Littérature scientifique sur le sujet « Censored failure time outcome »

My PhD dissertation deals with statistical methods for cut-point finding for continuous biomarkers. Categorization is often needed for clinical decision making when dealing with diagnostic (or prognostic) biomarkers and a dichotomous or censored failure time outcome. This allows the definition of two or more prognostic risk groups, or also patient’s stratifications for inclusion in randomized clinical trials (RCTs). We investigate the following cut-point finding methods: minimum P-value, Youden index, concordance probability and point closest to-(0,1) corner in the ROC plane. We compare them by assuming both Normal and Gamma biomarker distributions, showing whether they lead to the identification of the same true cut-point and further investigating their performance by simulation. Within the framework of censored survival data, we will consider here new estimation approaches of the optimal cut-point, which use a conditional weighting method to estimate the true positive and false positive fractions. Motivating examples on real datasets are discussed within the dissertation for both the dichotomous and censored failure time outcome. In all simulation scenarios, the point closest-to-(0,1) corner in the ROC plane and concordance probability approaches outperformed the other methods. Both these methods showed good performance in the estimation of the optimal cut-point of a biomarker. However, to improve results communicability, the Youden index or the concordance probability associated to the estimated cut-point could be reported to summarize the associated classification accuracy. The use of the minimum P-value approach for cut-point finding is not recommended because its objective function is computed under the null hypothesis of absence of association between the true disease status and X. This is in contrast with the presence of some discrimination potential of the biomarker X that leads to the dichotomization issue. The investigated cut-point finding methods are based on measures, i.e. sensitivity and specificity, defined conditionally on the outcome. My PhD dissertation opens the question on whether these methods could be applied starting from predictive values, that typically represent the most useful information for clinical decisions on treatments. However, while sensitivity and specificity are invariant to disease prevalence, predictive values vary across populations with different disease prevalence. This is an important drawback of the use of predictive values for cut-point finding. More in general, great care should be taken when establishing a biomarker cut-point for clinical use. Methods for categorizing new biomarkers are often essential in clinical decision-making even if categorization of a continuous biomarker is gained at a considerable loss of power and information. In the future, new methods involving the study of the functional form between the biomarker and the outcome through regression techniques such as fractional polynomials or spline functions should be considered to alternatively define cut-points for clinical use. Moreover, in spite of the aforementioned drawback related to the use of predictive values, we also think that additional new methods for cut-point finding should be developed starting from predictive values.

Title from PDF of title page (University of Missouri--Columbia, viewed on Feb 26, 2010). The entire thesis text is included in the research.pdf file; the official abstract appears in the short.pdf file; a non-technical public abstract appears in the public.pdf file. Dissertation advisor: Dr. (Tony) Jianguo Sun. Includes bibliographical references.

Thesis (Ph.D.)--University of Missouri-Columbia, 2006.<br>The entire dissertation/thesis text is included in the research.pdf file; the official abstract appears in the short.pdf file (which also appears in the research.pdf); a non-technical general description, or public abstract, appears in the public.pdf file. Title from title screen of research.pdf file viewed on (May 2, 2007) Vita. Includes bibliographical references.

Thesis (Ph.D.)--University of Missouri-Columbia, 2007.<br>The entire dissertation/thesis text is included in the research.pdf file; the official abstract appears in the short.pdf file (which also appears in the research.pdf); a non-technical general description, or public abstract, appears in the public.pdf file. Title from title screen of research.pdf file (viewed on March 6, 2009) Includes bibliographical references.

Thesis (Ph.D.)--University of Missouri-Columbia, 2006.<br>The entire dissertation/thesis text is included in the research.pdf file; the official abstract appears in the short.pdf file (which also appears in the research.pdf); a non-technical general description, or public abstract, appears in the public.pdf file. Title from title screen of research.pdf file viewed on (May 2, 2007) Vita. Includes bibliographical references.

