Gabriel Escarela
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Gabriel Escarela has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
- Alejandro Román Vásquez
Departamento de Matemáticas Universidad Autónoma Metropolitana–Iztapalapa CDMX Mexico
5 shared publications - José Ulises Márquez Urbina
Department of Statistics Universidad Carlos III de Madrid Leganés Spain
4 shared publications - Graciela González-Farías
Probabilidad y Estadística Centro de Investigación en Matemáticas A.C. Guanajuato City Mexico
3 shared publications Alberto Castillo-Morales
Departamento de Matemáticas, Universidad Autónoma Metropolitana - Iztapalapa, Mexico City, Mexico
2 shared publicationsBrian Francis
Lancaster University, Fylde College
2 shared publicationsKeith Soothill
Lancaster University, Cartmel College
2 shared publications
Latest Publications
- Open Access
- 6 August 2026
- Published by Wiley in Canadian Journal of Statistics
Vine copula knockoffs for variable selection in gene expression studies
Identifying clinical and genetic markers is essential for stratifying cancer patients by survival outcomes and guiding personalized treatment strategies. However, gene expression studies often involve high‐dimensional predictors with mixed data types and complex dependence, which complicates reliable variable selection. We propose a vine copula–based knockoffs framework that enables conditional variable selection in mixed clinical and genetic data with complex dependence structures. The efficacy of the proposed method is validated through extensive simulation scenarios. Its application to an ovarian cancer dataset identifies clinical and genetic markers associated with patient survival, with results consistent with survival meta‐analyses. This approach enhances biomarker discovery by linking methodological innovation with genomic applications.
- 6 August 2026
- Published by Wiley in Canadian Journal of Statistics
Open AccessVine copula knockoffs for variable selection in gene expression studies
Identifying clinical and genetic markers is essential for stratifying cancer patients by survival outcomes and guiding personalized treatment strategies. However, gene expression studies often involve high‐dimensional predictors with mixed data types and complex dependence, which complicates reliable variable selection. We propose a vine copula–based knockoffs framework that enables conditional variable selection in mixed clinical and genetic data with complex dependence structures. The efficacy of the proposed method is validated through extensive simulation scenarios. Its application to an ovarian cancer dataset identifies clinical and genetic markers associated with patient survival, with results consistent with survival meta‐analyses. This approach enhances biomarker discovery by linking methodological innovation with genomic applications.
- Open Access
- 20 January 2025
- Published by Springer Nature in Computational Statistics
- Vol. 40 (7) , 3843-3866
Derandomized Truncated D-vine Copula Knockoffs with e-values to control the false discovery rate
The Model-X knockoffs is a practical methodology for variable selection, which stands out from other selection strategies since it allows for the control of the false discovery rate, relying on finite-sample guarantees. In this article, we propose a Truncated D-vine Copula Knockoffs (TDCK) algorithm for sampling approximate knockoffs from complex multivariate distributions. Our algorithm enhances and improves features of previous attempts to sample knockoffs under the multivariate setting, with the three main contributions being: (1) the truncation of the D-vine copula, which reduces the dependence between the original variables and their corresponding knockoffs, thus improving the statistical power; (2) the employment of a straightforward non-parametric formulation for marginal transformations, eliminating the need for a specific parametric family or a kernel density estimator; (3) the use of the “rvinecopulib” R package offers better flexibility than the existing fitting vine copula knockoff methods. To eliminate the randomness from the different sets of selected variables in distinct realizations, we wrap the TDCK method with an existing derandomizing procedure for knockoffs, leading to a Derandomized Truncated D-vine Copula Knockoffs with e-values (DTDCKe) procedure. We demonstrate the robustness of the DTDCKe procedure under various scenarios with extensive simulation studies. We further illustrate its efficacy using a gene expression dataset, showing it achieves a more reliable gene selection than other competing methods when the findings are compared with those of a meta-analysis. The results indicate that our Truncated D-vine copula approach is robust and has superior power, representing an appealing approach for variable selection in different multivariate applications, particularly in gene expression analysis.
