Search

The analysis of recurrent events in the presence of terminal events requires special attention. Several approaches have been suggested for such analyses either using intensity models or marginal models. When analysing treatment effects on recurrent events in controlled trials, special attention should be paid to competing deaths and their impact on interpretation. This paper proposes a method that formulates a marginal model for recurrent events and terminal events simultaneously. Estimation is based on pseudo-observations for both the expected number of events and survival probabilities. Various relevant hypothesis tests in the framework are explored. Theoretical derivations and simulation studies are conducted to investigate the behaviour of the method. The method is applied to two real data examples. The bivariate marginal pseudo-observation model carries the strength of a two-dimensional modelling procedure and performs well in comparison with available models. Finally, an extension to a three-dimensional model, which decomposes the terminal event per death cause, is proposed and exemplified., (© 2021. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.)

Pseudo-values provide a method to perform regression analysis for complex quantities with right-censored data. A further complication, interval-censored data, appears when events such as dementia are studied in an epidemiological cohort. We propose an extension of the pseudo-value approach for interval-censored data based on a semi-parametric estimator computed using penalised likelihood and splines. This estimator takes interval-censoring and competing risks into account in an illness-death model. We apply the pseudo-value approach to three mean value parameters of interest in studies of dementia: the probability of staying alive and non-demented, the restricted mean survival time without dementia and the absolute risk of dementia. Simulation studies are conducted to examine properties of pseudo-values based on this semi-parametric estimator. The method is applied to the French cohort PAQUID, which included more than 3,000 non-demented subjects, followed for dementia for more than 25 years.

We study regression models for mean value parameters in survival analysis based on pseudo-observations. Such parameters include the survival probability and the cumulative incidence in a single point as well as the restricted mean life time and the cause-specific number of years lost. Goodness of fit techniques for such models based on cumulative sums of pseudo-residuals are derived including asymptotic results and Monte Carlo simulations. Practical examples from liver cirrhosis and bone marrow transplantation are also provided.

Poisson regression is an important tool in register-based epidemiology where it is used to study the association between exposure variables and event rates. In this paper, we will discuss the situation with 'large n and small p', where n is the sample size and p is the number of available covariates. Specifically, we are concerned with modeling options when there are time-varying covariates that can have time-varying effects. One problem is that tests of the proportional hazards assumption, of no interactions between exposure and other observed variables, or of other modeling assumptions have large power due to the large sample size and will often indicate statistical significance even for numerically small deviations that are unimportant for the subject matter. Another problem is that information on important confounders may be unavailable. In practice, this situation may lead to simple working models that are then likely misspecified. To support and improve conclusions drawn from such models, we discuss methods for sensitivity analysis, for estimation of average exposure effects using aggregated data, and a semi-parametric bootstrap method to obtain robust standard errors. The methods are illustrated using data from the Danish national registries investigating the diabetes incidence for individuals treated with antipsychotics compared with the general unexposed population., (Copyright © 2015 John Wiley & Sons, Ltd.)

A method based on pseudo-observations has been proposed for direct regression modeling of functionals of interest with right-censored data, including the survival function, the restricted mean and the cumulative incidence function in competing risks. The models, once the pseudo-observations have been computed, can be fitted using standard generalized estimating equation software. Regression models can however yield problematic results if the number of covariates is large in relation to the number of events observed. Guidelines of events per variable are often used in practice. These rules of thumb for the number of events per variable have primarily been established based on simulation studies for the logistic regression model and Cox regression model. In this paper we conduct a simulation study to examine the small sample behavior of the pseudo-observation method to estimate risk differences and relative risks for right-censored data. We investigate how coverage probabilities and relative bias of the pseudo-observation estimator interact with sample size, number of variables and average number of events per variable.