Failure time data analysis, or survival analysis, is involved in various research fields, such as medicine and public health. One basic assumption in standard survival analysis is that every individual in the study population will eventually experience the event of interest. However, this assumption is usually violated in practice, for example when the variable of interest is the time to relapse of a curable disease resulting in the existence of long-term survivors. Also, presence of unobservable risk factors in the group of susceptible individuals may introduce heterogeneity to the population, which is not properly addressed in standard survival models. Moreover, the individuals in the population may be grouped in clusters, where there are associations among observations from a cluster. There are methodologies in the literature to address each of these problems, but there is yet no natural and satisfactory way to accommodate the coexistence of a non-susceptible group and the heterogeneity in the susceptible group under a univariate setting. Also, various kinds of associations among survival data with a cure are not properly accommodated. To address the above-mentioned problems, a class of models is introduced to model univariate and multivariate data with long-term survivors. A semiparametric cure model for univariate failure time data with long-term survivors is introduced. It accommodates a proportion of non-susceptible individuals and the heterogeneity in the susceptible group using a compound- Poisson distributed random effect term, which is commonly called a frailty. It is a frailty-Cox model which does not place any parametric assumption on the baseline hazard function. An estimation method using multiple imputation is proposed for right-censored data, and the method is naturally extended to accommodate interval-censored data. The univariate cure model is extended to a multivariate setting by introducing correlations among the compound- Poisson frailties for individuals from the same cluster. This multivariate cure model is similar to a shared frailty model where the degree of association among each pair of observations in a cluster is the same. The model is further extended to accommodate repeated measurements from a single individual leading to serially correlated observations. Similar estimation methods using multiple imputation are developed for the multivariate models. The univariate model is applied to a breast cancer data and the multivariate models are applied to the hypobaric decompression sickness data from National Aeronautics and Space Administration, although the methodologies are applicable to a wide range of data sets.<br>published_or_final_version<br>Statistics and Actuarial Science<br>Master<br>Master of Philosophy

Kendall and Gehan estimating functions are used to estimate the regression parameter in accelerated failure time (AFT) model with censored observations. The accelerated failure time model is the preferred survival analysis method because it maintains a consistent association between the covariate and the survival time. The jackknife empirical likelihood method is used because it overcomes computation difficulty by circumventing the construction of the nonlinear constraint. Jackknife empirical likelihood turns the statistic of interest into a sample mean based on jackknife pseudo-values. U-statistic approach is used to construct the confidence intervals for the regression parameter. We conduct a simulation study to compare the Wald-type procedure, the empirical likelihood, and the jackknife empirical likelihood in terms of coverage probability and average length of confidence intervals. Jackknife empirical likelihood method has a better performance and overcomes the under-coverage problem of the Wald-type method. A real data is also used to illustrate the proposed methods.

In clinical trials and cohort studies the event of interest is often not observable, and is known only to have occurred between the visit when the event was first observed and the previous visit such data are called interval-censored. This thesis develops three pieces of research that build upon published methods for interval-censored data. Novel methods are developed which can be applied via self-written macros in the standard packages, with the aim of increasing the use of appropriate methods in applied medical research. The non-parametric maximum likelihood estimator 1,2 (NPMLE) is the most common statistical method for estimating of the survivor function for interval-censored data. However, the choice of method for obtaining confidence intervals for the survivor function is unclear. Three methods are assessed and compared using simulated data and data from the MRC Delta trial 3. Non- or semi-parametric methods that correctly account for interval-censoring are not readily available in statistical packages. Typically the event time is taken to be the right endpoint of the censoring interval and standard methods (e.g. Kaplan-Meier) for the analysis of right-censored failure time data are used, giving biased estimates of the survival curve. A simulation study compared simple imputation using the right endpoint and interval midpoint to the NPMLE and a proposed smoothed version of the NPMLE that extends the work of Pan and Chappell 4. These methods were also applied to data from the CHIPS study 5. Different approaches to the estimation of a binary covariate are compared: (i) a proportional hazards model 6, (ii) a piecewise exponential model 7, (iii) a simpler proportional hazards model based on imputed event times, and (iv) a proposed approximation to the piecewise exponential model that is a more rigorous alternative to simple imputation methods whilst simple to fit using standard software.