Imputing missing covariate values for the Cox model

Abstract

Multiple imputation is commonly used to impute missing data, and is typically more efficient than complete cases analysis in regression analysis when covariates have missing values. Imputation may be performed using a regression model for the incomplete covariates on other covariates and, importantly, on the outcome. With a survival outcome, it is a common practice to use the event indicator D and the log of the observed event or censoring time T in the imputation model, but the rationale is not clear. We assume that the survival outcome follows a proportional hazards model given covariates X and Z . We show that a suitable model for imputing binary or Normal X is a logistic or linear regression on the event indicator D , the cumulative baseline hazard H 0 ( T ), and the other covariates Z . This result is exact in the case of a single binary covariate; in other cases, it is approximately valid for small covariate effects and/or small cumulative incidence. If we do not know H 0 ( T ), we approximate it by the Nelson–Aalen estimator of H ( T ) or estimate it by Cox regression. We compare the methods using simulation studies. We find that using log T biases covariate‐outcome associations towards the null, while the new methods have lower bias. Overall, we recommend including the event indicator and the Nelson–Aalen estimator of H ( T ) in the imputation model. Copyright © 2009 John Wiley & Sons, Ltd.

Shown for reading. Its publisher allows quoting it, not republishing it.

References 19

1. 1.

   Rubin DB. Multiple Imputation for Nonresponse in Surveys. 1987.

   Not in our library

2. 2.

   Royston P. Multiple imputation of missing values. The Stata Journal. 2004;4:227-241.

   Not in our library

3. 3.

   Royston P. Multiple imputation of missing values: update. The Stata Journal. 2005;5:188-201.

   Not in our library

4. 4.

   SAS/STAT 9.1 User's Guide. 2004.

   Not in our library

5. 5.

   van Buuren S, Oudshoorn CGM. 2000.

   Not in our library

6. 6.

   Rubin DB, Schenker N. Multiple imputation for interval estimation from simple random samples with ignorable nonresponse. Journal of the American Statistical Association. 1986;81:366-374.

   Not in our library

7. 7.

   Fay RE. When are inferences from multiple imputation valid? Proceedings of the Survey Research Methods Section of the American Statistical Association. 1992;81:227-232.

   Not in our library

8. 8.

   Schafer JL. Analysis of Incomplete Multivariate Data. 1997.

   Not in our library

9. 9.

   Moons KG, Donders RA, Stijnen T, Harrell FE. Using the outcome for imputation of missing predictor values was preferred. Journal of Clinical Epidemiology. 2006;59:1092-1101.

   Not in our library10.1016/j.jclinepi.2006.01.009

10. 10.

    van Buuren S, Boshuizen HC, Knook DL. Multiple imputation of missing blood pressure covariates in survival analysis. Statistics in Medicine. 1999;18:681-694.

    Not in our library10.1002/(sici)1097-0258(19990330)18:6<681::aid-sim71>3.0.co;2-r

11. 11.

    Clark TG, Altman DG. Developing a prognostic model in the presence of missing data: an ovarian cancer case study. Journal of Clinical Epidemiology. 2003;56:28-37.

    Not in our library10.1016/s0895-4356(02)00539-5

12. 12.

    Barzi F, Woodward M. Imputations of missing values in practice: results from imputations of serum cholesterol in 28 cohort studies. American Journal of Epidemiology. 2004;160:34-45.

    Not in our library10.1093/aje/kwh175

13. 13.

    Schafer JL. 2008.

    Not in our library

14. 14.

    Little RJA, Rubin DB. Statistical Analysis with Missing Data. 2002.

    Not in our library

15. 15.

    Interferon-alpha and survival in metastatic renal carcinoma: early results of a randomised controlled trial. Lancet. 1999;353:14-17.

    Not in our library

16. 16.

    Royston P, Sauerbrei W. Multivariable Model-building: A Pragmatic Approach to Regression Analysis Based on Fractional Polynomials for Modelling Continuous Variables. 2008.

    Not in our library

17. 17.

    Li KH, Raghunathan TE, Rubin DB. Large-sample significance levels from multiply imputed data using mopment-based statistics and an F reference distribution. Journal of the American Statistical Association. 1991;86:1065-1073.

    Not in our library

18. 18.

    Collett D. Modelling Survival Data in Medical Research. 1994.

    Not in our library

19. 19.

    Little RJA. Missing-data adjustments in large surveys. Journal of Business and Economic Statistics. 1988;6:287-296.

    Not in our library

Cited by papers we hold 14

• Association of social isolation, loneliness and genetic risk with incidence of dementia: UK Biobank Cohort Study · 2022
  Not checked
• Physical Fitness and Dyslipidemia Among Japanese: A Cohort Study From the Niigata Wellness Study · 2021
  Not checked
• Serum neurofilament light and tau as prognostic markers for all-cause mortality in the elderly general population—an analysis from the MEMO study · 2021
  Not checked
• Missing Data in Clinical Research: A Tutorial on Multiple Imputation · 2021
  No notice
• Accounting for missing data in statistical analyses: multiple imputation is not always the answer · 2019
  No notice
• Using simulation studies to evaluate statistical methods · 2019
  No notice
• Tobacco smoking and alcohol drinking at diagnosis of head and neck cancer and all‐cause mortality: Results from head and neck 5000, a prospective observational cohort of people with head and neck cancer · 2018
  No notice
• Contribution of risk factors to excess mortality in isolated and lonely individuals: an analysis of data from the UK Biobank cohort study · 2017
  No notice
• Multiple imputation of covariates by fully conditional specification: Accommodating the substantive model · 2014
  No notice
• Physical capability in mid-life and survival over 13 years of follow-up: British birth cohort study · 2014
  No notice
• Tuning multiple imputation by predictive mean matching and local residual draws · 2014
  No notice
• Reporting recommendations for tumor marker prognostic studies (REMARK): explanation and elaboration · 2012
  No notice
• Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK): Explanation and Elaboration · 2012
  No notice
• Comparison of imputation methods for handling missing covariate data when fitting a Cox proportional hazards model: a resampling study · 2010
  No notice

Reads

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.