70 research outputs found
High Breakdown Inference in the Mixed Linear Model
No full textMixed linear models are used to analyze data in many settings. These models have a multivariate normal formulation in most cases. The maximum likelihood estimator (MLE) or the residual MLE (REML) is usually chosen to estimate the parameters. However, the latter are based on the strong assumption of exact multivariate normality. Welsh and Richardson have shown that these estimators are not robust to small deviations from the multivariate normality. This means that in practice a small proportion of data (even only one) can drive the value of the estimates on their own. Because the model is multivariate, we propose a high-breakdown robust estimator for very general mixed linear models that include, for example, covariates. This robust estimator belongs to the class of S-estimators, from which we can derive the asymptotic properties for inference. We also use it as a diagnostic tool to detect outlying subjects. We discuss the advantages of this estimator compared with other robust estimators proposed previously and illustrate its performance with simulation studies and analysis of three datasets. We also consider robust inference for multivariate hypotheses as an alternative to the classical F-test by using a robust score-type test statistic proposed by Heritier and Ronchetti, and study its properties through simulations and analysis of real data
Robust MM-Estimation and Inference in Mixed Linear Models
No full textMixed linear models are used to analyse data in many settings. These models generally rely on the normality assumption and are often fitted by means of the maximum likelihood estimator (MLE) or the restricted maximum likelihood estimator (REML). However, the sensitivity of these estimation techniques and related tests to this underlying assumption has been identified as a weakness that can even lead to wrong interpretations. Recently Copt and Victoria-Feser(2005) proposed a high breakdown estimator, namely an S-estimator, for general mixed linear models. It has the advantage of being easy to compute - even for highly structured variance matrices - and allow the computation of a robust score test. However this proposal cannot be used to define a likelihood ratio type test which is certainly the most direct route to robustify an F-test. As the latter is usually a key tool to test hypothesis in mixed linear models, we propose two new robust estimators that allow the desired extension. They also lead to resistant Wald-type tests useful for testing contrasts and covariate efects. We study their properties theoretically and by means of simulations. An analysis of a real data set illustrates the advantage of the new approach in the presence of outlying observations.
Robust Methods in Biostatistics
No full textRobust statistics is an extension of classical statistics that specifically takes into account the concept that the underlying models used to describe data are only approximate. Its basic philosophy is to produce statistical procedures which are stable when the data do not exactly match the postulated models as it is the case for example with outliers. Robust Methods in Biostatistics proposes robust alternatives to common methods used in statistics in general and in biostatistics in particular and illustrates their use on many biomedical datasets. The methods introduced include robust estimation, testing, model selection, model check and diagnostics. They are developed for the following general classes of models: Linear regression. Generalized linear models. Linear mixed models. Marginal longitudinal data models. Cox survival analysis model. The methods are introduced both at a theoretical and applied level within the framework of each general class of models, with a particular emphasis put on practical data analysis. This book is of particular use for research students, applied statisticians and practitioners in the health field interested in more stable statistical techniques. An accompanying website provides R code for computing all of the methods described, as well as for analyzing all the datasets used in the book
Fast Algorithms for Computing High Breakdown Covariance Matrices with Missing Data
No full textRobust estimation of covariance matrices when some of the data at hand are missing is an important problem. It has been studied by Little and Smith (1987) and more recently by Cheng and Victoria-Feser (2002). The latter propose the use of high breakdown estimators and so-called hybrid algorithms (see, e.g., Woodruff and Rocke, 1994). In particular, the minimum volume ellipsoid of Rousseeuw (1984) is adapted to the case of missing data. To compute it, they use (a modified version of) the forward search algorithm (see e.g. Atkinson, 1994). In this paper, we propose to use instead a modification of the C-step algorithm proposed by Rousseeuw and Van Driessen (1999) which is actually a lot faster. We also adapt the orthogonalized Gnanadesikan-Kettenring (OGK) estimator proposed by Maronna and Zamar (2002) to the case of missing data and use it as a starting point for an adapted S-estimator. Moreover, we conduct a simulation study to compare different robust estimators in terms of their efficiency and breakdown
