Woncheol Jang

Development and validation of the good outcome following attempted resuscitation (GO-FAR) score to predict neurologically intact survival after in-hospital cardiopulmonary resuscitation

IMPORTANCE Informing patients and providers of the likelihood of survival after in-hospital cardiac arrest (IHCA), neurologically intact or with minimal deficits, may be useful when discussing do-not-attempt-resuscitation orders. OBJECTIVE To develop a simple prearrest point score that can identify patients unlikely to survive IHCA, neurologically intact or with minimal deficits. DESIGN, SETTING, AND PARTICIPANTS The study included 51,240 inpatients experiencing an index episode of IHCA between January 1, 2007, and December 31, 2009, in 366 hospitals participating in the Get With the Guidelines-Resuscitation registry. Dividing data into training (44.4%), test (22.2%), and validation (33.4%) data sets, we used multivariate methods to select the best independent predictors of good neurologic outcome, created a series of candidate decision models, and used the test data set to select the model that best classified patients as having a very low (<1%), low (1%-3%), average (>3%-15%), or higher than average (>15%) likelihood of survival after in-hospital cardiopulmonary resuscitation for IHCA with good neurologic status. The final model was evaluated using the validation data set. MAIN OUTCOMES AND MEASURES Survival to discharge after in-hospital cardiopulmonary resuscitation for IHCA with good neurologic status (neurologically intact or with minimal deficits) based on a Cerebral Performance Category score of 1. RESULTS The best performing model was a simple point score based on 13 prearrest variables. The C statistic was 0.78 when applied to the validation set. It identified the likelihood of a good outcome as very low in 9.4% of patients (good outcome in 0.9%), low in 18.9% (good outcome in 1.7%), average in 54.0% (good outcome in 9.4%), and above average in 17.7% (good outcome in 27.5%). Overall, the score can identify more than one-quarter of patients as having a low or very low likelihood of survival to discharge, neurologically intact or with minimal deficits after IHCA (good outcome in 1.4%). CONCLUSIONS AND RELEVANCE The Good Outcome Following Attempted Resuscitation (GO-FAR) scoring system identifies patients who are unlikely to benefit from a resuscitation attempt should they experience IHCA. This information can be used as part of a shared decision regarding do-not-attempt-resuscitation orders.

A Numerical Study of PQL Estimation Biases in Generalized Linear Mixed Models Under Heterogeneity of Random Effects

The penalized quasi-likelihood (PQL) approach is the most common estimation procedure for the generalized linear mixed model (GLMM). However, it has been noticed that the PQL tends to underestimate variance components as well as regression coefficients in the previous literature. In this article, we numerically show that the biases of variance component estimates by PQL are systematically related to the biases of regression coefficient estimates by PQL, and also show that the biases of variance component estimates by PQL increase as random effects become more heterogeneous.

Incorporating spatial dependence into Bayesian multiple testing of statistical parametric maps in functional neuroimaging.

The analysis of functional neuroimaging data often involves the simultaneous testing for activation at thousands of voxels, leading to a massive multiple testing problem. This is true whether the data analyzed are time courses observed at each voxel or a collection of summary statistics such as statistical parametric maps (SPMs). It is known that classical multiplicity corrections become strongly conservative in the presence of a massive number of tests. Some more popular approaches for thresholding imaging data, such as the Benjamini-Hochberg step-up procedure for false discovery rate control, tend to lose precision or power when the assumption of independence of the data does not hold. Bayesian approaches to large scale simultaneous inference also often rely on the assumption of independence. We introduce a spatial dependence structure into a Bayesian testing model for the analysis of SPMs. By using SPMs rather than the voxel time courses, much of the computational burden of Bayesian analysis is mitigated. Increased power is demonstrated by using the dependence model to draw inference on a real dataset collected in a fMRI study of cognitive control. The model also is shown to lead to improved identification of neural activation patterns known to be associated with eye movement tasks.

