Search

Background: New methods such as machine learning techniques have been increasingly used to enhance the performance of risk predictions for clinical decision-making. However, commonly reported performance metrics may not be sufficient to capture the advantages of these newly proposed models for their adoption by health care professionals to improve care. Machine learning models often improve risk estimation for certain subpopulations that may be missed by these metrics.Methods and Results: This article addresses the limitations of commonly reported metrics for performance comparison and proposes additional metrics. Our discussions cover metrics related to overall performance, discrimination, calibration, resolution, reclassification, and model implementation. Models for predicting acute kidney injury after percutaneous coronary intervention are used to illustrate the use of these metrics.Conclusions: We demonstrate that commonly reported metrics may not have sufficient sensitivity to identify improvement of machine learning models and propose the use of a comprehensive list of performance metrics for reporting and comparing clinical risk prediction models. [ABSTRACT FROM AUTHOR]

Background: The use of transcatheter aortic valve replacement (TAVR) to treat aortic stenosis in the United States is growing, yet little is known about the variation in procedural outcomes in community practice. We developed a TAVR in-hospital mortality risk model and used it to quantify variation in mortality rates across United States (US) TAVR centers.Methods and Results: We analyzed data from 22 248 TAVR procedures performed at 318 sites participating in the Society of Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy Registry (November 2011 to October 2014). A Bayesian hierarchical model was developed to estimate hospital-specific risk-adjusted mortality rates adjusting for 40 patient baseline factors. A total of 1130 in-hospital deaths (5.1%) were observed. Reliability-adjusted risk-adjusted mortality rate estimates ranged from 3.4% to 7.7% with an interquartile range of 4.8% to 5.4%. A patient's predicted odds of dying was 80% higher if treated by a hospital 1 standard deviation above the mean compared with a hospital 1 standard deviation below the mean (odds ratio =1.8; 95% credible interval, 1.4%-2.2%).Conclusions: Risk modeling of TAVR in-hospital mortality revealed variation in risk-adjusted mortality rates during the US early commercial experience. Transcatheter Valve Therapy Registry analyses using this model will support research, feedback reporting, and the identification of factors associated with quality. [ABSTRACT FROM AUTHOR]

Background: Several antiplatelet medications used during and after percutaneous coronary intervention (PCI) are contraindicated for specific patient groups. A broad assessment of contraindicated medication use and associated clinical outcomes is not well described.Methods and Results: Using national Veterans Affairs Clinical Assessment, Reporting, and Tracking Program data for all PCI between 2007 and 2013, we evaluated patients with contraindications to commonly used antiplatelet medications during and after PCI, defined in accordance with package inserts. Adjusted association between contraindicated medication use and outcomes of periprocedural bleeding and 30-day mortality were assessed using Cox proportional hazards with inverse probability weighting. Among 64 294 patients undergoing PCI, 11 315(17.6%) had a contraindication to a common antiplatelet medication and 737 (6.5%) of these patients received a contraindicated medication. In unadjusted analyses, any contraindicated medication use was associated with both increased bleeding and 30-day mortality. In adjusted models, contraindicated abciximab use in patients with thrombocytopenia (hazard ratio, 2.23; 95% confidence interval, 1.58-3.16) and in patients with a previous stroke (hazard ratio, 1.93; 95% confidence interval, 1.37-2.71) remained significantly associated with increased bleeding. Contraindicated abciximab use was not significantly associated with 30-day mortality in adjusted models. Use of eptifibatide in dialysis patients was not significantly associated with an increased risk of bleeding or mortality.Conclusions: In this national cohort, ≈18% of patients undergoing PCI had contraindications to common antiplatelet medications. Approximately 6% of those patients received a contraindicated medication with attendant bleeding risk, although this did not translate into significantly higher risk of 30-day mortality. Continued efforts to reduce contraindicated medication use may help avoid periprocedural complications. [ABSTRACT FROM AUTHOR]