- 20 January 2025
- Published by Springer Nature in Computational Statistics
- Vol. 40 (7) , 3843-3866
Open AccessDerandomized Truncated D-vine Copula Knockoffs with e-values to control the false discovery rate
The Model-X knockoffs is a practical methodology for variable selection, which stands out from other selection strategies since it allows for the control of the false discovery rate, relying on finite-sample guarantees. In this article, we propose a Truncated D-vine Copula Knockoffs (TDCK) algorithm for sampling approximate knockoffs from complex multivariate distributions. Our algorithm enhances and improves features of previous attempts to sample knockoffs under the multivariate setting, with the three main contributions being: (1) the truncation of the D-vine copula, which reduces the dependence between the original variables and their corresponding knockoffs, thus improving the statistical power; (2) the employment of a straightforward non-parametric formulation for marginal transformations, eliminating the need for a specific parametric family or a kernel density estimator; (3) the use of the “rvinecopulib” R package offers better flexibility than the existing fitting vine copula knockoff methods. To eliminate the randomness from the different sets of selected variables in distinct realizations, we wrap the TDCK method with an existing derandomizing procedure for knockoffs, leading to a Derandomized Truncated D-vine Copula Knockoffs with e-values (DTDCKe) procedure. We demonstrate the robustness of the DTDCKe procedure under various scenarios with extensive simulation studies. We further illustrate its efficacy using a gene expression dataset, showing it achieves a more reliable gene selection than other competing methods when the findings are compared with those of a meta-analysis. The results indicate that our Truncated D-vine copula approach is robust and has superior power, representing an appealing approach for variable selection in different multivariate applications, particularly in gene expression analysis.
- Open Access
Derandomized Truncated D-vine Copula Knockoffs with e-values to control the false discovery rate
The Model-X knockoffs is a practical methodology for variable selection, which stands out from other selection strategies since it allows for the control of the false discovery rate (FDR), relying on finite-sample guarantees. In this article, we propose a Truncated D-vine Copula Knockoffs (TDCK) algorithm for sampling approximate knockoffs from complex multivariate distributions. Our algorithm enhances and improves features of previous attempts to sample knockoffs under the multivariate setting, with the three main contributions being: 1) the truncation of the D-vine copula, which reduces the dependence between the original variables and their corresponding knockoffs, improving the statistical power; 2) the employment of a straightforward non-parametric formulation for marginal transformations, eliminating the need for a specific parametric family or a kernel density estimator; 3) the use of the "rvinecopulib'' R package offers better flexibility than the existing fitting vine copula knockoff methods. To eliminate the randomness in distinct realizations resulting in different sets of selected variables, we wrap the TDCK method with an existing derandomizing procedure for knockoffs, leading to a Derandomized Truncated D-vine Copula Knockoffs with e-values (DTDCKe) procedure. We demonstrate the robustness of the DTDCKe procedure under various scenarios with extensive simulation studies. We further illustrate its efficacy using a gene expression dataset, showing it achieves a more reliable gene selection than other competing methods, when the findings are compared with those of a meta-analysis. The results indicate that our Truncated D-vine copula approach is robust and has superior power, representing an appealing approach for variable selection in different multivariate applications, particularly in gene expression analysis.
Open AccessDerandomized Truncated D-vine Copula Knockoffs with e-values to control the false discovery rate
The Model-X knockoffs is a practical methodology for variable selection, which stands out from other selection strategies since it allows for the control of the false discovery rate (FDR), relying on finite-sample guarantees. In this article, we propose a Truncated D-vine Copula Knockoffs (TDCK) algorithm for sampling approximate knockoffs from complex multivariate distributions. Our algorithm enhances and improves features of previous attempts to sample knockoffs under the multivariate setting, with the three main contributions being: 1) the truncation of the D-vine copula, which reduces the dependence between the original variables and their corresponding knockoffs, improving the statistical power; 2) the employment of a straightforward non-parametric formulation for marginal transformations, eliminating the need for a specific parametric family or a kernel density estimator; 3) the use of the "rvinecopulib'' R package offers better flexibility than the existing fitting vine copula knockoff methods. To eliminate the randomness in distinct realizations resulting in different sets of selected variables, we wrap the TDCK method with an existing derandomizing procedure for knockoffs, leading to a Derandomized Truncated D-vine Copula Knockoffs with e-values (DTDCKe) procedure. We demonstrate the robustness of the DTDCKe procedure under various scenarios with extensive simulation studies. We further illustrate its efficacy using a gene expression dataset, showing it achieves a more reliable gene selection than other competing methods, when the findings are compared with those of a meta-analysis. The results indicate that our Truncated D-vine copula approach is robust and has superior power, representing an appealing approach for variable selection in different multivariate applications, particularly in gene expression analysis.
