Covariate-adjusted non-parametric survival curve estimation

Eli Lilly and Company, US Commercial Information Sciences, IN 46285, USA.
Statistics in Medicine (Impact Factor: 1.83). 05/2011; 30(11):1243-53. DOI: 10.1002/sim.4216
Source: PubMed


Kaplan-Meier survival curve estimation is a commonly used non-parametric method to evaluate survival distributions for groups of patients in the clinical trial setting. However, this method does not permit covariate adjustment which may reduce bias and increase precision. The Cox proportional hazards model is a commonly used semi-parametric method for conducting adjusted inferences and may be used to estimate covariate-adjusted survival curves. However, this model relies on the proportional hazards assumption that is often difficult to validate. Research work has been carried out to introduce a non-parametric covariate-adjusted method to estimate survival rates for certain given time intervals. We extend the non-parametric covariate-adjusted method to develop a new model to estimate the survival rates for treatment groups at any time point when an event occurs. Simulation studies are conducted to investigate the model's performance. This model is illustrated with an oncology clinical trial example.

18 Reads
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Survival is an end point of immense interest in cardiothoracic research. In observational studies, the comparison of survival between groups of patients is usually accomplished using a toolbox that includes Kaplan-Meier survival curves and the Cox model. The Cox model yields comparisons between groups adjusted for case-mix differences, whereas the Kaplan-Meier is a plot of survival over time without adjustment. During the past decade, new methods have emerged for case-mix adjustment of survival curves and are increasingly being used in cardiothoracic research. The purpose of this report is to describe, illustrate, and review several approaches to case-mix adjusted survival (or event-free) curves.
    The Annals of thoracic surgery 05/2012; 93(5):1416-25. DOI:10.1016/j.athoracsur.2011.12.094 · 3.85 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: In randomized clinical trials, improving efficiency and reducing bias due to chance imbalance in covariates among groups are always of considerable interest. The two purposes are often achieved by some type of covariate adjustment. In trials involving time-to-an-event, Kaplan-Meier and Nelson-Aalen estimators are the most popular nonparametric estimation of survival curves. However, these methods do not permit direct covariate adjustment, missing the important chance of improving efficiency and reducing bias. In this article, we propose robust, covariate adjusted analogues of the Nelson-Aalen and Kaplan-Meier estimators. The method is robust in that it does not require any additional modeling assumptions and hence the resulting estimators are again nonparametric. The robustness is achieved by taking advantage of the study design, i.e., treatments are randomized. Large-sample properties of the proposed estimators are developed, which show that the improvement in efficiency is guaranteed asymptotically. Simulation studies using reasonably small sample sizes further demonstrate the efficiency gain and the ability to reduce or remove bias resulted from chance imbalance to a large degree, e.g., more than 10-fold reduction in bias is achieved. Efficiency improvement and bias reduction are also illustrated by application to a cancer clinical trial. The proposed methods may help to resolve the tension between the need to make best use of data and the unwillingness to make additional assumptions in analyzing data from clinical trials.
    Lifetime Data Analysis 02/2014; 21(1). DOI:10.1007/s10985-014-9291-y · 0.65 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Increasingly often in epidemiologic research, associations between survival time and predictors of interest are measured by differences between distribution functions rather than hazard functions. For example, differences in percentiles of survival time, expressed in absolute time units (e.g., weeks), may complement the popular risk ratios, which are unitless measures. When analyzing time to an event of interest (e.g., death) in prospective cohort studies, the time scale can be set to start at birth or at study entry. The advantages of one time origin over the other have been thoroughly explored for the estimation of risks but not for the estimation of survival percentiles. In this paper, we analyze the use of different time scales in the estimation of survival percentiles with Laplace regression. Using this regression method, investigators can estimate percentiles of survival time over levels of an exposure of interest while adjusting for potential confounders. Our findings may help to improve modeling strategies and ease interpretation in the estimation of survival percentiles in prospective cohort studies. © The Author 2015. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail:
    American journal of epidemiology 06/2015; 182(3). DOI:10.1093/aje/kwv033 · 5.23 Impact Factor