Background: Stress testing after percutaneous coronary intervention (PCI) in fee-for-service settings is common and rates vary by hospital. Rates of stress testing after PCI within integrated healthcare systems, such as the Veterans Affairs (VA) are unknown.Methods and Results: We evaluated all VA patients who underwent PCI from October 2007 through June 2010. To avoid the influence of Medicare eligibility on rates of stress testing use in the VA, we excluded Medicare eligible patients during the follow-up period. Hospital-level variation in risk-standardized rates of stress testing and the association with 1-year mortality and myocardial infarction was determined from Markov chain Monte Carlo methods. Among 10 293 patients undergoing PCI at 55 VA hospitals, 2239 (21.8%) had a stress test performed within 1 year of PCI and 3902 (37.9%) within 2 years. Most stress tests after PCI were performed with nuclear imaging (79.8%). The hospital-level risk-standardized rate of stress testing differed significantly from the average at 14 hospitals, with 8 (14.5%) hospitals significantly below and 6 (10.9%) hospitals significantly above the average stress testing rate. Hospital-level risk-standardized stress testing rates were not significantly correlated with risk-standardized mortality (Spearman ρ=-0.24; P=0.08) or myocardial infarction rates (Spearman ρ=0.20; P=0.14).Conclusions: In the VA, nearly 40% of patients underwent stress testing in the 2 years after PCI, which is a third less than published studies from other healthcare systems. However, stress testing rates varied across VA hospitals, suggesting opportunities to optimize the use of stress testing are still present in integrated healthcare systems. [ABSTRACT FROM AUTHOR]

Background: The Dual Antiplatelet Therapy Study is large streamlined clinical trial designed to evaluate antiplatelet treatment strategies in a broadly inclusive population of subjects treated with coronary stents. Whether large streamlined trials can successfully include a representative group of study sites and patients has not been formally assessed.Methods and Results: Within the National Cardiovascular Data Registry CathPCI Registry, we compared characteristics and outcomes of hospitals participating versus not participating in the Dual Antiplatelet Therapy Study. We also compared clinical and procedural characteristics of trial subjects undergoing percutaneous coronary intervention (PCI) with drug-eluting stents to contemporaneous patients within the National Cardiovascular Data Registry CathPCI Registry. Standardized differences between groups were estimated. Between September 2009 and July 2011, 1.1 million PCIs were performed among 1276 hospitals, of which 309 (24.2%) participated in the Dual Antiplatelet Therapy Study. Participating hospitals were larger (468 versus 311 beds), more frequently located in urban settings (61.2% versus 42.6%), and had higher annual PCI volumes (858 versus 378) compared with nonparticipating hospitals, although hospital case mix and procedural outcomes were similar. Compared with CathPCI patients, trial patients undergoing PCI with drug-eluting stents were similar with respect to race, sex, and rates of diabetes mellitus, hypertension, and smoking, although they had lower rates of prior cardiovascular disease.Conclusions: Within the Dual Antiplatelet Therapy Study, clinical trial sites had similar patient case mix and clinical outcomes as nonparticipating sites. Although trial participants were representative of PCI patients with respect to race, sex and most comorbidities, they had a lower prevalence of chronic cardiovascular disease compared with registry patients. Although a streamlined cardiovascular clinical trial may successfully involve a large number of hospitals and rapidly enroll a diverse population of patients, differences between eligible patients and those actually enrolled remained.Clinical Trial Registration Url: http://www.clinicaltrials.gov. Unique identifier: NCT00977938. [ABSTRACT FROM AUTHOR]

