Accounting for Propensity Score Variability in IPTW Weighted Cox Proportional Hazards Regression and Risk Estimation

Michael Richard Crager

doi:doi:10.11648/j.ajam.20221005.11

| Peer-Reviewed

Accounting for Propensity Score Variability in IPTW Weighted Cox Proportional Hazards Regression and Risk Estimation

Michael Richard Crager

Published in American Journal of Applied Mathematics (Volume 10, Issue 5)

Received: 6 September 2022 Accepted: 4 October 2022 Published: 17 October 2022

Views: Downloads:

Download PDF

Share This Article

Twitter
Linked In
Facebook

Abstract

Under the assumption of no unmeasured confounders, Cox proportional hazards regression with inverse probability of treatment (IPTW) weighting based on propensity scores can be used to produce approximately unbiased estimates of treatment effect hazard ratios and event risks using observational cohorts. Often the weights are treated as fixed even though they are random variables, typically derived from a logistic regression analysis applied to the same cohort with treatment use as the outcome. Bootstrapping the entire process of weight-derivation, Cox regression analysis and estimation produces valid confidence intervals that account for the variability in the weights, but this method may be time- and resource-intensive for large cohorts. Here the delta method is used to derive large sample interval estimates of treatment effects and event risks that account for variability in the weights analytically. External time-dependent covariates, left truncation, and cohort sampling study designs are accommodated. Simulation studies show that this method provides confidence interval coverage probabilities at or above nominal level for small and moderate sample sizes. Stabilization of the weights by multiplying them by the overall treatment rate noticeably improves confidence interval coverage probabilities. Software to perform the calculations is freely available.

Published in	American Journal of Applied Mathematics (Volume 10, Issue 5)
DOI	10.11648/j.ajam.20221005.11
Page(s)	176-204
Creative Commons	This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium or format, provided the original work is properly cited.
Copyright	Copyright © The Author(s), 2022. Published by Science Publishing Group