Are higher operator volumes for unprotected left main stem percutaneous coronary intervention associated with improved patient outcomes?: A survival analysis of 6724 procedures from the British Cardiovascular Intervention Society National Database
No full textBackground: The relationship between operator volume and survival after unprotected left main stem percutaneous coronary intervention (uLMS-PCI) is poorly defined. Methods: Data from the British Cardiovascular Intervention Society national PCI database were analyzed for all uLMS-PCI procedures performed in England and Wales between 2012 and 2014 and 4 quartiles of annualized uLMS-PCI volume (Q1-Q4) generated. Individual logistic regressions were performed for 12-month mortality to quantify the independent association between operator quartile and outcomes. Results: In total, 6724 uLMS-PCI procedures were analyzed with a negatively skewed distribution and an annualized median of 3 procedures per year. Operator volume ranged from 1 to 54 uLMS-PCI procedures/year. Within Q1, 347 operators performed a median of 2 procedures/year (interquartile range, 1-3); in Q2, 134 operators performed a median of 5 procedures/year (interquartile range, 4-6); in Q3, 59 operators performed a mean of 10 procedures/year (interquartile range, 8-12); and in Q4, 29 operators performed a mean of 21 procedures/year (interquartile range, 17-29). Higher volume operators undertook uLMS-PCI in patients with greater comorbid burden and performed more complex procedures compared with lower operator volumes. Adjusted in-hospital survival (odds ratio, 0.39 [95% CI, 0.24-0.67]; P<0.001), in-hospital major adverse cardiac and cerebral events (odds ratio, 0.41 [95% CI, 0.27-0.62]; P<0.001), and 12-month survival (odds ratio, 0.54 [95% CI, 0.39-0.73]; P<0.001) were lower in Q4 operators compared with Q1 operators. A close association between operator volume/case and superior 12-month survival was observed (P<0.001). The lower volume threshold of minimum operator uLMS-PCI volume associated with improved survival was ≥16 cases/year. Conclusions: These data suggest that operator volume is an important factor in determining outcome after uLMS-PCI.</p
Operator volumes and in-hospital outcomes: an analysis of 7,740 rotational atherectomy procedures from the BCIS national database
No full textObjectives
The aims of this study were to use a national percutaneous coronary intervention (PCI) registry to study temporal changes in procedure volumes of PCI using rotational atherectomy (ROTA-PCI), the patient and procedural factors associated with differing quartiles of operator ROTA-PCI volume, and the relationship between operator ROTA-PCI volumes and in-hospital patient outcomes.
Background
Whether higher operator volume is associated with improved outcomes after ROTA-PCI is poorly defined.
Methods
Data from the British Cardiovascular Intervention Society national PCI database were analyzed for all ROTA-PCI procedures performed in the United Kingdom between 2013 and 2016. Individual logistic regressions were performed to quantify the independent association between annual operator ROTA-PCI volume and in-hospital outcomes.
Results
In total, 7,740 ROTA-PCI procedures were performed, with a negatively skewed distribution and an annualized operator volume median of 2.5 procedures/year (range: 0.25 to 55.25). Higher volume operators undertook more complex procedures in patients with greater comorbid burdens than lower volume operators. A significant inverse association was observed between operator ROTA-PCI volume and in-hospital mortality (odds ratio [OR]: 0.986/case; 95% confidence interval (CI): 0.975 to 0.996; p = 0.007) and major adverse cardiac and cerebral events (OR: 0.983/case; 95% CI: 0.975 to 0.993; p < 0.001). Additionally, lower rates of emergency cardiac surgery (OR: 0.964/case; 95% CI: 0.939 to 0.991; p = 0.008), arterial complications (OR: 0.975/case; 95% CI: 0.975 to 0.982; p < 0.001) and in-hospital major bleeding (OR: 0.985/case; 95% CI: 0.977 to 0.993; p < 0.001) were associated with higher ROTA-PCI operator volume. Sensitivity analyses in several subgroups demonstrated a consistency of improved outcomes as annual ROTA-PCI volume increased. An annual volume of <4 ROTA-PCI procedures/year was observed to be associated with increased major adverse cardiac and cerebral events, with 239 of 432 operators (55%) not exceeding this threshold.