Background: Whether increasing use of radial access has improved percutaneous coronary intervention outcomes remains unknown. We sought to determine the relationship between increasing facility-level use of transradial percutaneous coronary intervention (TRI) and periprocedural outcomes.Methods and Results: Within the National Cardiovascular Data Registry CathPCI Registry, we estimated the risk-adjusted association between hospital category of change in TRI use (during the 3-year period from 2009 to 2012) and trends in access site and overall bleeding, fluoroscopy time, and contrast use among 818 facilities with low baseline TRI use. There were 4 categories of hospital change in TRI use: very low (baseline, 0.2% increasing to 1.8% at the end of 3 years), low (0.9% increasing to 8.9%), moderate (1.6% increasing to 27.2%), and high (1.0% increasing to 45.1%). Risk-adjusted access site bleeding decreased over time for all hospital categories; however, the rate of decline varied across hospital categories (P for interaction, <0.001). The decrease in access site bleeding was significantly greater for hospitals with moderate or high increases in TRI use (relative risk, 0.45, 95% confidence interval, 0.36-0.56) when compared with that of very low or low hospitals (relative risk, 0.65; 95% confidence interval, 0.58-0.74; P for comparison, 0.002). Similar findings were observed for overall bleeding. An increase in fluoroscopy time (≈1.3 minutes) was noted at hospitals with moderate and high use of TRI (P=0.01). Trends in contrast use were similar across hospital categories.Conclusions: In a national sample of hospitals performing percutaneous coronary intervention, bleeding rates decreased over time for all hospital categories of change in TRI use. The decline in bleeding outcomes was larger at hospitals with increased adoption of TRI when compared with hospitals with minimal or no change in TRI use. [ABSTRACT FROM AUTHOR]

Background: Contrast left ventriculography is a method of measuring left ventricular function usually performed at the discretion of the invasive cardiologist during cardiac catheterization. We sought to determine variation in the use of left ventriculography in the Veterans Affairs (VA) Health Care System. Methods and Results: We identified adult patients who underwent cardiac catheterization including coronary angiography between 2000 and 2009 in the VA Health Care System. We determined patient and hospital predictors of the use of left ventriculography as well as the variation in use across VA facilities. Results were validated using data from the VA's Clinical Assessment, Reporting, and Tracking (CART) program. Of 457 170 cardiac catheterization procedures among 336 853 patients, left ventriculography was performed on 263 695 (58%) patients. Use of left ventriculography decreased over time (64% in 2000 to 50% in 2009) and varied markedly across facilities (<1->95% of cardiac catheterizations). Patient factors explained little of the large variation in use between facilities. When the cohort was restricted to those with an echocardiogram in the prior 30 days and no intervening event, left ventriculography was still performed in 50% of cases. Conclusions: There is large variation in the use of left ventriculography across VA facilities that is not explained by patient characteristics. [ABSTRACT FROM AUTHOR]

Background: Variation in outcomes after percutaneous coronary interventions (PCI) may reflect differences in quality of care. To date, however, we lack a methodology to monitor and improve national hospital 30-day mortality rates among patients undergoing PCI.Methods and Results: We developed hierarchical logistic regression models to calculate hospital risk-standardized 30-day all-cause PCI mortality rates. Due to differences in risk, patients were divided into 2 cohorts: those with ST-segment elevation myocardial infarction or cardiogenic shock, and those with no ST-segment elevation myocardial infarction and no cardiogenic shock. The models were derived using 2006 data from the CathPCI Registry linked with administrative claims data, and validated using comparable 2005 data. In the derivation cohort of the ST-segment elevation myocardial infarction or shock model (n=15 123), the unadjusted 30-day mortality rate was 9.2%. The final model included 13 variables with the observed mortality rates ranging from 1.4% to 40.3% across deciles of the predicted patient mortality rates. The 25th and 75th percentiles of the risk-standardized mortality rate were 8.5% and 9.7%, with 5th and 95th percentiles of 7.6% and 11.0%. In the derivation cohort of the no ST-segment elevation myocardial infarction and no shock model (n=110 529), the unadjusted 30-day mortality rate was 1.4%. The final model included 16 variables with the observed predicted mortality rates ranging from 0.1% to 7.0% across deciles of the predicted patient mortality rates. The 25th and 75th percentiles of the risk-standardized mortality rate across 612 hospitals were 1.3% and 1.6%, with 5th and 95th percentiles of 1.0% and 2.0%.Conclusions: These National Quality Forum endorsed registry-based models produce estimates of hospital risk-standardized mortality rates for patients undergoing PCI. [ABSTRACT FROM AUTHOR]

The article presents a study which examined contemporary rates of clopidogrel use and associated factors, with a particular focus on hospital quality of myocardial infarction (MI) care, among patients with non-ST segment elevation myocardial infarction (NSTEMI) discharged from U.S. hospitals between October 2009 and March 2011. Patient-level factors such as male sex, white race, cardiac risk factors, and higher troponin levels were associated with higher prescription rates.

