Tree-based machine learning models such as random forests, decision trees, and gradient boosted trees are popular non-linear predictive models, yet comparatively little attention has been paid to explaining their predictions. Here, we improve the interpretability of tree-based models through three main contributions: 1) The first polynomial time algorithm to compute optimal explanations based on game theory. 2) A new type of explanation that directly measures local feature interaction effects. 3) A new set of tools for understanding global model structure based on combining many local explanations of each prediction. We apply these tools to three medical machine learning problems and show how combining many high-quality local explanations allows us to represent global structure while retaining local faithfulness to the original model. These tools enable us to i) identify high magnitude but low frequency non-linear mortality risk factors in the US population, ii) highlight distinct population subgroups with shared risk characteristics, iii) identify non-linear interaction effects among risk factors for chronic kidney disease, and iv) monitor a machine learning model deployed in a hospital by identifying which features are degrading the model's performance over time. Given the popularity of tree-based machine learning models, these improvements to their interpretability have implications across a broad set of domains.

overestimated. Irrespectively, it is important to point out that there is little information with regards to real-life battery longevity for the studied CRT-D device (Quadra Assura MP, St Jude) given its fairly recent commercialization. Nonetheless, we believe the main value of the present study lies in the reported proportional differences in battery longevity between the different pacing programming protocols. Our results may therefore help clinicians make more informed decisions when considering MPP activation, given the current scarcity of information regarding its impact on battery longevity. ConclusionsIn most cases, MPP activation significantly reduces battery longevity compared with that for conventional CRT configuration. However, when reasonable MPP LV vector PCTs (< _4.0 V) are achieved, the decrease in battery longevity is relatively small and this may prompt the clinician to activate MPP in such scenarios.

Background The incidence and etiology of sudden cardiac death (SCD) in athletes is debated with hypertrophic cardiomyopathy (HCM) often reported as the most common etiology. Methods and Results A database of all NCAA deaths (2003 – 2013) was developed. Additional information and autopsy reports were obtained when possible. Cause of death was adjudicated by an expert panel. There were 4,242,519 athlete-years (AY) and 514 total student athlete deaths. Accidents were the most common cause of death (257, 50%, 1:16,508 AY) followed by medical causes (147, 29%, 1:28,861 AY). The most common medical cause of death was SCD (79, 15%, 1:53,703 AY). Males were at higher risk than females 1:37,790 AY vs. 1:121,593 AY (IRR 3.2, 95% CI, 1.9-5.5, p < .00001), and black athletes were at higher risk than white athletes 1:21,491 AY vs. 1:68,354 AY (IRR 3.2, 95% CI, 1.9-5.2, p < .00001). The incidence of SCD in Division 1 male basketball athletes was 1:5,200 AY. The most common findings at autopsy were autopsy negative sudden unexplained death (AN-SUD) in 16 (25%) and definitive evidence for HCM was seen in 5 (8%). Media reports identified more deaths in higher divisions (87%, 61%, and 44%) while percentages from the internal database did not vary (87%, 83%, and 89%). Insurance claims identified only 11% of SCDs. Conclusions The rate of SCD in NCAA athletes is high, with males, black athletes and basketball players at substantially higher risk. The most common finding at autopsy is AN-SUD. Media reports are more likely to capture high profile deaths, while insurance claims are not a reliable method for case identification.

Sudden cardiac death (SCD) is the leading cause of death in athletes during sport. Whether obtained for screening or diagnostic purposes, an ECG increases the ability to detect underlying cardiovascular conditions that may increase the risk for SCD. In most countries, there is a shortage of physician expertise in the interpretation of an athlete's ECG. A critical need exists for physician education in modern ECG interpretation that distinguishes normal physiological adaptations in athletes from abnormal findings suggestive of pathology. On 13-14 February 2012, an international group of experts in sports cardiology and sports medicine convened in Seattle, Washington, to define contemporary standards for ECG interpretation in athletes. The objective of the meeting was to develop a comprehensive training resource to help physicians distinguish normal ECG alterations in athletes from abnormal ECG findings that require additional evaluation for conditions associated with SCD.

Sudden cardiac death (SCD) is the leading cause of mortality in athletes during sport. A variety of mostly hereditary, structural or electrical cardiac disorders are associated with SCD in young athletes, the majority of which can be identified or suggested by abnormalities on a resting 12-lead electrocardiogram (ECG). Whether used for diagnostic or screening purposes, physicians responsible for the cardiovascular care of athletes should be knowledgeable and competent in ECG interpretation in athletes. However, in most countries a shortage of physician expertise limits wider application of the ECG in the care of the athlete. A critical need exists for physician education in modern ECG interpretation that distinguishes normal physiological adaptations in athletes from distinctly abnormal findings suggestive of underlying pathology. Since the original 2010 European Society of Cardiology recommendations for ECG interpretation in athletes, ECG standards have evolved quickly, advanced by a growing body of scientific data and investigations that both examine proposed criteria sets and establish new evidence to guide refinements. On 26-27 February 2015, an international group of experts in sports cardiology, inherited cardiac disease, and sports medicine convened in Seattle, Washington (USA), to update contemporary standards for ECG interpretation in athletes. The objective of the meeting was to define and revise ECG interpretation standards based on new and emerging research and to develop a clear guide to the proper evaluation of ECG abnormalities in athletes. This statement represents an international consensus for ECG interpretation in athletes and provides expert opinion-based recommendations linking specific ECG abnormalities and the secondary evaluation for conditions associated with SCD.

