Debajyoti Sinha

Beyond kappa: A review of interrater agreement measures

In 1960, Cohen introduced the kappa coefficient to measure chance-corrected nominal scale agreement between two raters. Since then, numerous extensions and generalizations of this interrater agreement measure have been proposed in the literature. This paper reviews and critiques various approaches to the study of interrater agreement, for which the relevant data comprise either nominal or ordinal categorical ratings from multiple raters. It presents a comprehensive compilation of the main statistical approaches to this problem, descriptions and characterizations of the underlying models, and discussions of related statistical methodologies for estimation and confidence-interval construction. The emphasis is on various practical scenarios and designs that underlie the development of these measures, and the interrelationships between them.

A New Bayesian Model for Survival Data with a Surviving Fraction

Abstract We consider Bayesian methods for right-censored survival data for populations with a surviving (cure) fraction. We propose a model that is quite different from the standard mixture model for cure rates. We provide a natural motivation and interpretation of the model and derive several novel properties of it. First, we show that the model has a proportional hazards structure, with the covariates depending naturally on the cure rate. Second, we derive several properties of the hazard function for the proposed model and establish mathematical relationships with the mixture model for cure rates. Prior elicitation is discussed in detail, and classes of noninformative and informative prior distributions are proposed. Several theoretical properties of the proposed priors and resulting posteriors are derived, and comparisons are made to the standard mixture model. A real dataset from a melanoma clinical trial is discussed in detail.

Semiparametric Bayesian analysis of survival data

Abstract This review article investigates the potential of Bayes methods for the analysis of survival data using semiparametric models based on either the hazard or the intensity function. The nonparametric part of every model is assumed to be a realization of a stochastic process. The parametric part, which may include a regression parameter or a parameter quantifying the heterogeneity of a population, is assumed to have a prior distribution with possibly unknown hyperparameters. Careful applications of some recently popular computational tools, including sampling-based algorithms, are used to find posterior estimates of several quantities of interest even when dealing with complex models and unusual data structures. The methodologies developed herein are motivated and aimed at analyzing some common types of survival data from different medical studies; here we focus on univariate survival data in the presence of fixed and time-dependent covariates, multiple event-time data for repeated nonfatal events, ...

Flexible Cure Rate Modeling Under Latent Activation Schemes

With rapid improvements in medical treatment and health care, many datasets dealing with time to relapse or death now reveal a substantial portion of patients who are cured (i.e., who never experience the event). Extended survival models called cure rate models account for the probability of a subject being cured and can be broadly classified into the classical mixture models of Berkson and Gage (BG type) or the stochastic tumor models pioneered by Yakovlev and extended to a hierarchical framework by Chen, Ibrahim, and Sinha (YCIS type). Recent developments in Bayesian hierarchical cure models have evoked significant interest regarding relationships and preferences between these two classes of models. Our present work proposes a unifying class of cure rate models that facilitates flexible hierarchical model building while including both existing cure model classes as special cases. This unifying class enables robust modeling by accounting for uncertainty in underlying mechanisms leading to cure. Issues such as regressing on the cure fraction and propriety of the associated posterior distributions under different modeling assumptions are also discussed. Finally, we offer a simulation study and also illustrate with two datasets (on melanoma and breast cancer) that reveal our frameworks ability to distinguish among underlying mechanisms that lead to relapse and cure.

Potent Antitumor Activity of a Novel Cationic Pyridinium-Ceramide Alone or in Combination with Gemcitabine against Human Head and Neck Squamous Cell Carcinomas in Vitro and in Vivo

In this study, a cationic water-soluble ceramide analog l-threo-C6-pyridinium-ceramide-bromide (l-t-C6-Pyr-Cer), which exhibits high solubility and bioavailability, inhibited the growth of various human head and neck squamous cell carcinoma (HNSCC) cell lines at low IC50 concentrations, independent of their p53 status. Consistent with its design to target negatively charged intracellular compartments, l-t-C6-Pyr-Cer accumulated mainly in mitochondria-, and nuclei-enriched fractions upon treatment of human UM-SCC-22A cells [human squamous cell carcinoma (SCC) of the hypopharynx] at 1 to 6 h. In addition to its growth-inhibitory function as a single agent, the supra-additive interaction of l-t-C6-Pyr-Cer with gemcitabine (GMZ), a chemotherapeutic agent used in HNSCC, was determined using isobologram studies. Then, the effects of this ceramide, alone or in combination with GMZ, on the growth of UM-SCC-22A xenografts in SCID mice was assessed following the determination of preclinical parameters, such as maximum tolerated dose, clearance from the blood, and bioaccumulation. Results demonstrated that treatment with l-t-C6-Pyr-Cer in combination with GMZ significantly prevented the growth of HNSCC tumors in vivo. The therapeutic efficacy of l-t-C6-Pyr-Cer/GMZ combination against HNSCC tumors was approximately 2.5-fold better than that of the combination of 5-fluorouracil/cis-platin. In addition, liquid chromatography/mass spectroscopy analysis showed that the levels of l-t-C6-Pyr-Cer in HNSCC tumors weresignificantly higher than its levels in the liver and intestines; interestingly, the combination with GMZ increased the sustained accumulation of this ceramide by approximately 40%. Moreover, treatment with l-t-C6-Pyr-Cer/GMZ combination resulted in a significant inhibition of telomerase activity and decrease in telomere length in vivo, which are among downstream targets of ceramide.