- Open Access
- 19 January 2024
- Published by Wiley in International Journal of Mathematics and Mathematical Sciences
- Vol. 2024, 1-13
Gaussian Copula Regression Modeling for Marker Classification Metrics with Competing Risk Outcomes
Decisions regarding competing risks are usually based on a continuous-valued marker toward predicting a cause-specific outcome. The classification power of a marker can be summarized using the time-dependent receiver operating characteristic curve and the corresponding area under the curve (AUC). This paper proposes a Gaussian copula-based model to represent the joint distribution of the continuous-valued marker, the overall survival time, and the cause-specific outcome. Then, it is used to characterize the time-varying ROC curve in the context of competing risks. Covariate effects are incorporated by linking linear components to the skewed normal distribution for the margin of the marker, a parametric proportional hazards model for the survival time, and a logit model for the cause of failure. Estimation is carried out using maximum likelihood, and a bootstrap technique is implemented to obtain confidence intervals for the AUC. Information-criteria strategies are employed to find a parsimonious model. The performance of the proposed model is evaluated in simulation studies, considering different sample sizes and censoring distributions. The methods are illustrated with the reanalysis of a prostate cancer clinical trial. The joint regression strategy produces a straightforward and flexible model of the time-dependent ROC curve in the presence of competing risks, enhancing the understanding of how covariates may affect the discrimination of a marker.
- 19 January 2024
- Published by Wiley in International Journal of Mathematics and Mathematical Sciences
- Vol. 2024, 1-13
Open AccessGaussian Copula Regression Modeling for Marker Classification Metrics with Competing Risk Outcomes
Decisions regarding competing risks are usually based on a continuous-valued marker toward predicting a cause-specific outcome. The classification power of a marker can be summarized using the time-dependent receiver operating characteristic curve and the corresponding area under the curve (AUC). This paper proposes a Gaussian copula-based model to represent the joint distribution of the continuous-valued marker, the overall survival time, and the cause-specific outcome. Then, it is used to characterize the time-varying ROC curve in the context of competing risks. Covariate effects are incorporated by linking linear components to the skewed normal distribution for the margin of the marker, a parametric proportional hazards model for the survival time, and a logit model for the cause of failure. Estimation is carried out using maximum likelihood, and a bootstrap technique is implemented to obtain confidence intervals for the AUC. Information-criteria strategies are employed to find a parsimonious model. The performance of the proposed model is evaluated in simulation studies, considering different sample sizes and censoring distributions. The methods are illustrated with the reanalysis of a prostate cancer clinical trial. The joint regression strategy produces a straightforward and flexible model of the time-dependent ROC curve in the presence of competing risks, enhancing the understanding of how covariates may affect the discrimination of a marker.
- Open Access
- 19 April 2023
- Published by SAGE Publications in Statistical Methods in Medical Research
- Vol. 32 (6) , 1203-1216
Copula modeling for the estimation of measures of marker classification and predictiveness performance with survival outcomes
The discriminative and predictive power of a continuous-valued marker for survival outcomes can be summarized using the receiver operating characteristic and predictiveness curves, respectively. In this paper, fully parametric and semi-parametric copula-based constructions of the joint model of the marker and the survival time are developed for characterizing, plotting, and analyzing both curves along with other underlying performance measures. The formulations require a copula function, a parametric specification for the margin of the marker, and either a parametric distribution or a non-parametric estimator for the margin of the time to event, to respectively characterize the fully parametric and semi-parametric joint models. Estimation is carried out using maximum likelihood and a two-stage procedure for the parametric and semi-parametric models, respectively. Resampling-based methods are used for computing standard errors and confidence bounds for the various parameters, curves, and associated measures. Graphical inspection of residuals from each conditional distribution is employed as a guide for choosing a copula from a set of candidates. The performance of the estimators of various classification and predictiveness measures is assessed in simulation studies, assuming different copula and censoring scenarios. The methods are illustrated with the analysis of two markers using the familiar primary biliary cirrhosis data set.
- 19 April 2023
- Published by SAGE Publications in Statistical Methods in Medical Research
- Vol. 32 (6) , 1203-1216
Open AccessCopula modeling for the estimation of measures of marker classification and predictiveness performance with survival outcomes
The discriminative and predictive power of a continuous-valued marker for survival outcomes can be summarized using the receiver operating characteristic and predictiveness curves, respectively. In this paper, fully parametric and semi-parametric copula-based constructions of the joint model of the marker and the survival time are developed for characterizing, plotting, and analyzing both curves along with other underlying performance measures. The formulations require a copula function, a parametric specification for the margin of the marker, and either a parametric distribution or a non-parametric estimator for the margin of the time to event, to respectively characterize the fully parametric and semi-parametric joint models. Estimation is carried out using maximum likelihood and a two-stage procedure for the parametric and semi-parametric models, respectively. Resampling-based methods are used for computing standard errors and confidence bounds for the various parameters, curves, and associated measures. Graphical inspection of residuals from each conditional distribution is employed as a guide for choosing a copula from a set of candidates. The performance of the estimators of various classification and predictiveness measures is assessed in simulation studies, assuming different copula and censoring scenarios. The methods are illustrated with the analysis of two markers using the familiar primary biliary cirrhosis data set.
