Abstract
In this paper we propose a threshold regression (TR) model for time to event data related to subject health using a latent Ornstein–Uhlenbeck (OU) process that fails once it hits a boundary value for the first time. Baseline covariates are incorporated into the analysis using a log-link function for the initial state of the health process. The model provides clinically meaningful covariate effects and does not require the proportional hazards assumption of the commonly used Cox model. Unlike TR models based on the Wiener process, the OU model allows increments in the health process to depend on previous values and drifts toward a state of equilibrium or homeostasis, which are present in many biological applications. We also extend our model to incorporate a cure rate for applications with improper survival functions, such as time to tumor recurrence in a cancer clinical trial. Our models are applied to overall and relapse-free survival data of melanoma patients undergoing definitive surgery.
Similar content being viewed by others
References
Aalen OO, Gjessing HK (2001) Understanding the shape of the hazard rate: a process point of view. Stat Sci 16:1–22
Aalen OO, Gjessing HK (2004) Survival models based on the Ornstein–Uhlenbeck process. Lifetime Data Anal 10:407–423
Aalen OO, Borgan O, Gjessing HK (2008) Survival and event history analysis. Springer, New York
American Cancer Society (2012) Melanoma skin cancer. http://www.cancer.org/Cancer/SkinCancer-Melanoma/DetailedGuide/melanoma-skin-cancer-staging. Retrieved 26 Mar 2012
Balka J, Desmond A, McNicholas P (2009) Review and implementation of cure models based on first hitting times for Wiener processes. Lifetime Data Anal 15:147–176
Blessing W (1987) The lower brainstem and bodily homeostasis. Oxford University Press, New York
Chen M-H, Ibrahim J, Sinha D (2002) Bayesian inference for multivariate survival data with a cure fraction. J Multivar Anal 80:101–126
Chhikara RS, Folks JL (1989) The inverse Gaussian distribution: theory, methods, and applications. Marcel Dekker, New York
Chiras D (2005) Hum Biol. Jones and Bartlett Learning, Sudbury, MA
Cox DR, Miller HD (1968) The theory of stochastic processes. Chapman and Hall, London
Farewell VT (1982) The use of mixture models for the analysis of survival data with long-term survivors. Biometrics 38:1041–1046
Graf E, Schmoor C, Sauerbrei W, Schumacher M (1999) Assessment and comparison of prognostic classification schemes for survival data. Stat Med 18:2529–2545
Horrocks J, Thompson M (2004) Modelling event times with multiple outcomes using the Wiener process with drift. Lifetime Data Anal 10:29–49
Ibrahim JG, Chen M, Sinha D (2001) Bayesian survival analysis. Springer, New York
Kirkwood J, Ibrahim J, Sondak V, Richards J, Flaherty L, Ernstoff M, Smith T, Rao U, Steele M, Blum R (2000) High and low-dose interferon alfa-2b in high-risk melanoma: first analysis of intergroup trial E1690/S9111/C9190. J Clin Oncol 18:2444–2458
Kuk AYC, Chen C-H (1992) A mixture model combining logistic regression with proportional hazards regression. Biometrika 79:531–541
Lancaster T (1972) A stochastic model for the duration of a strike. J R Stat Soc Ser A 135:257–271
Larralde H (2004) A first passage time distribution for a discrete version of the Ornstein–Uhlenbeck process. J Phys A Math Gen 37:3759–3767
Lawless J, Crowder M (2004) Covariates and random effects in a gamma process model with application to degradation and failure. Lifetime Data Anal 10:213–227
Lee MLT, DeGruttola V, Schoenfeld D (2000) A model for markers and latent health status. J R Stat Soc Ser B 62:747–762
Lee MLT, Garshick E, Whitmore GA, Laden F, Hart J (2004) Assessing lung cancer risk to rail workers using a first hitting time regression model. Environmetrics 15:1–12
Lee MLT, Whitmore GA (2006) Threshold regression for survival analysis: modeling event times by a stochastic process reaching a boundary. Stat Sci 21:501–513
Lee MLT, Whitmore GA, Rosner B (2010) Threshold regression for survival data with time-varying covariates. Stat Med 29:896–905
Li Q, Shen X, Pearl D (2007) Bayesian modelling of the dynamics of hepatotoxicity. Stat Med 26:3591–3611
Li J, Lee MLT (2011) Analysis of failure time using threshold regression with semi-parametric varying coefficients. Stat Neerl 65:164–182
Myers L (1981) Survival functions induced by stochastic covariate processes. J Appl Probab 18:523–529
Park C, Padgett WJ (2005) Accelerated degradation models for failure based on geometric Brownian motion and gamma processes. Lifetime Data Anal 11:511–527
Pennell ML, Whitmore GA, Lee MLT (2010) Bayesian random-effects threshold regression with application to survival data with non proportional hazards. Biostatistics 11:111–126
Ricciardi L, Sato S (1988) First-passage-time density and moments of the Ornstein–Uhlenbeck process. J Appl Probab 25:43–57
Taylor JMG (1995) Semi-parametric estimation in failure time mixture models. Biometrics 51:899–907
Taylor J, Law N (1998) Does the covariance structure matter in longitudinal modelling for the prediction of future CD4 counts? Stat Med 17:2381–2394
Thomas M (1975) Some mean first-passage time approximations for the Ornstein–Uhlenbeck process. J Appl Probab 12:600–604
Woodbury MA, Manton KG (1977) A random-walk model of human mortality and aging. Theor Popul Biol 11:37–48
Xiao T, Whitmore G, He X and Lee M-L T (2012) Threshold regression for time-to-event analysis: the stthreg package. Stata J 12:257–283
Yamaguchi K (1992) Accelerated failure-time regression models with a regression model of surviving fraction: an application to the analysis of “permanent employment” in Japan. J Am Stat Assoc 87:284–292
Yashin A (1985) Dynamics in survival analysis: conditional Gaussian property versus Cameron–Martin formula. In: Krylov N, Liptser R, Novikov A (eds) Statistics and control of stochastic processes. Springer, New York, pp 466–475
Yashin A, Arbeev K, Akushevich I, Kulminski A, Akushevich L, Ukraintseva S (2007) Stochastic model for analysis of longitudinal data on aging and mortality. Math Biosci 208:538–551
Acknowledgments
This work was supported by the National Institute of Occupational Safety and Health [OH008649].
Author information
Authors and Affiliations
Corresponding author
Electronic supplementary material
Below is the link to the electronic supplementary material.
Rights and permissions
About this article
Cite this article
Erich, R., Pennell, M.L. Ornstein–Uhlenbeck threshold regression for time-to-event data with and without a cure fraction. Lifetime Data Anal 21, 1–19 (2015). https://doi.org/10.1007/s10985-014-9306-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10985-014-9306-8