Robust Estimation in AFT Models and a Covariate Adjusted Mann–Whitney Statistic for Comparing Two Sojourn Times

  • Sutirtha Chakraborty
  • Somnath DattaEmail author
Conference paper
Part of the Springer Proceedings in Mathematics & Statistics book series (PROMS, volume 175)


A Mann–Whitney statistic may be used to compare two sets of state waiting times; however, wrong conclusions may be reached if there are confounding covariates. We develop a Mann–Whitney type test statistic based on the residuals from an accelerated failure time model fitted to two groups of sojourn times with a common set of covariates. This covariate adjusted test statistics handles right censoring via the inverse probability of censoring weights. These weights were devised to improve efficiency in the sense that certain pairs in which at least one state entry time is uncensored could be compared. Extensive simulation studies were undertaken to evaluate the performance of this test. A real data illustration of our methodology is also provided.


Confounding U-statistic Two sample Sojourn time AFT model 



This research was supported in parts by grants from the US National Science Foundation (DMS-0706965) and the National Security Agency (H98230-11-1-0168). Chakraborty acknowledges fellowship support and a dissertation completion award from University of Louisville. The authors thank an anonymous reviewer for helpful comments; they also thank the Christopher and Dana Reeve Foundation and all current and past members of the NeuroRecovery Network.


  1. 1.
    Beran, R.: Nonparametric regression with randomly censored survival data. Technical report. University of California, Berkeley (1981)Google Scholar
  2. 2.
    Cox, D.R.: Regression models and life-tables. J. R. Stat. Soc. Ser. B 34, 187–202 (1972)MathSciNetzbMATHGoogle Scholar
  3. 3.
    Datta, S. Estimating the mean life time using right censored data. Stat. Methodol. 2, 65–69 (2005)Google Scholar
  4. 4.
    Datta, S., Satten, G.A.: Validity of the Aalen-Johansen estimators of stage occupation probabilities and integrated transition hazards for non-Markov models. Stat. Probab. Lett. 55, 403–411 (2001)MathSciNetCrossRefzbMATHGoogle Scholar
  5. 5.
    Datta, S., Bandyopadhyay, D., Satten, G.A.: Inverse probability of censoring weighted U-statistics for right censored data with applications. Scandinavian J. Stat. 37, 680–700 (2010)MathSciNetCrossRefzbMATHGoogle Scholar
  6. 6.
    Datta, S., Lorenz, D.J., Harkema, S.J.: A dynamic longitudinal evaluation of the utility of the Berg Balance Scale in patients with motor incomplete spinal cord injury. Arch. Phys. Med. Rehabil. 93, 1565–1573 (2012)CrossRefGoogle Scholar
  7. 7.
    Fan, J., Datta, S.: Mann-Whitney tests for comparing sojourn time distributions when the transition times are right censored. Ann. Inst. Stat. Math. 65, 149–166 (2013)MathSciNetCrossRefzbMATHGoogle Scholar
  8. 8.
    Harkema, S.J., Schmidt-Read, M., Behrman, A.L., Bratta, A., Sisto, S.A., Edgerton, V.R.: Establishing the NeuroRecovery Network: multisite rehabilitation centers that provide activity-based therapies and assessments for neurologic disorders. Arch. Phys. Med. Rehabil. 93, 1498–1507 (2011)Google Scholar
  9. 9.
    Harkema, S.J., Schmidt-Read, M., Lorenz, D., Edgerton, V.R., Behrman, A.L.: Balance and ambulation improvements in individuals with chronic incomplete spinal cord injury using locomotor training based rehabilitation. Arch. Phys. Med. Rehabil. 93, 1508–1517 (2011b)Google Scholar
  10. 10.
    Hoeffding, W.: A class of statistics with asymptotically normal distribution. Ann. Math. Stat. 19, 293–325 (1948)MathSciNetCrossRefzbMATHGoogle Scholar
  11. 11.
    Huang, Y.: Censored regression with the multistate accelerated sojourn times model. J. R. Stat. Soc. Ser. B 64, 17–29 (2002)MathSciNetCrossRefzbMATHGoogle Scholar
  12. 12.
    Mann, H.B., Whitney, D.R.: On a test of whether one of two random variables is stochastically larger than the other. Ann. Math. Stat. 18, 50–60 (1947)MathSciNetCrossRefzbMATHGoogle Scholar
  13. 13.
    Satten, G.A., Datta, S.: Marginal estimation for multistage models: waiting time distributions and competing analyses. Stat. Med. 21, 3–19 (2002)CrossRefGoogle Scholar
  14. 14.
    Satten, G.A., Datta, S., Robins, J.M.: An estimator for the survival function when data are subject to dependent censoring. Stat. Probab. Lett. 54, 397–403 (2001)MathSciNetCrossRefzbMATHGoogle Scholar
  15. 15.
    Schisterman, E., Rotnizky, A.: Estimation of the mean of a \(K\)-sample U-statistic with missing outcomes and auxiliaries. Biometrika 88, 713–725 (2001)MathSciNetCrossRefzbMATHGoogle Scholar
  16. 16.
    Serfling, R.J.: Approximation Theorems of Mathematical Statistics. Wiley, New York (1980)CrossRefzbMATHGoogle Scholar
  17. 17.
    Strawderman, R.L.: The accelerated gap times model. Biometrika 92, 647–666 (2005)MathSciNetCrossRefzbMATHGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2016

Authors and Affiliations

  1. 1.National Institute of Biomedical GenomicsKalyaniIndia
  2. 2.University of FloridaGainesvilleUSA

Personalised recommendations