- 14 September 2020
- Published by Wiley in Statistics in Medicine
- Vol. 39 (28) , 4252-4266
Copula modeling of receiver operating characteristic and predictiveness curves
Receiver operating characteristic (ROC) and predictiveness curves are graphical tools to study the discriminative and predictive power of a continuous‐valued marker in a binary outcome. In this paper, a copula‐based construction of the joint density of the marker and the outcome is developed for plotting and analyzing both curves. The methodology only requires a copula function, the marginal distribution of the marker, and the prevalence rate for the model to be characterized. The adoption of the Gaussian copula and the customization of the margin for the marker are proposed for such characterization. The computation of both curves is numerically more feasible than methods that attempt to obtain one curve in terms of the other. Estimation is carried out using maximum likelihood and resampling‐based methods. Randomized quantile residuals from each conditional distribution are employed for both assessing the adequacy of the model and identifying outliers. The performance of the estimators of both curves and their underlying quantities is evaluated in simulation studies that assume different dependence structures and sample sizes. The methods are illustrated with an analysis of the level of progesterone receptor gene expression for the diagnosis and prediction of estrogen receptor‐positive breast cancer.
- 14 September 2020
- Published by Wiley in Statistics in Medicine
- Vol. 39 (28) , 4252-4266
Copula modeling of receiver operating characteristic and predictiveness curves
Receiver operating characteristic (ROC) and predictiveness curves are graphical tools to study the discriminative and predictive power of a continuous‐valued marker in a binary outcome. In this paper, a copula‐based construction of the joint density of the marker and the outcome is developed for plotting and analyzing both curves. The methodology only requires a copula function, the marginal distribution of the marker, and the prevalence rate for the model to be characterized. The adoption of the Gaussian copula and the customization of the margin for the marker are proposed for such characterization. The computation of both curves is numerically more feasible than methods that attempt to obtain one curve in terms of the other. Estimation is carried out using maximum likelihood and resampling‐based methods. Randomized quantile residuals from each conditional distribution are employed for both assessing the adequacy of the model and identifying outliers. The performance of the estimators of both curves and their underlying quantities is evaluated in simulation studies that assume different dependence structures and sample sizes. The methods are illustrated with an analysis of the level of progesterone receptor gene expression for the diagnosis and prediction of estrogen receptor‐positive breast cancer.
- 20 October 2019
- Published by Taylor & Francis in Communications in Statistics - Theory and Methods
- Vol. 50 (12) , 2831-2847
Parametric and semiparametric copula-based models for the regression analysis of competing risks
Copula-based constructions of the joint distribution of the overall survival time and the cause-specific failure are developed for the analysis of competing risks data. Covariate effects are incorporated by characterizing the margins with parametric and semiparametric proportional hazards models for the survival times, and a logistic regression model for the cause of failure. Appealing aspects of the methods are the use of single-outcome techniques for the adequate modeling of the proportional hazards component, the ability to find parsimonious models using information-criterion metrics, and the inherent simplicity and usefulness of the interpretation of the parameters. Estimation is carried out using maximum likelihood and two-stage pseudolikelihood procedures for the parametric and semiparametric models, respectively. The performance of the estimators is evaluated in simulation studies. The methods are illustrated with the reanalysis of a follicular cell lymphoma dataset.
- 20 October 2019
- Published by Taylor & Francis in Communications in Statistics - Theory and Methods
- Vol. 50 (12) , 2831-2847
Parametric and semiparametric copula-based models for the regression analysis of competing risks
Copula-based constructions of the joint distribution of the overall survival time and the cause-specific failure are developed for the analysis of competing risks data. Covariate effects are incorporated by characterizing the margins with parametric and semiparametric proportional hazards models for the survival times, and a logistic regression model for the cause of failure. Appealing aspects of the methods are the use of single-outcome techniques for the adequate modeling of the proportional hazards component, the ability to find parsimonious models using information-criterion metrics, and the inherent simplicity and usefulness of the interpretation of the parameters. Estimation is carried out using maximum likelihood and two-stage pseudolikelihood procedures for the parametric and semiparametric models, respectively. The performance of the estimators is evaluated in simulation studies. The methods are illustrated with the reanalysis of a follicular cell lymphoma dataset.