Background: Measurement of hospital quality has traditionally focused on processes of care and postprocedure outcomes. Appropriateness measures for percutaneous coronary intervention (PCI) assess quality as it relates to patient selection and the decision to perform PCI. The association between patient selection for PCI and processes of care and postprocedural outcomes is unknown.Methods and Results: We included 203 531 patients undergoing nonacute (elective) PCI from 779 hospitals participating in the National Cardiovascular Data Registry (NCDR) CathPCI Registry between July 2009 and April 2011. We examined the association between a hospital's proportion of nonacute PCIs categorized as inappropriate by the 2009 Appropriate Use Criteria (AUC) for Coronary Revascularization and in-hospital mortality, bleeding complications, and use of optimal guideline-directed medical therapy at discharge (ie, aspirin, thienopyridines, and statins). When categorized as hospital tertiles, the range of inappropriate PCI was 0.0% to 8.1% in the lowest tertile, 8.1% to 15.2% in the middle tertile, and 15.2% to 58.6% in the highest tertile. Compared with lowest-tertile hospitals, mortality was not significantly different at middle-tertile (adjusted odds ratio [OR], 0.93; 95% confidence interval [CI], 0.73-1.19) or highest-tertile hospitals (OR, 1.12; 95% CI, 0.88-1.43; P=0.35 for differences between tertiles). Similarly, risk-adjusted bleeding did not vary significantly (middle-tertile OR, 1.13; 95% CI, 1.02-1.16; highest-tertile OR, 1.02; 95% CI, 0.91-1.16; P=0.07 for differences between tertiles) nor did use of optimal medical therapy at discharge (85.3% versus 85.7% versus 85.2%; P=0.58).Conclusions: In a national cohort of nonacute PCIs, a hospital's proportion of inappropriate PCIs was not associated with in-hospital mortality, bleeding, or medical therapy at discharge. This suggests PCI appropriateness measures aspects of hospital PCI quality that are independent of how well the procedure is performed. Therefore, PCI appropriateness and postprocedural outcomes are both important metrics to inform PCI quality. [ABSTRACT FROM AUTHOR]

The article reports on the link between the National Cardiovascular Data Registry CathPCI Registry and longitudinal claims data from the Centers for Medicare & Medicaid (CMS) in the U.S. The study focused on hospitals and patients that are representative of the elderly population, those age 65 years and above. It concluded that a representative cohort of data was created by using deterministic matching strategies.

Background: Blood pressure (BP) control among coronary artery disease patients remains suboptimal in clinical practice, potentially due to gaps in treatment intensification and medication adherence. However, longitudinal studies evaluating these relationships and outcomes are limited.Methods and Results: We assessed BP trajectories among health maintenance organization patients with hypertension and incident coronary artery disease. BP trajectories were modeled over the year after coronary artery disease diagnosis, stratified by target BP goal. Treatment intensification (increase in BP therapies in the setting of an elevated BP), medication adherence (percentage of days covered with BP therapies), and outcomes (all-cause mortality, myocardial infarction, and revascularization) were evaluated in multivariable models: 9569 patients had a <140/90 mm Hg BP target and 12,861 had a <130/80 mm Hg BP target. Within each group, 4 trajectories were identified: good, borderline, improved, and poor control. After adjustment, increasing BP treatment intensity was significantly associated with better BP trajectories in both groups. Medication adherence had inconsistent effects. There were no significant differences in combined outcomes by BP trajectory, but among the diabetes and renal disease cohort, borderline control patients were less likely to have myocardial infarction (odds ratio, 0.61; 95% confidence interval, 0.40-0.93), and good control patients were less likely to have myocardial infarction (odds ratio, 0.53; 95% confidence interval, 0.34-0.84) or a revascularization procedure (odds ratio, 0.66; 95% confidence interval, 0.47-0.93) compared with poor control patients.Conclusions: In this health maintenance organization population, treatment intensification but not medication adherence significantly affects BP trajectories in the year after coronary artery disease diagnosis. Better BP trajectories are associated with lower rates of myocardial infarction and revascularization. [ABSTRACT FROM AUTHOR]