Sudden cardiac death (SCD) is the leading cause of mortality in athletes during sport. A variety of mostly hereditary, structural, or electrical cardiac disorders are associated with SCD in young athletes, the majority of which can be identified or suggested by abnormalities on a resting 12-lead electrocardiogram (ECG). Whether used for diagnostic or screening purposes, physicians responsible for the cardiovascular care of athletes should be knowledgeable and competent in ECG interpretation in athletes. However, in most countries a shortage of physician expertise limits wider application of the ECG in the care of the athlete. A critical need exists for physician education in modern ECG interpretation that distinguishes normal physiological adaptations in athletes from distinctly abnormal findings suggestive of underlying pathology. Since the original 2010 European Society of Cardiology recommendations for ECG interpretation in athletes, ECG standards have evolved quickly over the last decade; pushed by a growing body of scientific data that both tests proposed criteria sets and establishes new evidence to guide refinements. On 26-27 February 2015, an international group of experts in sports cardiology, inherited cardiac disease, and sports medicine convened in Seattle, Washington, to update contemporary standards for ECG interpretation in athletes. The objective of the meeting was to define and revise ECG interpretation standards based on new and emerging research and to develop a clear guide to the proper evaluation of ECG abnormalities in athletes. This statement represents an international consensus for ECG interpretation in athletes and provides expert opinion-based recommendations linking specific ECG abnormalities and the secondary evaluation for conditions associated with SCD.

Sudden cardiac death (SCD) is the leading cause of mortality in athletes during sport. A variety of mostly hereditary, structural, or electrical cardiac disorders are associated with SCD in young athletes, the majority of which can be identified or suggested by abnormalities on a resting 12-lead electrocardiogram (ECG). Whether used for diagnostic or screening purposes, physicians responsible for the cardiovascular care of athletes should be knowledgeable and competent in ECG interpretation in athletes. However, in most countries a shortage of physician expertise limits wider application of the ECG in the care of the athlete. A critical need exists for physician education in modern ECG interpretation that distinguishes normal physiological adaptations in athletes from distinctly abnormal findings suggestive of underlying pathology. Since the original 2010 European Society of Cardiology recommendations for ECG interpretation in athletes, ECG standards have evolved quickly over the last decade; pushed by a growing body of scientific data that both tests proposed criteria sets and establishes new evidence to guide refinements. On February 26-27, 2015, an international group of experts in sports cardiology, inherited cardiac disease, and sports medicine convened in Seattle, Washington, to update contemporary standards for ECG interpretation in athletes. The objective of the meeting was to define and revise ECG interpretation standards based on new and emerging research and to develop a clear guide to the proper evaluation of ECG abnormalities in athletes. This statement represents an international consensus for ECG interpretation in athletes and provides expert opinion-based recommendations linking specific ECG abnormalities and the secondary evaluation for conditions associated with SCD.

IMPORTANCE Trends and in-hospital outcomes associated with early adoption of the subcutaneous implantable cardioverter defibrillator (S-ICD) in the United States have not been described. OBJECTIVES To describe early use of the S-ICD in the United States and to compare in-hospital outcomes among patients undergoing S-ICD vs transvenous (TV)-ICD implantation. DESIGN, SETTING, AND PARTICIPANTS A retrospective analysis of 393 734 ICD implants reported to the National Cardiovascular Data Registry ICD Registry, a nationally representative US ICD registry, between September 28, 2012 (US Food and Drug Administration S-ICD approval date), and March 31, 2015, was conducted. A 1:1:1 propensity-matched analysis of 5760 patients was performed to compare in-hospital outcomes among patients with S-ICD with those of patients with single-chamber (SC)–ICD and dual-chamber (DC)–ICD. MAIN OUTCOMES AND MEASURES Analysis of trends in S-ICD adoption as a function of total ICD implants and comparison of in-hospital outcomes (death, complications, and defibrillation threshold [DFT] testing) among S-ICD and TV-ICD recipients. RESULTS Of the 393 734 ICD implants evaluated during the study period, 3717 were S-ICDs (0.9%). A total of 109 445 (27.8%) of the patients were female; the mean (SD) age was 67.03 (13.10) years. Use of ICDs increased from 0.2%during the fourth quarter of 2012 to 1.9% during the first quarter of 2015. Compared with SC-ICD and DC-ICD recipients, those with S-ICDs were more often younger, female, black, undergoing dialysis, and had experienced prior cardiac arrest. Among 2791 patients with S-ICD who underwent DFT testing, 2588 (92.7%), 2629 (94.2%), 2635 (94.4%), and 2784 (99.7%) were successfully defibrillated (≤65, ≤70, ≤75, and≤80 J, respectively). In the propensity-matched analysis of 5760 patients, in-hospital complication rates associated with S-ICDs (0.9%) were comparable to those of SC-ICDs (0.6%) (P = .27) and DC-ICD rates (1.5%) (P = .11). Mean (SD) length of stay after S-ICD implantation was comparable to that after SC-ICD implantation (1.1 [1.5] vs 1.0 [1.2] days; P = .77) and less than after DC-ICD implantation (1.1 [1.5] vs 1.2 [1.5] days; P < .001). CONCLUSIONS AND RELEVANCE The use of S-ICDs is rapidly increasing in the United States. Early adoption has been associated with low complication rates and high rates of successful DFT testing despite frequent use in patients with a high number of comorbidities.