- 15 February 2018
- Published by Taylor & Francis in Journal of Applied Statistics
- Vol. 45 (15) , 2741-2759
Joint regression modeling for missing categorical covariates in generalized linear models
Missing covariates data is a common issue in generalized linear models (GLMs). A model-based procedure arising from properly specifying joint models for both the partially observed covariates and the corresponding missing indicator variables represents a sound and flexible methodology, which lends itself to maximum likelihood estimation as the likelihood function is available in computable form. In this paper, a novel model-based methodology is proposed for the regression analysis of GLMs when the partially observed covariates are categorical. Pair-copula constructions are used as graphical tools in order to facilitate the specification of the high-dimensional probability distributions of the underlying missingness components. The model parameters are estimated by maximizing the weighted log-likelihood function by using an EM algorithm. In order to compare the performance of the proposed methodology with other well-established approaches, which include complete-cases and multiple imputation, several simulation experiments of Binomial, Poisson and Normal regressions are carried out under both missing at random and non-missing at random mechanisms scenarios. The methods are illustrated by modeling data from a stage III melanoma clinical trial. The results show that the methodology is rather robust and flexible, representing a competitive alternative to traditional techniques.
- 15 February 2018
- Published by Taylor & Francis in Journal of Applied Statistics
- Vol. 45 (15) , 2741-2759
Joint regression modeling for missing categorical covariates in generalized linear models
Missing covariates data is a common issue in generalized linear models (GLMs). A model-based procedure arising from properly specifying joint models for both the partially observed covariates and the corresponding missing indicator variables represents a sound and flexible methodology, which lends itself to maximum likelihood estimation as the likelihood function is available in computable form. In this paper, a novel model-based methodology is proposed for the regression analysis of GLMs when the partially observed covariates are categorical. Pair-copula constructions are used as graphical tools in order to facilitate the specification of the high-dimensional probability distributions of the underlying missingness components. The model parameters are estimated by maximizing the weighted log-likelihood function by using an EM algorithm. In order to compare the performance of the proposed methodology with other well-established approaches, which include complete-cases and multiple imputation, several simulation experiments of Binomial, Poisson and Normal regressions are carried out under both missing at random and non-missing at random mechanisms scenarios. The methods are illustrated by modeling data from a stage III melanoma clinical trial. The results show that the methodology is rather robust and flexible, representing a competitive alternative to traditional techniques.
- Open Access
- 1 January 2017
- Published by SAGE Publications in Breast Cancer: Basic and Clinical Research
- Vol. 11
Long-Term Cause-Specific Mortality After Surgery for Women With Breast Cancer: A 20-Year Follow-Up Study From Surveillance, Epidemiology, and End Results Cancer Registries
Background: Research into long-term cause-specific mortality of women diagnosed with breast cancer is important because it allows for the splitting of the population into patients who eventually die from breast cancer and from other causes. The adoption of this approach helps to identify patients with an elevated risk of eventual death from breast cancer. Objective: The primary aim of this study was to examine the associations between both sociodemographic and clinicopathologic characteristics and the underlying risks of death from breast cancer and from other causes for women diagnosed with breast cancer. A second aim was to propose a predictive biomarker of cause-specific mortality in terms of treatment and several important characteristics of a patient. Methods: A cohort of 16 511 female patients diagnosed with breast cancer in 1990 was obtained from the Surveillance, Epidemiology, and End Results cancer registries and followed for 20 years. A mixture model for the regression analysis of competing risks was used to identify factors and confounders that affected either the eventual cause-specific mortality or conditional cause-specific hazard rates, or both. Missing data were handled with multiple imputation. Results: Curvilinear relationships of age at diagnosis along with race, marital status, breast cancer type, tumor size, estrogen receptor status, extension, lymph node status, type of surgery, and radiotherapy status were significant risk factors for the cause-specific mortality, with extension and lymph node status appearing to be confounded with the effects of both type of surgery and radiotherapy status. The score obtained from combining a set of predictors showed to be an accurate predictive biomarker. Conclusions: In cause-specific mortality of women diagnosed breast cancer, prognosis appears to depend on both sociodemographic and clinicopathologic factors. The predictive biomarker proposed in this study may help identifying the level of seriousness of the disease earlier than traditional methods, potentially guiding future allocation of resources for better patient care and management strategies.