Keywords

Cox Regression, IPTW, Propensity Score, Risk Estimation, Variability

References

[1]	Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika 76, 41–55.
[2]	Austin, P. C. (2013). The performance of different propensity score methods for estimating marginal hazard ratios. Statistics in Medicine 32, 2837–2849.
[3]	Lin, D. Y., & Wei, L. J. (1989). The robust inference for the Cox proportional hazards model. Journal of the American Statistical Association 84, 1074–1078.
[4]	Austin, P. C. (2016). Variance estimation when using inverse probability of treatment weighting (IPTW) with survival analysis. Statistics in Medicine 35, 5642–5655.
[5]	Gray, R. (2009). Weighted analyses for cohort sampling designs. Lifetime Data Analysis 15, 24–40.
[6]	Kalbfleisch, J. D., & Prentice, R. L. (2002). The Statistical Analysis of Failure Time Data, 2nd ed. Wiley, New York.
[7]	Andersen, P. K., Borgan, O., Gill, R. D., & Keiding, N. (1993). Statistical Models Based on Counting Processes, 1^st ed. Spring-Verlag, New York.
[8]	Cortese, G., & Andersen, P. K. (2009). Competing risks and time-dependent covariates. Biometrical Journal 51, 138–158.
[9]	Cain, K., C., Harlow, S. D., Little, R., J., Nan, B., Yosef, M., Taffe, J. R., & Elliott MR (2011). Bias due to left truncation and left censoring in longitudinal studies of developmental and disease processes. American Journal of Epidemiology 173, 1078–1084.
[10]	Hajage, D., Chauvet, G., Belin, L., Lafourcade, A., Tubach, F., & De Rycke, Y. (2018). Closed-form variance estimator for weighted propensity score estimators with survival outcome. Biometrical Journal 60: 1151–1163.
[11]	Lee, M.-J., & Lee S. (2022). Review and comparison of treatment effect estimators using propensity and prognostic scores. International Journal of Biostatistics. doi: 10.1515/ijb-2021-0005.
[12]	Li, F., Morgan, K. L., & Zaslavsky, A. M. (2018). Balancing covariates via propensity score weighting. Journal of the American Statistical Association 113, 390-400. https:/doi.org/10.1080/02621459.2016.1260466
[13]	Li, F., Thomas, L. E., & Li, F. (2019). Addressing extreme propensity scores via the overlap weights. American Journal of Epidemiology 188, 250–257.
[14]	Mao, H., Li, L., Wei, Y., & Shen, Y. (2018). On the propensity score weighting analysis with survival outcome: Estimands, estimation, and inference. Statistics in Medicine 36, 3745-3763.
[15]	Cheng, C., Li, F., Thomas, L. E., & Li, F. F. (2022). Addressing extreme propensity scores in estimating counterfactual survival functions via the overlap weights. American Journal of Epidemiology 191, 1140-1151.
[16]	Thomas, L. E., Li, F., & Pencina, M. J. (2020). Overlap weighting: a propensity score method that mimics attributes of a randomized clinical trial. Journal of the American Medical Association 323, 2417–2418.
[17]	Cole, S. R., & Hernán, M. A. (2008). Constructing inverse probability weights for marginal structural models. American Journal of Epidemiology 168, 656–664.
[18]	Austin, P. C., & Stuart, E. A. (2015). Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Statistics in Medicine 34, 3661–3679.
[19]	Pugh, M., Robbins, J., Lipsitz, S., & Harrington, D. (1993). Inference in the Cox Proportional Hazards Model with Missing Covariates. Technical Report 758Z. Department of Biostatistics, Harvard School of Public Health.
[20]	Therneau, T. M., & Grambsch, P. M. (2000). Modeling Survival Data: Extending the Cox Model. Springer, New York.
[21]	Czepiel, S. A. (2002). Maximum Likelihood Estimation of Logistic Regression Models: Theory and Implementation. http://czep.net
[22]	Tsiatis, A. (1981). A large sample study of the estimates for the integrated hazard function in Cox's regression model for survival data. The Annals of Statistics 9, 93–108.
[23]	Peduzzi, P., Concato, J., Feinstein, A. R., & Holford, T. R. (1995). Importance of events per independent variable in proportional hazards regression analysis. II. Accuracy and precision of regression estimates. Journal of Clinical Epidemiology 48, 1503–1510.
[24]	Brookhart, A. M., Schneeweiss, S., Rothman, K. J., Glynn, R. J., Avorn, J., & Stürmer, T. (2006). Variable selection for propensity score models. American Journal of Epidemiology 163, 1149–1156.

Cite This Article

Plain Text BibTeX RIS

APA Style

Michael Richard Crager. (2022). Accounting for Propensity Score Variability in IPTW Weighted Cox Proportional Hazards Regression and Risk Estimation. American Journal of Applied Mathematics, 10(5), 176-204. https://doi.org/10.11648/j.ajam.20221005.11

Copy | Download

ACS Style

Michael Richard Crager. Accounting for Propensity Score Variability in IPTW Weighted Cox Proportional Hazards Regression and Risk Estimation. Am. J. Appl. Math. 2022, 10(5), 176-204. doi: 10.11648/j.ajam.20221005.11

Copy | Download

AMA Style

Michael Richard Crager. Accounting for Propensity Score Variability in IPTW Weighted Cox Proportional Hazards Regression and Risk Estimation. Am J Appl Math. 2022;10(5):176-204. doi: 10.11648/j.ajam.20221005.11

