Masoudi, Frederick A.

Observational medical data present unique opportunities for analysis of medical outcomes and treatment decision making. However, because these datasets do not contain the strict pairing of randomized control trials, matching techniques are to draw comparisons among patients. A key limitation to such techniques is verification that the variables used to model treatment decision making are also relevant in identifying the risk of major adverse events. This article explores a deep mixture of experts approach to jointly learn how to match patients and model the risk of major adverse events in patients. Although trained with information regarding treatment and outcomes, after training, the proposed model is decomposable into a network that clusters patients into phenotypes from information available before treatment. This model is validated on a dataset of patients with acute myocardial infarction complicated by cardiogenic shock. The mixture of experts approach can predict the outcome of mortality with an area under the receiver operating characteristic curve of 0.85 ± 0.01 while jointly discovering five potential phenotypes of interest. The technique and interpretation allow for identifying clinically relevant phenotypes that may be used both for outcomes modeling as well as potentially evaluating individualized treatment effects.

Background: The Food and Drug Administration approved left atrial appendage occlusion with the Watchman device for patients who are at increased stroke risk and are suitable for oral anticoagulation but who have an appropriate reason to seek a nondrug alternative. These broad criteria raise the question of their interpretation in clinical practice. There is a lack of studies comprehensively evaluating the indications for Watchman implantation among a large series of patients from contemporary, real-world practice in the United States. Methods: We used the National Cardiovascular Data Registry Left Atrial Appendage Occlusion Registry to identify Watchman procedures performed between 2016 and 2018. We assessed procedural indications for Watchman implantation in the United States and evaluated the association between procedural indications and in-hospital adverse events. Results: A total of 38 314 procedures were included. The mean patient age was 76.1±8.1 years, and 58.9% were men. The mean CHA 2 DS 2 -VASc score was 4.8±1.5, whereas the mean hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio, elderly, drugs/alcohol (HAS-BLED) score was 3.0±1.1. Prior stroke or transient ischemic attack was reported in 40.2% and prior bleeding in 70.1%, with gastrointestinal bleeding being most common (41.9%). The most common site-reported procedural indications for Watchman implantation were increased thromboembolic risk (64.8%) and history of major bleed (64.3%), followed by high fall risk (35.5%). Most (71.9%) had ≥2 procedural indications. Patients with high fall risk had increased risk of in-hospital adverse events (adjusted OR, 1.12; P =0.025), but no other differences were found in the risk of in-hospital adverse events by procedural indication. Conclusions: Among patients in the National Cardiovascular Data Registry Left Atrial Appendage Occlusion Registry, the most common procedural indications for Watchman implantation were increased thromboembolic risk, history of major bleed, and high fall risk. A majority of patients had multiple procedural indications. High fall risk conferred a modestly increased risk of in-hospital adverse events.

Introduction: Primary prevention implantable cardioverter defibrillators (ICDs) reduce mortality in select patients with left ventricular systolic dysfunction (LVSD). The occurrence of device therapies after ICD implantation in contemporary clinical practice is not well described, especially in patient subgroups designated in the Centers for Medicare and Medicaid Services (CMS) 2005 Coverage with Evidence Development (CED). Methods: The Longitudinal Study of ICDs assessed rates and correlates of device therapies (overall and those requiring shock) up to 3 years post-implant among 2540 patients with LVSD receiving first-time primary prevention ICDs in 7 US health care systems from 2006-2009. Implant data from the National Cardiovascular Data Registry ICD Registry were linked to electronic longitudinal health care system data and a novel centrally adjudicated repository of device therapies abstracted from medical records. Proportional hazard models evaluated associations with device therapies by appropriateness, adjusted for demographic and clinical factors. Results: Subjects were 26% women, 35% <65 years old, and 59% non-Hispanic white. Over a mean 26 months, 738 (29.1%) received at least 1 therapy (median 2). Estimated 3-year risk of any device therapy was 36% (24% appropriate, 12% inappropriate); for therapy requiring shock, corresponding values were 24%, 14%, 9%. The rate of appropriate therapy was higher in men than women (adjusted hazard ratio 1.84, 95% confidence interval 1.43-2.35). The rate of inappropriate therapy was higher for subjects <65 than for those >=65 years (1.40, 1.04-1.88), and lower among 2009 implants compared to 2006 (0.66, 0.46-0.95). Regarding the 3 CMS CED patient subgroups, neither ejection fraction nor heart failure symptom severity was associated with device therapy; for LVSD etiology, the rate of inappropriate therapy was nominally but not significantly higher among subjects with non-ischemic cardiomyopathy of < 9 months duration (1.38, 0.88-2.18). Conclusions: In a representative cohort of primary prevention ICD patients in usual care, rates of device therapies were somewhat lower than in landmark trials, differed by some patient characteristics, but did not differ meaningfully for CMS CED subgroups.