A Weibull Regression Model with Gamma Frailties for Multivariate Survival Data

Frequently in the analysis of survival data, survival times within the same group are correlated due to unobserved co-variates. One way these co-variates can be included in the model is as frailties. These frailty random block effects generate dependency between the survival times of the individuals which are conditionally independent given the frailty. Using a conditional proportional hazards model, in conjunction with the frailty, a whole new family of models is introduced. By considering a gamma frailty model, often the issue is to find an appropriate model for the baseline hazard function. In this paper a flexible baseline hazard model based on a correlated prior process is proposed and is compared with a standard Weibull model. Several model diagnostics methods are developed and model comparison is made using recently developed Bayesian model selection criteria. The above methodologies are applied to the McGilchrist and Aisbett (1991) kidney infection data and the analysis is performed using Markov Chain Monte Carlo methods.

Semiparametric Bayesian Analysis of Multiple Event Time Data

Abstract Multiple event time data (e.g., carcinogenic growths in different times and locations, multiple attacks of cardiac arrest) arise in various medical studies. A Bayesian analysis of such data based on proportional intensity model of multiple event time data is presented in this paper. The Bayesian structure is somewhat analogous to that used by Kalbfleisch in a proportional hazard model. An unobserved random frailty component is used in the proportional intensity model to take care of heterogeneity among the intensity processes in different subjects. The Monte Carlo method of sampling from multivariate distributions, the so-called Gibbs sampler, is used to sample from the joint posterior distribution of the unknown parameters. The methodology developed here is exemplified with the well-known data set on rat tumors of Gail, Santner, and Brown.

Bayesian spatial prediction

This paper presents a complete Bayesian methodology for analyzing spatial data, one which employs proper priors and features diagnostic methods in the Bayesian spatial setting. The spatial covariance structure is modeled using a rich class of covariance functions for Gaussian random fields. A general class of priors for trend, scale, and structural covariance parameters is considered. In particular, we obtain analytic results that allow easy computation of the predictive distribution for an arbitrary prior on the parameters of the covariance function using importance sampling. The computations, as well as model diagnostics and sensitivity analysis, are illustrated with a set of precipitation data.

Effects of South Carolina's sex offender registration and notification policy on adult recidivism

Some sex offender registration and notification (SORN) policies subject all registered sex offenders to Internet notification. The present study examined the effects of one such broad notification policy on sex crime recidivism. Secondary data were analyzed for a sample of 6,064 male offenders convicted of at least one sex crime between 1990 and 2004. Across a mean follow-up of 8.4 years, 490 (8%) offenders had new sex crime charges and 299 (5%) offenders had new sex crime convictions. Cox’s relative risks and competing risks models estimated the influence of registration status on risk of sexual recidivism while controlling for time at risk. Registration status did not predict recidivism in any model. These results cast doubt on the effectiveness of broad SORN policies in preventing repeat sexual assault. Policy implications, particularly with respect to the federal Adam Walsh Child Protection and Safety Act, which requires broad notification, are discussed.

On Optimality Properties of the Power Prior

The power prior is a useful general class of priors that can be used for arbitrary classes of regression models, including generalized linear models, generalized linear mixed models semiparametric survival models with censored data, frailty models, multivariate models, and nonlinear models. The power prior specification for the regression coefficients focuses on observable quantities in that the elicitation is based on historical data, D0, and a scalar quantity, a0, quantifying the heterogeneity between the current data, D, and the historical data D0. The power prior distribution is then constructed by raising the likelihood function of the historical data to the power a0, where 0 ≤ a0 ≤ 1. The scalar a0 is a precision parameter that can be viewed as a measure of compatibility between the historical and current data. In this article we give a formal justification of the power prior and show that it is an optimal class of informative priors in the sense that it minimizes a convex sum of Kullback-Leibler (KL) divergences between two specific posterior densities, in which one density is based on no incorporation of historical data and the other density is based on pooling the historical and current data. This result provides a strong motivation for using the power prior as an informative prior in Bayesian inference. In addition, we derive a formal relationship between this convex sum of KL divergences and the information-processing rules proposed by others. Specifically, we show that the power prior is a 100% efficient information-processing rule in the sense defined earlier. Several examples involving simulations as well as real datasets are examined to demonstrate the proposed methodology.