A predicted risk of 17% can be called reliable if it can be expected that the event will occur to about 17 of 100 patients who all received a predicted risk of 17%. Statistical models can predict the absolute risk of an event such as cardiovascular death in the presence of competing risks such as death due to other causes. For personalized medicine and patient counseling, it is necessary to check that the model is calibrated in the sense that it provides reliable predictions for all subjects. There are three often encountered practical problems when the aim is to display or test if a risk prediction model is well calibrated. The first is lack of independent validation data, the second is right censoring, and the third is that when the risk scale is continuous, the estimation problem is as difficult as density estimation. To deal with these problems, we propose to estimate calibration curves for competing risks models based on jackknife pseudo-values that are combined with a nearest neighborhood smoother and a cross-validation approach to deal with all three problems., (Copyright © 2014 John Wiley & Sons, Ltd.)

When using the co-twin control design for analysis of event times, one needs a model to address the possible within-pair association. One such model is the shared frailty model in which the random frailty variable creates the desired within-pair association. Standard inference for this model requires independence between the random effect and the covariates. We study how violations of this assumption affect inference for the regression coefficients and conclude that substantial bias may occur. We propose an alternative way of making inference for the regression parameters by using a fixed-effects models for survival in matched pairs. Fitting this model to data generated from the frailty model provides consistent and asymptotically normal estimates of regression coefficients, no matter whether the independence assumption is met.

We study the competing risks model and show that the cause j cumulative incidence function integrated from 0 to τ has a natural interpretation as the expected number of life years lost due to cause j before time τ. This is analogous to the τ-restricted mean lifetime, which is the survival function integrated from 0 to τ. It is discussed how the number of years lost may be related to subject-specific explanatory variables in a regression model based on pseudo-observations, and the method is exemplified using data from a bone marrow transplantation study. Finally, inclusion of standard mortality rates is discussed., (Copyright © 2013 John Wiley & Sons, Ltd.)

Prediction of cumulative incidences is often a primary goal in clinical studies with several endpoints. We compare predictions among competing risks models with time-dependent covariates. For a series of landmark time points, we study the predictive accuracy of a multi-state regression model, where the time-dependent covariate represents an intermediate state, and two alternative landmark approaches. At each landmark time point, the prediction performance is measured as the t-year expected Brier score where pseudovalues are constructed in order to deal with right-censored event times. We apply the methods to data from a bone marrow transplant study where graft versus host disease is considered a time-dependent covariate for predicting relapse and death in remission., (Copyright © 2013 John Wiley & Sons, Ltd.)

The basic parameters in both survival analysis and more general multistate models, including the competing risks model and the illness-death model, are the transition hazards. It is often necessary to supplement the analysis of such models with other model parameters, which are all functionals of the transition hazards. Unfortunately, not all such functionals are equally meaningful in practical contexts, even though they may be mathematically well defined. We have found it useful to check whether the functionals satisfy three simple principles, which may be used as criteria for practical interpretability., (Copyright © 2011 John Wiley & Sons, Ltd.)

Life-course epidemiology seeks to better understand the mechanisms that lead to the development of chronic diseases. An example is the mechanism leading from body size to coronary heart disease (CHD); one way to acquire a better understanding of this mechanism is to investigate to what extent it works through other risk factors. In this paper, the dynamic path analysis model is presented as a tool to analyze these dynamic mechanisms in life-course epidemiology. A key feature of dynamic path analysis is its ability to decompose the total effect of a risk factor into a direct effect (not mediated by other variables) and indirect effects (mediated through other variables). This is illustrated by examining the associations between repeated measurements of body mass index (BMI) and systolic blood pressure (SBP) and the risk of CHD in a sample of Danish men between 1976 and 2006. The effect of baseline BMI on the risk of CHD is decomposed into a direct effect and indirect effects going through later BMI, concurrent SBP, or later SBP. In conclusion, dynamic path analysis is a flexible tool that by the decomposition of effects can be used to increase the understanding of mechanisms that underlie the etiology of chronic disease.