The article discusses a study that assesses the risk as well as the absolute benefits of an intervention in patients with heart failure (HF) and left ventricular systolic dysfunction (LVSD). The study used data from the American Heart Association's Get With the Guidelines-HF module. It suggests the need to expand evidence base for treating highest-risk patients and to develop strategies to optimize HF outcomes.

The article discusses a study that assesses the possible temporary relationship between stopping clopidogrel and death or myocardial infarction (MI) among patients who survived an acute coronary syndrome (ACS) hospitalization. Patients with acute coronary syndrome discharged on clopidogrel from an integrated care delivery system were evaluated. It concludes that there is a rise in the risk of death/MI in the 0-90 day interval after stopping clopidogrel and suggests for the development of strategies to cut this effect after clopidogrel treatment course.

The article discusses a study on the risk of death or myocardial infarction (MI) for patients who delayed filling their initial clopidogrel prescription on hospital discharge after stent implantation. The study used clinical registry and administrative data from three integrated health care delivery systems. It concludes that patients who delay filling their initial clopidogrel prescription after hospital discharge faces increased risk of death/MI and stresses the need to improve patients care during the transition process after stent implantation.

The article discusses the findings of a research which validated risk score for in-hospital mortality in patients with heart failure from the American Heart Association's Get With the Guidelines-Heart Failure program. The identified independent predictors of in-hospital death were age, admission systolic blood pressure, blood urea nitrogen, heart rate, sodium, the presence of chronic obstructive pulmonary disease, and non-African American race. Results showed that the predicted risk of in-hospital death varied by greater than 24-fold across deciles.

Background: Women have an unexplained worse outcome after myocardial infarction (MI) compared with men in many studies. Depressive symptoms predict adverse post-MI outcomes and are more prevalent among women than men. We examined whether depressive symptoms contribute to women's worse outcomes after MI.Methods and Results: In a prospective multicenter study (PREMIER), 2411 (807 women) MI patients were enrolled. Depressive symptoms were assessed with the Patient Health Questionnaire. Outcomes included 1-year rehospitalization, presence of angina using the Seattle Angina Questionnaire, and 2-year mortality. Multivariable analyses were used to evaluate the association between sex and these outcomes, adjusting for clinical characteristics. The depressive symptoms score was added to the models to evaluate whether it attenuated the association between sex and outcomes. Depressive symptoms were more prevalent in women compared with men (29% versus 18.8%, P<0.001). After adjusting for demographic factors, comorbidities, and MI severity, women had a mildly higher risk of rehospitalization (hazard ratio, 1.20; 95% CI, 1.04 to 1.40), angina (odds ratio, 1.32; 95% CI, 1.00 to 1.75), and mortality (hazard ratio, 1.27; 95% CI, 0.98 to 1.64). After adding depressive symptoms to the multivariable models, the relationship further declined toward the null, particularly for rehospitalization (hazard ratio, 1.14; 95% CI, 0.98 to 1.34) and angina (odds ratio, 1.22; 95% CI, 0.91 to 1.63), whereas there was little change in the estimate for mortality (hazard ratio, 1.24; 95% CI, 0.95 to 1.62). Depressive symptoms were significantly associated with each of the study outcomes with a similar magnitude of effect in both women and men.Conclusions: A higher prevalence of depressive symptoms in women modestly contributes to their higher rates of rehospitalization and angina compared with men but not mortality after MI. Our results support the recent recommendations of improving recognition of depressive symptoms after MI. [ABSTRACT FROM AUTHOR]