CI: 0.93-1.54)]; however, the interaction between policy and densities was not significant (p = .14). Policy changes were not associated with risk of death post-transplant [metropolitan: HR 1.04 (95% CI: 0.80-1.34); micropolitan: HR 1.10 (95% CI: 0.55-2.23); rural: HR 1.04 (95% CI: 0.52-2.08); interaction p = .99]. CONCLUSIONS: The 2018 UNOS heart transplant policy was associated with earlier receipt of heart transplant and no difference in post-transplant survival within population densities. Additional follow-up is needed to determine whether improvements are sustained.

Information Extraction From Electronic Health Records to Predict Readmission Following Acute Myocardial Infarction: Does Natural Language Processing Using Clinical Notes Improve Prediction of Readmission?

Background Social risk factors influence rehospitalization rates yet are challenging to incorporate into prediction models. Integration of social risk factors using natural language processing (NLP) and machine learning could improve risk prediction of 30-day readmission following an acute myocardial infarction. Methods and Results Patients were enrolled into derivation and validation cohorts. The derivation cohort included inpatient discharges from Vanderbilt University Medical Center between January 1, 2007, and December 31, 2016, with a primary diagnosis of acute myocardial infarction, who were discharged alive, and not transferred from another facility. The validation cohort included patients from Dartmouth-Hitchcock Health Center between April 2, 2011, and December 31, 2016, meeting the same eligibility criteria described above. Data from both sites were linked to Centers for Medicare & Medicaid Services administrative data to supplement 30-day hospital readmissions. Clinical notes from each cohort were extracted, and an NLP model was deployed, counting mentions of 7 social risk factors. Five machine learning models were run using clinical and NLP-derived variables. Model discrimination and calibration were assessed, and receiver operating characteristic comparison analyses were performed. The 30-day rehospitalization rates among the derivation (n=6165) and validation (n=4024) cohorts were 15.1% (n=934) and 10.2% (n=412), respectively. The derivation models demonstrated no statistical improvement in model performance with the addition of the selected NLP-derived social risk factors. Conclusions Social risk factors extracted using NLP did not significantly improve 30-day readmission prediction among hospitalized patients with acute myocardial infarction. Alternative methods are needed to capture social risk factors.

An explainable artificial intelligence approach for predicting cardiovascular outcomes using electronic health records.

Understanding the conditionally-dependent clinical variables that drive cardiovascular health outcomes is a major challenge for precision medicine. Here, we deploy a recently developed massively scalable comorbidity discovery method called Poisson Binomial based Comorbidity discovery (PBC), to analyze Electronic Health Records (EHRs) from the University of Utah and Primary Children's Hospital (over 1.6 million patients and 77 million visits) for comorbid diagnoses, procedures, and medications. Using explainable Artificial Intelligence (AI) methodologies, we then tease apart the intertwined, conditionally-dependent impacts of comorbid conditions and demography upon cardiovascular health, focusing on the key areas of heart transplant, sinoatrial node dysfunction and various forms of congenital heart disease. The resulting multimorbidity networks make possible wide-ranging explorations of the comorbid and demographic landscapes surrounding these cardiovascular outcomes, and can be distributed as web-based tools for further community-based outcomes research. The ability to transform enormous collections of EHRs into compact, portable tools devoid of Protected Health Information solves many of the legal, technological, and data-scientific challenges associated with large-scale EHR analyses.

Maternal cardiovascular complications at the time of delivery and subsequent re-hospitalization in the USA, 2010-16.

AIMS: Cardiovascular (CV) complications are the leading cause of maternal morbidity and mortality. The objective was to estimate trends in the incidence of peripartum CV complications in the USA between 2010 and 2016. METHODS AND RESULTS: This was a retrospective analyses using data from the Healthcare Cost and Utilization Project. We included women with delivery codes consistent with delivery, weighted to a national estimate. The primary outcome was the age-adjusted incidence of CV complications among all deliveries, including complications that occurred during re-hospitalizations. Complications were identified using International Classification of Diseases (ICD) codes. Joinpoint regression was used to evaluate time trends and complications were stratified by type. The secondary outcome was in-hospital maternal death among women with a CV complication. We identified a weighted estimate of 27 408 652 women hospitalized for delivery from 2010 to 2016. Including all years, the complication incidence was 7.36/1000 births [95% confidence interval (CI) 7.18-7.54], with an estimated annual percentage change of 5.8% (95% CI 3.7-7.8%). Cardiac dysrhythmia was the most common complication [3.98/1000 births (95% CI 3.88-4.08)] and acute myocardial infarction was the least common complication [0.11/1000 births (95% CI 0.10-0.11)]. The incidence of hypertension, acute myocardial infarction, and cardiac arrest increased over time, the incidence of congestive heart failure and acute cerebrovascular disease remained stable, the incidence of pulmonary heart disease increased from 2015 onward, and the incidence of cardiac dysrhythmia decreased in 2016. Complications during re-hospitalization accounted for 13.6% (95% CI 13.2-14.1%) of all complications and was highest for acute myocardial infarction [28.1% (95% CI 23.2-33.1)]. Among women with any complication, the mortality rate was 1.20 (95% CI 1.11-1.29) per 100 complications. CONCLUSION: Our analyses suggest the rate of peripartum CV complications are increasing in the USA, which highlights the need for active efforts in research and prevention.

Regional Systems of Care in ST Elevation Myocardial Infarction.

ST-segment elevation myocardial infarction is a medical emergency with significant health care delivery challenges to ensure rapid triage and treatment. Several developments over the past decades have led to improved care delivery, decreased time to reperfusion, and decreased mortality. Still, significant challenges remain to further optimize the delivery of care for this patient population.