The effect of event-dependent sampling of processes consisting of recurrent events is investigated when analyzing whether the risk of recurrence increases with event count. We study the situation where processes are selected for study if an event occurs in a certain selection interval. Motivation comes from psychiatric epidemiology where repeated hospital admissions are studied for patients with affective disease, as seen in Kessing et al. (Acta Psychiatr Scand 109:339-344, 2004b). For the selected processes, either only disease course from selection and onwards is used in the analysis, or, both retrospective and prospective disease course histories are used. We examine two methods to correct for the selection depending on which data are used in the analysis. In the first case, the conditional distribution of the process given the pre-selection history is determined. In the second case, an inverse-probability-of-selection weighting scheme is suggested. The ability of the methods to correct for the bias due to selection is investigated with simulations. Furthermore, the methods are applied to affective disease data from a register-based study (Kessing et al. Br J Psychiatry 185:372-377, 2004a) and from a long-term clinical study (Kessing et al. Acta Psychiatr Scand 109:339-344, 2004b).

Time-dependent covariates are frequently encountered in regression analysis for event history data and competing risks. They are often essential predictors, which cannot be substituted by time-fixed covariates. This study briefly recalls the different types of time-dependent covariates, as classified by Kalbfleisch and Prentice [The Statistical Analysis of Failure Time Data, Wiley, New York, 2002] with the intent of clarifying their role and emphasizing the limitations in standard survival models and in the competing risks setting. If random (internal) time-dependent covariates are to be included in the modeling process, then it is still possible to estimate cause-specific hazards but prediction of the cumulative incidences and survival probabilities based on these is no longer feasible. This article aims at providing some possible strategies for dealing with these prediction problems. In a multi-state framework, a first approach uses internal covariates to define additional (intermediate) transient states in the competing risks model. Another approach is to apply the landmark analysis as described by van Houwelingen [Scandinavian Journal of Statistics 2007, 34, 70-85] in order to study cumulative incidences at different subintervals of the entire study period. The final strategy is to extend the competing risks model by considering all the possible combinations between internal covariate levels and cause-specific events as final states. In all of those proposals, it is possible to estimate the changes/differences of the cumulative risks associated with simple internal covariates. An illustrative example based on bone marrow transplant data is presented in order to compare the different methods.

The experience of a patient in a survival study may be modelled as a process with two states and one possible transition from an "alive" state to a "dead" state. In some studies, however, the "alive" state may be partitioned into two or more intermediate (transient) states, each of which corresponding to a particular stage of the illness. In such studies, multi-state models can be used to model the movement of patients among the various states. In these models issues, of interest include the estimation of progression rates, assessing the effects of individual risk factors, survival rates or prognostic forecasting. In this article, we review modelling approaches for multi-state models, and we focus on the estimation of quantities such as the transition probabilities and survival probabilities. Differences between these approaches are discussed, focussing on possible advantages and disadvantages for each method. We also review the existing software currently available to fit the various models and present new software developed in the form of an R library to analyse such models. Different approaches and software are illustrated using data from the Stanford heart transplant study and data from a study on breast cancer conducted in Galicia, Spain.

In bone marrow transplantation studies, patients are followed over time and a number of events may be observed. These include both ultimate events like death and relapse and transient events like graft versus host disease and graft recovery. Such studies, therefore, lend themselves for using an analytic approach based on multi-state models. We will give a review of such methods with emphasis on regression models for both transition intensities and transition- and state occupation probabilities. Both semi-parametric models, like the Cox regression model, and parametric models based on piecewise constant intensities will be discussed.

We adopt the total time on test procedure to investigate monotone time trends in the intensity in a repeated event setting. The correct model is assumed to be a proportional hazards model, with a random effect to account for dependence within subjects. The method offers a simple routine for testing relevant hypotheses for recurrent event processes, without making distributional assumptions about the frailty. Such assumptions may severely affect conclusions concerning regression coefficients and cause bias in the estimated heterogeneity. The method is illustrated by re-analyzing Danish registry data and a long-term Swiss clinical study on recurrence in affective disorder., (Copyright (c) 2008 John Wiley & Sons, Ltd.)