Asymptotic properties of the maximum likelihood estimator for the proportional hazards model with doubly censored data

Doubly censored data, which include left as well as right censored observations, are frequently met in practice. Though estimation of the distribution function with doubly censored data has seen much study, relatively little is known about the inference of regression coefficients in the proportional hazards model for doubly censored data. In particular, theoretical properties of the maximum likelihood estimator of the regression coefficients in the proportional hazards model have not been proved yet. In this paper, we show the consistency and asymptotic normality of the maximum likelihood estimator and prove its semiparametric efficiency. The proposed methods are illustrated with simulation studies and analysis of an application from a medical study.

Density estimation for grouped data with application to line transect sampling

Line transect sampling is a method used to estimate wildlife populations, with the resulting data often grouped in intervals. Estimating the density from grouped data can be challenging. In this paper we propose a kernel density estimator of wildlife population density for such grouped data. Our method uses a combined cross-validation and smoothed bootstrap approach to select the optimal bandwidth for grouped data. Our simulation study shows that with the smoothing parameter selected with this method, the estimated density from grouped data matches the true density more closely than with other approaches. Using smoothed bootstrap, we also construct bias-adjusted confidence intervals for the value of the density at the boundary. We apply the proposed method to two grouped data sets, one from a wooden stake study where the true density is known, and the other from a survey of kangaroos in Australia.

Permutation test for incomplete paired data with application to cDNA microarray data

A paired data set is common in microarray experiments, where the data are often incompletely observed for some pairs due to various technical reasons. In microarray paired data sets, it is of main interest to detect differentially expressed genes, which are usually identified by testing the equality of means of expressions within a pair. While much attention has been paid to testing mean equality with incomplete paired data in previous literature, the existing methods commonly assume the normality of data or rely on the large sample theory. In this paper, we propose a new test based on permutations, which is free from the normality assumption and large sample theory. We consider permutation statistics with linear mixtures of paired and unpaired samples as test statistics, and propose a procedure to find the optimal mixture that minimizes the conditional variances of the test statistics, given the observations. Simulations are conducted for numerical power comparisons between the proposed permutation tests and other existing methods. We apply the proposed method to find differentially expressed genes for a colorectal cancer study.

An EM algorithm for the proportional hazards model with doubly censored data

In this paper, we consider a new procedure for estimating parameters in the proportional hazards model with doubly censored data. Computing the maximum likelihood estimator with doubly censored data is often nontrivial and requires a certain constraint optimization procedure, which is computationally unstable and sometimes fails to converge. We propose an approximated likelihood and study the maximum approximated likelihood estimator, which is obtained by maximizing the approximated likelihood. In comparison to the maximum likelihood estimator, this new estimator is stable and always converges with an efficient EM algorithm we develop. The stability of the new estimator even with moderate sample sizes is amply demonstrated through simulated and real data. For theoretical justification of the approximated likelihood, we show the consistency of the maximum approximated likelihood estimator.

Analysis of Long Period Variable Stars With Nonparametric Tests for Trend Detection

In astronomy the study of variable stars—that is, stars characterized by showing significant variation in their brightness over time—has made crucial contributions to our understanding of many phenomena, from stellar birth and evolution to the calibration of the extragalactic distance scale. In this article, we develop a method for analyzing multiple, (pseudo)-periodic time series with the goal of detecting temporal trends in their periods. We allow for nonstationary noise and for clustering among the various time series. We apply this method to the long-standing astronomical problem of identifying variable stars whose regular brightness fluctuations have periods that change over time. The results of our analysis show that such changes can be substantial, raising the possibility that astronomers’ estimates of galactic distances can be refined. Two significant contributions of our approach, relative to existing methods for this problem, are as follows: 1. The method is nonparametric, making minimal assumptions about both the temporal trends themselves but also the covariance structure of the nonstationary noise. 2. Our proposed test has higher power than existing methods. The test is based on inference for a high-dimensional normal mean, with control of the false discovery rate to account for multiplicity. We present theory and simulations to demonstrate the performance of our method. We also analyze data from the American Association of Variable Star Observers and find a monotone relationship between mean period and strength of trend similar to that identified by Hart, Koen, and Lombard (2007).