- 1 January 2017
- Published by SAGE Publications in Breast Cancer: Basic and Clinical Research
- Vol. 11
Open AccessLong-Term Cause-Specific Mortality After Surgery for Women With Breast Cancer: A 20-Year Follow-Up Study From Surveillance, Epidemiology, and End Results Cancer Registries
Background: Research into long-term cause-specific mortality of women diagnosed with breast cancer is important because it allows for the splitting of the population into patients who eventually die from breast cancer and from other causes. The adoption of this approach helps to identify patients with an elevated risk of eventual death from breast cancer. Objective: The primary aim of this study was to examine the associations between both sociodemographic and clinicopathologic characteristics and the underlying risks of death from breast cancer and from other causes for women diagnosed with breast cancer. A second aim was to propose a predictive biomarker of cause-specific mortality in terms of treatment and several important characteristics of a patient. Methods: A cohort of 16 511 female patients diagnosed with breast cancer in 1990 was obtained from the Surveillance, Epidemiology, and End Results cancer registries and followed for 20 years. A mixture model for the regression analysis of competing risks was used to identify factors and confounders that affected either the eventual cause-specific mortality or conditional cause-specific hazard rates, or both. Missing data were handled with multiple imputation. Results: Curvilinear relationships of age at diagnosis along with race, marital status, breast cancer type, tumor size, estrogen receptor status, extension, lymph node status, type of surgery, and radiotherapy status were significant risk factors for the cause-specific mortality, with extension and lymph node status appearing to be confounded with the effects of both type of surgery and radiotherapy status. The score obtained from combining a set of predictors showed to be an accurate predictive biomarker. Conclusions: In cause-specific mortality of women diagnosed breast cancer, prognosis appears to depend on both sociodemographic and clinicopathologic factors. The predictive biomarker proposed in this study may help identifying the level of seriousness of the disease earlier than traditional methods, potentially guiding future allocation of resources for better patient care and management strategies.
- 11 July 2016
- Published by SAGE Publications in Statistical Methods in Medical Research
- Vol. 25 (4) , 1579-1595
Addressing missing covariates for the regression analysis of competing risks: Prognostic modelling for triaging patients diagnosed with prostate cancer
Competing risks arise in medical research when subjects are exposed to various types or causes of death. Data from large cohort studies usually exhibit subsets of regressors that are missing for some study subjects. Furthermore, such studies often give rise to censored data. In this article, a carefully formulated likelihood-based technique for the regression analysis of right-censored competing risks data when two of the covariates are discrete and partially missing is developed. The approach envisaged here comprises two models: one describes the covariate effects on both long-term incidence and conditional latencies for each cause of death, whilst the other deals with the observation process by which the covariates are missing. The former is formulated with a well-established mixture model and the latter is characterised by copula-based bivariate probability functions for both the missing covariates and the missing data mechanism. The resulting formulation lends itself to the empirical assessment of non-ignorability by performing sensitivity analyses using models with and without a non-ignorable component. The methods are illustrated on a 20-year follow-up involving a prostate cancer cohort from the National Cancer Institutes Surveillance, Epidemiology, and End Results program.
- 11 July 2016
- Published by SAGE Publications in Statistical Methods in Medical Research
- Vol. 25 (4) , 1579-1595
Addressing missing covariates for the regression analysis of competing risks: Prognostic modelling for triaging patients diagnosed with prostate cancer
Competing risks arise in medical research when subjects are exposed to various types or causes of death. Data from large cohort studies usually exhibit subsets of regressors that are missing for some study subjects. Furthermore, such studies often give rise to censored data. In this article, a carefully formulated likelihood-based technique for the regression analysis of right-censored competing risks data when two of the covariates are discrete and partially missing is developed. The approach envisaged here comprises two models: one describes the covariate effects on both long-term incidence and conditional latencies for each cause of death, whilst the other deals with the observation process by which the covariates are missing. The former is formulated with a well-established mixture model and the latter is characterised by copula-based bivariate probability functions for both the missing covariates and the missing data mechanism. The resulting formulation lends itself to the empirical assessment of non-ignorability by performing sensitivity analyses using models with and without a non-ignorable component. The methods are illustrated on a 20-year follow-up involving a prostate cancer cohort from the National Cancer Institutes Surveillance, Epidemiology, and End Results program.