Objectives: Sickness absence is the outcome in many epidemiologic studies and is often based on summary measures such as the number of sickness absences per year. In this study the use of modern statistical methods was examined by making better use of the available information. Since sickness absence data deal with events occurring over time, the use of statistical models for survival data has been reviewed, and the use of frailty models has been proposed for the analysis of such data., Methods: Three methods for analyzing data on sickness absences were compared using a simulation study involving the following: (i) Poisson regression using a single outcome variable (number of sickness absences), (ii) analysis of time to first event using the Cox proportional hazards model, and (iii) frailty models, which are random effects proportional hazards models. Data from a study of the relation between the psychosocial work environment and sickness absence were used to illustrate the results., Results: Standard methods were found to underestimate true effect sizes by approximately one-tenth [method i] and one-third [method ii] and to have lower statistical power than frailty models., Conclusions: An uncritical use of standard methods may underestimate the effect of work environment exposures or leave predictors of sickness absence undiscovered.

The associations in mortality of adult adoptees and their biological or adoptive parents have been studied in order to separate genetic and environmental influences. The 1003 Danish adoptees born 1924-26 have previously been analysed in a Cox regression model, using dichotomised versions of the parents' lifetimes as covariates. This model will be referred to as the conditional Cox model, as it analyses lifetimes of adoptees conditional on parental lifetimes. Shared frailty models may be more satisfactory by using the entire observed lifetime of the parents. In a simulation study, sample size, distribution of lifetimes, truncation- and censoring patterns were chosen to illustrate aspects of the adoption dataset, and were generated from the conditional Cox model or a shared frailty model with gamma distributed frailties. First, efficiency was compared in the conditional Cox model and a shared frailty model, based on the conditional approach. For data with type 1 censoring the models showed no differences, whereas in data with random or no censoring, the models had different power in favour of the one from which data were generated. Secondly, estimation in the shared frailty model by a conditional approach or a two-stage copula approach was compared. Both approaches worked well, with no sign of dependence upon the truncation pattern, but some sign of bias depending on the censoring. For frailty parameters close to zero, we found bias when the estimation procedure used did not allow negative estimates. Based on this evaluation, we prefer to use frailty models allowing for negative frailty parameter estimates. The conclusions from earlier analyses of the adoption study were confirmed, though without greater precision than using the conditional Cox model. Analyses of associations between parental lifetimes are also presented.

The copula of a bivariate distribution, constructed by making marginal transformations of each component, captures all the information in the bivariate distribution about the dependence between two variables. For frailty models for bivariate data the choice of a family of distributions for the random frailty corresponds to the choice of a parametric family for the copula. A class of tests of the hypothesis that the copula is in a given parametric family, with unspecified association parameter, based on bivariate right censored data is proposed. These tests are based on first making marginal Kaplan-Meier transformations of the data and then comparing a non-parametric estimate of the copula to an estimate based on the assumed family of models. A number of options are available for choosing the scale and the distance measure for this comparison. Significance levels of the test are found by a modified bootstrap procedure. The procedure is used to check the appropriateness of a gamma or a positive stable frailty model in a set of survival data on Danish twins.

The J-shaped relation between alcohol intake and mortality is well established, whereas the nadir of the curve is not determined. Due to non-linearity of the relation, categorical alcohol variables have been used to model the relation. In Generalized Additive Models (GAM) non-linear relations can be modelled without the disadvantages of categorization and without assumptions regarding the functional form. The aim of this study was to use GAM to evaluate the relation between alcohol intake, amount and type, and mortality. The relation was investigated using data from the Copenhagen City Heart Study (11,920 participants of whom 5552 died during 20 years follow-up). Using GAM, a smooth J-shaped relation between alcohol and mortality was found. However, if non-drinkers were categorized separately there was a positive association between alcohol and mortality even for low alcohol intake. For equal total alcohol intake, men and women drinking wine or spirits had lower mortality than beer drinkers. The nadir of the relation between alcohol and mortality was sensitive to the handling of non-drinkers. When non-drinkers were categorized separately we found no indication of a beneficial influence of low alcohol intake on mortality.

