Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Dec;31(12):2352-2367.
doi: 10.1177/09622802221122406. Epub 2022 Sep 13.

Flexible semiparametric mode regression for time-to-event data

Alexander Seipp  1 , Verena Uslar  2 , Dirk Weyhe  2 , Antje Timmer  1 , Fabian Otto-Sobotka  1
Affiliations

Flexible semiparametric mode regression for time-to-event data

Alexander Seipp et al. Stat Methods Med Res. 2022 Dec.

Abstract

The distribution of time-to-event outcomes is usually right-skewed. While for symmetric and moderately skewed data the mean and median are appropriate location measures, the mode is preferable for heavily skewed data as it better represents the center of the distribution. Mode regression has been introduced for uncensored data to model the relationship between covariates and the mode of the outcome. Starting from nonparametric kernel density based mode regression, we examine the use of inverse probability of censoring weights to extend mode regression to handle right-censored data. We add a semiparametric predictor to add further flexibility to the model and we construct a pseudo Akaike's information criterion to select the bandwidth and smoothing parameters. We use simulations to evaluate the performance of our proposed approach. We demonstrate the benefit of adding mode regression to one's toolbox for analyzing survival data on a pancreatic cancer data set from a prospectively maintained cancer registry.

Keywords: Iteratively weighted least squares; P-splines; inverse probability of censoring; inverse probability weights; pancreatic cancer.

PubMed Disclaimer

Conflict of interest statement

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Figures

Figure 1.
Figure 1.
Kaplan-Meier estimates stratified by chemotherapy with pointwise confidence interval (left) and smoothed density estimates (right) of survival times.
Figure 2.
Figure 2.
Boxplots of patients age seperated by chemotherapy and R status.
Figure 3.
Figure 3.
Estimate (solid line) and 95% confidence interval (dashed lines) for the relationship between age and mode of the log survival time.
Figure 4.
Figure 4.
Further simulation results on the convergence of the algorithm. (a) Computation time of the algorithm in dependence of sample size, with either linear effects or with P-splines. Points and lines represent the median, the shaded areas are intervals from Q1 to Q3. (b) Progression of objective values in three cases where the hyperparameter optimization did not converge within 200 optimization steps.
See this image and copyright information in PMC

References

    1. Hyndman RJ. Computing and graphing highest density regions. Am Stat 1995; 50: 120–126.
    1. Robert Koch-Institut and Gesellschaft der epidemiologischen Krebsregister in Deutschland e.V. Krebs in Deutschland 2015/2016 [Cancer in Germany 2015/2016], 2019.DOI: 10.25646/5977.2. - DOI
    1. Sasaki H, Ono Y, Sugiyama M. Modal regression via direct log-density derivative estimation. In Hirose A, Ozawa S, Doya K et al. (eds.) Neural Information Processing. Springer International Publishing. ISBN 978-3-319-46672-9, pp. 108–116.
    1. Ota H, Kato K, Hara S. Quantile regression approach to conditional mode estimation. Electron J Stat 2019; 13: 3120–3160.
    1. Chen YC. Modal regression using kernel density estimation: a review. WIREs Comput Stat 2018; 10: e1431.

Publication types