Objective: To characterize changes in rates of hospitalization for digoxin toxicity and trends in the associated mortality and readmission among older adults over a 12-year period in the United States. Methods: We studied 33,952,331 Medicare fee-for-service beneficiaries 65 years or older with a hospital discharge diagnosis of digoxin toxicity in the United States from 1999 to 2011. Outcome measures were rates of hospitalization for digoxin toxicity; in-hospital mortality; 30-day mortality; and 30-day readmission. Results: There were 20,957 hospitalizations for a principal or secondary diagnosis of digoxin toxicity between 1999 and 2011. The rate declined significantly from 15.2 per 100,000 person-years (95% confidence interval [CI]: 14.7-15.7) in 1999 to 2.1 per 100,000 person-years (95% CI: 1.9-2.3) in 2011 (p<0.001), representing an adjusted annual decline of 17.0% (95% CI: 16.2-17.0) (Figure 1). Between 1999 and 2011, the observed in-hospital and 30-day mortality rates associated with hospitalization for digoxin toxicity declined significantly, from 6.0% (95% CI: 5.2-6.8) to 3.3% (95% CI: 2.0-5.1) (p<0.01) and 14.0% (95% CI: 13.0-15.2) to 10.6% (95% CI: 8.2-13.4) (p<0.05), respectively, representing an annual decline for in-hospital mortality of 5.0% (95% CI: 3.7-7.2) and for 30-day mortality of 4.0% (95% CI: 3.1-5.7). The overall observed 30-day readmission rate declined significantly from 23.5% (95% CI: 22.1-24.9) in 1999 to 18.9% (95% CI: 15.6-22.3) in 2011 (p<0.05), but there was no significant decline in the adjusted annual change in 30-day readmission (1.0%, 95% CI: 0.0-1.7). Conclusions: In a national sample of Medicare beneficiaries, the rate of hospitalization for digoxin toxicity and subsequent mortality declined significantly between 1999 and 2011.

Introduction: The American College of Cardiology developed the National Cardiovascular Data Registry (NCDR) AFib Ablation Registry to assess prevalence, demographics, procedural management and outcomes of patients undergoing atrial fibrillation (AFib) ablation procedures. The objective of this registry’s first presentation is to summarize initial AFib Ablation Registry data regarding patient characteristics, procedural characteristics, and in-hospital adverse events and outcomes in those undergoing AFib ablation enrolled in the registry. Methods: Data were tabulated and summarized with simple proportions for categorical data, and mean ± standard deviation for continuous data. Results: A total of 31,345 patients who underwent AFib ablation were included in the registry (mean age 65.0 ± 10.4, 65.2% male, 57.6% paroxysmal AFib), with ablations performed at 120 hospitals, by 501 physicians. The mean CHA2DS2-VASc score of was 2.6 ± 1.6, and mean HAS-BLED score was 1.1 ± 0.9. Successful pulmonary vein isolation was able to be achieved in all veins in 94.5% of patients. The prevalence of any complication during procedural admission was 3.1%; the most common were hematoma (0.5%) and pericardial effusion requiring intervention (0.42%). The mean length of stay during the index admission (days) was 1.4 ± 7.6, and in-hospital death prevalence was 0.04% [FIGURE] . Conclusions: In conclusion, the NCDR AFib Ablation Registry is a large, national, multi-center, prospective cohort study to describe events outside the clinical trial or research setting, demonstrating successful pulmonary vein isolation in the majority of patients and a low prevalence of adverse events. Future studies from the registry will be able to help support and develop evidence-based guidelines for AFib ablation treatment and optimize patient outcomes.