Regression models for survival data are often specified from the hazard function while classical regression analysis of quantitative outcomes focuses on the mean value (possibly after suitable transformations). Methods for regression analysis of mean survival time and the related quantity, the restricted mean survival time, are reviewed and compared to a method based on pseudo-observations. Both Monte Carlo simulations and two real data sets are studied. It is concluded that while existing methods may be superior for analysis of the mean, pseudo-observations seem well suited when the restricted mean is studied.

This paper discusses the misclassification that occurs when relying solely on routine register data in family studies of disease clustering. A register study of familial aggregation of schizophrenia is used as an example. The familial aggregation is studied using a regression model for the disease in the child including the disease status of the parents as a risk factor. If all the information is found in the routine registers then the disease status of the parents is only known from the time when the register started and if this information is used unquestioningly the parents who have had the disease before this time are misclassified as disease-free. Two methods are presented to adjust for this misclassification: regression calibration and an EM-type algorithm. These methods are used in the schizophrenia example where the large effect of having a schizophrenic mother hardly shows any signs of bias due to misclassification. The methods are also studied in simulations showing that the misclassification problem increases with the disease frequency., (Copyright 2002 John Wiley & Sons, Ltd.)

The Cox regression model with a shared frailty factor allows for unobserved heterogeneity or for statistical dependence between the observed survival times. Estimation in this model when the frailties are assumed to follow a gamma distribution is reviewed, and we address the problem of obtaining variance estimates for regression coefficients, frailty parameter, and cumulative baseline hazards using the observed nonparametric information matrix. A number of examples are given comparing this approach with fully parametric inference in models with piecewise constant baseline hazards.

In survival analysis, deviations from proportional hazards may sometimes be explained by unaccounted random heterogeneity, or frailty. This paper recalls the literature on omitted covariates in survival analysis and shows in a case study how unstably frailty models might behave when asked to account for unobserved heterogeneity in standard survival analysis with no replications per heterogeneity unit. Accelerated failure time modelling seems to avoid these difficulties and also to yield easily interpretable results. We propose that it would be advantageous to upgrade the accelerated failure time approach alongside the hazard modelling approach to survival analysis.

The situation is considered where the status of the individuals under study is assessed only once during their lifetime, and as a consequence the observed time to an event of interest is either left- or right-censored. Nonparametric estimation of the time to event distribution is reviewed and a nonparametric two-sample test is proposed. The performance of the test is illustrated by simulations and by a numerical example.

This paper studies two classes of hazard-rate-based models for the mortality in a group of individuals taking normal life expectancy into account. In a multiplicative hazard model, the estimate for the relative mortality generalises the standardised mortality ratio, and the adequacy of a model with constant relative mortality can be tested using a type of total time on test statistic. In an additive hazard model, continuous-time generalisations of a "corrected" survival curve and a "normal" survival curve are obtained, and the adequacy of a model with constant excess mortality can again be tested using a type of total time on test statistic. A model including both the multiplicative hazard model and the additive hazard model is briefly considered. The use of the models is illustrated on a set of data concerning survival after operation for malignant melanoma.

We describe a bootstrap investigation of the stability of a Cox proportional hazards regression model resulting from the analysis of a clinical trial of azathioprine versus placebo in patients with primary biliary cirrhosis. We have considered stability to refer both to the choice of variables included in the model and, more importantly, to the predictive ability of the model. In stepwise Cox regression analyses of 100 bootstrap samples using 17 candidate variables, the most frequently selected variables were those selected in the original analysis, and no other important variable was identified. Thus there was no reason to doubt the model obtained in the original analysis. For each patient in the trial, bootstrap confidence intervals were constructed for the estimated probability of surviving two years. It is shown graphically that these intervals are markedly wider than those obtained from the original model.