Conclusions
In-hospital adverse outcomes occurred less frequently as ROTA-PCI operator volume increased. These data suggest that operator volume is an important factor determining outcome after ROTA-PCI
High breakdown inference for mixed linear models
No full textMixed linear models are used to analyze data in many settings. These models have in most cases a multivariate normal formulation. The maximum likelihood estimator (MLE) or the residual MLE (REML) are usually chosen to estimate the parameters. However, the latter are based on the strong assumption of exact multivariate normality. Welsh and Richardson (1997) have shown that these estimators are not robust to small deviations from the multivariate normality. This means, in practice, that a small proportion of data (even only one) can drive the value of the estimates on their own. We present some of the most used models in the analysis of variance. We introduce the mixed linear model formulation and see that in most cases it is possible to extract independent subvectors of observation. The structure of the covariance matrix is derived for a great variety of models. Since the model is multivariate, we propose in this thesis a high breakdown multivariate robust estimator for very general mixed linear models, that include, for example, covariates. This robust estimator belongs to the class of S-estimators (Rousseeuw and Yohai 1984) from which we can derive the asymptotic properties for inference. We also use it as a diagnostic tool to detect outlying subjects. We derive the estimating equation defining the high breakdown estimator and we describe how it can be computed via a simple iterative algorithm. We study the behavior of the robust estimator through an extensive simulation study. It is compared to the maximum likelihood estimator under a great variety of configuration implying different models, different contamination patterns and different samples size. We also discuss the advantages of this estimator and illustrate its performance with the analysis of four datasets. We also consider robust inference for multivariate hypotheses as an alternative to the classical F-test by using a robust score type test statistic proposed by Heritier and Ronchetti (1994) and study its properties by means of simulations and real data analysis
Blue phase stabilization by CoPt-decorated reduced-graphene oxide nanosheets dispersed in a chiral liquid crystal
No full textWe report on the effect of reduced-graphene oxide nanosheets decorated by CoPt nanoparticles on the blue phase range of a chiral liquid crystal. By means of high-resolution ac calorimetry and polarizing optical microscopy, it is demonstrated that a small concentration of these nanosheets induces the stabilization of a single blue phase structure in comparison to three blue phases existing in the pure compound. The results are compared with other liquid crystal-dispersed graphene studies, and, moreover, a short theoretical discussion of the stabilization effect is included. © 2020 Author(s)
Access site and outcomes for unprotected left main stem percutaneous coronary intervention: An analysis of the British Cardiovascular Intervention Society Database
No full textOBJECTIVES: Using the British Cardiovascular Intervention Society percutaneous coronary intervention (PCI) database, temporal trends, predictors, and outcomes of radial access (RA) versus femoral access (FA) for unprotected left main stem percutaneous coronary intervention (LMS-PCI) were studied.BACKGROUND: Data on arterial access site for LMS-PCI are poorly defined.METHODS: Data were analyzed from 19,482 LMS-PCI procedures performed in England and Wales between 2007 and 2014. Multivariate logistic regression was used to identify predictors of access site choice and its association with outcomes.RESULTS: The frequency of FA use fell from 77.7% in 2007 to 31.7% in 2014 (p < 0.001 for trend). In the most contemporary study years (2012 to 2014), the strongest associates of FA use for unprotected LMS-PCI were renal disease, PCI for restenosis, chronic total occlusion intervention, and female sex. Use of intravascular imaging and chronic anticoagulation were associated with a higher likelihood of RA use. Complexity of the PCI procedure in the RA cohort increased significantly during the study period. Length of stay was shorter (2.6 ± 9.2 vs. 3.6 ± 9.0; p < 0.001) and same day discharge greater (43.0% vs. 26.6%; p < 0.001) with RA use. After propensity matching, RA use was associated with significant reductions in in-hospital events including access site arterial complications, major bleeding, and major adverse cardiovascular events. Conversion to RA for LMS-PCI was associated with similar reductions in the whole patient cohort. RA use was not associated with lower 12-month mortality.CONCLUSIONS: In contemporary practice, the radial artery is the predominant access site for unprotected LMS-PCI, and its use is associated with shorter length of stay, less vascular complications, and less major bleeding than femoral access.</p
Temporal trends in in-hospital outcomes following unprotected left-main percutaneous coronary intervention: an analysis of 14 522 cases from British cardiovascular intervention society database 2009 to 2017
No full textBackground:
Percutaneous coronary intervention (PCI) is increasingly used as a treatment option for unprotected left main stem artery (unprotected left main stem percutaneous intervention) disease. However, whether patient outcomes have improved over time is uncertain.
Methods:
Using the United Kingdom national PCI database, we studied all patients undergoing unprotected left main stem percutaneous intervention between 2009 and 2017. We excluded patients who presented with ST-segment–elevation, cardiogenic shock, and with an emergency indication for PCI.
Results:
Between 2009 and 2017, in the study-indicated population, 14 522 unprotected left main stem percutaneous intervention procedures were performed. Significant temporal changes in baseline demographics were observed with increasing patient age and comorbid burden. Procedural complexity increased over time, with the number of vessels treated, bifurcation PCI, number of stents used, and use of intravascular imaging and rotational atherectomy increased significantly through the study period. After adjustment for baseline differences, there were significant temporal reductions in the occurrence of peri-procedural myocardial infarction (P<0.001 for trend), in-hospital major adverse cardiac or cerebrovascular events (P<0.001 for trend), and acute procedural complications (P<0.001 for trend). In multivariable analysis examining the associates of in-hospital major adverse cardiac or cerebrovascular events, while age per year (odds ratio, 1.02 [95% CIs, 1.01–1.03]), female sex (odds ratio, 1.47 [1.19–1.82]), 3 or more stents (odds ratio, 1.67 [05% [1.02–2.67]), and patient comorbidity were associated with higher rates of in-hospital major adverse cardiac or cerebrovascular events, by contrast use of intravascular imaging (odds ratio, 0.56 [0.45–0.70]), and year of PCI (odds ratio, 0.63 [0.46–0.87]) were associated with lower rates of in-hospital major adverse cardiac or cerebrovascular events.
Conclusions:
Despite trends for increased patient and procedural complexity, in-hospital patient outcomes have improved after unprotected left main stem percutaneous intervention over time
- …