OBJECTIVE To determine whether burden of multiple chronic conditions (MCCs) influences the risk of receiving inappropriate vs appropriate device therapies. DESIGN Retrospective cohort study. SETTING Seven US healthcare delivery systems. PARTICIPANTS Adults with left ventricular systolic dysfunction receiving an implantable cardioverter‐defibrillator (ICD) for primary prevention. MEASUREMENTS Data on 24 comorbid conditions were captured from electronic health records and categorized into quartiles of comorbidity burden (0‐3, 4‐5, 6‐7 and 8‐16). Incidence of ICD therapies (shock and antitachycardia pacing [ATP] therapies), including appropriateness, was collected for 3 years after implantation. Outcomes included time to first ICD therapy, total ICD therapy burden, and risk of inappropriate vs appropriate ICD therapy. RESULTS Among 2235 patients (mean age = 69 ± 11 years, 75% men), the median number of comorbidities was 6 (interquartile range = 4‐8), with 98% having at least two comorbidities. During a mean 2.2 years of follow‐up, 18.3% of patients experienced at least one appropriate therapy and 9.9% experienced at least one inappropriate therapy. Higher comorbidity burden was associated with an increased risk of first inappropriate therapy (adjusted hazard ratio [HR] = 1.94 [95% confidence interval {CI} = 1.14‐3.31] for 4‐5 comorbidities; HR = 2.25 [95% CI = 1.25‐4.05] for 6‐7 comorbidities; and HR = 2.91 [95% CI = 1.54‐5.50] for 8‐16 comorbidities). Participants with 8‐16 comorbidities had a higher total burden of ICD therapy (adjusted relative risk [RR] = 2.12 [95% CI = 1.43‐3.16]), a higher burden of inappropriate therapy (RR = 3.39 [95% CI = 1.67‐6.86]), and a higher risk of receiving inappropriate vs appropriate therapy (RR = 1.74 [95% CI = 1.07‐2.82]). Comorbidity burden was not significantly associated with receipt of appropriate ICD therapies. Patterns were similar when separately examining shock or ATP therapies. CONCLUSIONS In primary prevention ICD recipients, MCC burden was independently associated with an increased risk of inappropriate but not appropriate device therapies. Comorbidity burden should be considered when engaging patients in shared decision making about ICD implantation.

Background Primary prevention implantable cardioverter‐defibrillators (ICDs) reduce mortality in selected patients with left ventricular systolic dysfunction by delivering therapies (antitachycardia pacing or shocks) to terminate potentially lethal arrhythmias; inappropriate therapies also occur. We assessed device therapies among adults receiving primary prevention ICDs in 7 healthcare systems. Methods and Results We linked medical record data, adjudicated device therapies, and the National Cardiovascular Data Registry ICD Registry. Survival analysis evaluated therapy probability and predictors after ICD implant from 2006 to 2009, with attention to Centers for Medicare and Medicaid Services Coverage With Evidence Development subgroups: left ventricular ejection fraction, 31% to 35%; nonischemic cardiomyopathy <9 months’ duration; and New York Heart Association class IV heart failure with cardiac resynchronization therapy defibrillator. Among 2540 patients, 35% were <65 years old, 26% were women, and 59% were white. During 27 (median) months, 738 (29%) received ≥1 therapy. Three‐year therapy risk was 36% (appropriate, 24%; inappropriate, 12%). Appropriate therapy was more common in men (adjusted hazard ratio [HR], 1.84; 95% confidence interval [CI], 1.43–2.35). Inappropriate therapy was more common in patients with atrial fibrillation (adjusted HR, 2.20; 95% CI, 1.68–2.87), but less common among patients ≥65 years old versus younger (adjusted HR, 0.72; 95% CI, 0.54–0.95) and in recent implants (eg, in 2009 versus 2006; adjusted HR, 0.66; 95% CI, 0.46–0.95). In Centers for Medicare and Medicaid Services Coverage With Evidence Development analysis, inappropriate therapy was less common with cardiac resynchronization therapy defibrillator versus single chamber (adjusted HR, 0.55; 95% CI, 0.36–0.84); therapy risk did not otherwise differ for Centers for Medicare and Medicaid Services Coverage With Evidence Development subgroups. Conclusions In this community cohort of primary prevention patients receiving ICD, therapy delivery varied across demographic and clinical characteristics, but did not differ meaningfully for Centers for Medicare and Medicaid Services Coverage With Evidence Development subgroups.