Zero-inflated models for censored and overdispersed count data have received little attention so far, except for the zero-inflated Poisson (ZIP) model which assumes that overdispersion is entirely caused by zero-inflation. When additional overdispersion is present, useful alternatives to ZIP are given by the zero-inflated generalized Poisson (ZIGP) and zero-inflated negative binomial (ZINB) models. This paper investigates properties of the maximum likelihood estimator (MLE) in ZIGP and ZINB regression models when the count response is subject to right-censoring. Simulations are used to examine performance (bias, mean square error, coverage probabilities and standard error calculations) of the MLE. Results suggest that maximum likelihood yields accurate inference. A simple, efficient and easy-to-implement methodology for variable selection is also proposed. It is applicable even when the number of predictors is very large and yields interpretable and sound results. The proposed methods are applied to a dataset of healthcare demand.
Health subjectivity measurement is essential to understand the interconnection complexity between physical health and psychological state of an individual. For this, all the elements that make up a subjective measurement questionnaire should contribute to apprehending a single dimension. However, in practice unidimensionality of a questionnaire is rarely verified. In this context, this raises questions about the degree of relevance of studies based on these measures.
In order to anticipate preventive measures to support and maintain functional autonomy community-dwelling in elderly people followed prospectively and consequently delay loss of independence, a comprehension of the concept of functional independence trajectories is necessary. The aim of this paper is to create a readable synthesis of the resources available in the literature for different functional independence trajectories of older person and the methods of prevention proposed to maintain autonomy. The resources discussed in this document are clinical, mathematical and statistical.
This paper aims to establish some asymptotic properties of kernel estimators of the conditional distribution function and the conditional quantile when the lifetime observations and the covariates are associated.
Identifying the right, or the optimal, dose is one of the major difficulty during drug development. As a
consequence, the dedicated dose-finding study is a key milestone in the drug development and recent methodological progress has been made for the analysis of those studies, in putting more emphasis on the identification of the efficacy dose-response profile. The placebo controlled, parallel group, fixed design is still a standard for the dose-finding studies, but adaptive designs are becoming more attractive in some cirumstances. The aim of this paper is to review the most recent methodologies for adaptive dose-finding trials : those methods are applicable in the cases when one or several doses are dropped after the interim analysis as well when only the patient allocation is modified after the interim analysis. This review is complemented by simulations illustrating the methodologies, including a two-stage version of the MCP-Mod method proposed by the authors.