Copy | Download

@article{10.11648/j.ajam.20221005.11,
  author = {Michael Richard Crager},
  title = {Accounting for Propensity Score Variability in IPTW Weighted Cox Proportional Hazards Regression and Risk Estimation},
  journal = {American Journal of Applied Mathematics},
  volume = {10},
  number = {5},
  pages = {176-204},
  doi = {10.11648/j.ajam.20221005.11},
  url = {https://doi.org/10.11648/j.ajam.20221005.11},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajam.20221005.11},
  abstract = {Under the assumption of no unmeasured confounders, Cox proportional hazards regression with inverse probability of treatment (IPTW) weighting based on propensity scores can be used to produce approximately unbiased estimates of treatment effect hazard ratios and event risks using observational cohorts. Often the weights are treated as fixed even though they are random variables, typically derived from a logistic regression analysis applied to the same cohort with treatment use as the outcome. Bootstrapping the entire process of weight-derivation, Cox regression analysis and estimation produces valid confidence intervals that account for the variability in the weights, but this method may be time- and resource-intensive for large cohorts. Here the delta method is used to derive large sample interval estimates of treatment effects and event risks that account for variability in the weights analytically. External time-dependent covariates, left truncation, and cohort sampling study designs are accommodated. Simulation studies show that this method provides confidence interval coverage probabilities at or above nominal level for small and moderate sample sizes. Stabilization of the weights by multiplying them by the overall treatment rate noticeably improves confidence interval coverage probabilities. Software to perform the calculations is freely available.},
 year = {2022}
}

Copy | Download

TY  - JOUR
T1  - Accounting for Propensity Score Variability in IPTW Weighted Cox Proportional Hazards Regression and Risk Estimation
AU  - Michael Richard Crager
Y1  - 2022/10/17
PY  - 2022
N1  - https://doi.org/10.11648/j.ajam.20221005.11
DO  - 10.11648/j.ajam.20221005.11
T2  - American Journal of Applied Mathematics
JF  - American Journal of Applied Mathematics
JO  - American Journal of Applied Mathematics
SP  - 176
EP  - 204
PB  - Science Publishing Group
SN  - 2330-006X
UR  - https://doi.org/10.11648/j.ajam.20221005.11
AB  - Under the assumption of no unmeasured confounders, Cox proportional hazards regression with inverse probability of treatment (IPTW) weighting based on propensity scores can be used to produce approximately unbiased estimates of treatment effect hazard ratios and event risks using observational cohorts. Often the weights are treated as fixed even though they are random variables, typically derived from a logistic regression analysis applied to the same cohort with treatment use as the outcome. Bootstrapping the entire process of weight-derivation, Cox regression analysis and estimation produces valid confidence intervals that account for the variability in the weights, but this method may be time- and resource-intensive for large cohorts. Here the delta method is used to derive large sample interval estimates of treatment effects and event risks that account for variability in the weights analytically. External time-dependent covariates, left truncation, and cohort sampling study designs are accommodated. Simulation studies show that this method provides confidence interval coverage probabilities at or above nominal level for small and moderate sample sizes. Stabilization of the weights by multiplying them by the overall treatment rate noticeably improves confidence interval coverage probabilities. Software to perform the calculations is freely available.
VL  - 10
IS  - 5
ER  -

Copy | Download

Author Information

Michael Richard Crager

Department of Biostatistics, Exact Sciences Corporation, Redwood City, United States

Download PDF

Sections

Plain Text BibTeX RIS

APA Style

Michael Richard Crager. (2022). Accounting for Propensity Score Variability in IPTW Weighted Cox Proportional Hazards Regression and Risk Estimation. American Journal of Applied Mathematics, 10(5), 176-204. https://doi.org/10.11648/j.ajam.20221005.11

Copy | Download

ACS Style

Michael Richard Crager. Accounting for Propensity Score Variability in IPTW Weighted Cox Proportional Hazards Regression and Risk Estimation. Am. J. Appl. Math. 2022, 10(5), 176-204. doi: 10.11648/j.ajam.20221005.11

Copy | Download

AMA Style

Michael Richard Crager. Accounting for Propensity Score Variability in IPTW Weighted Cox Proportional Hazards Regression and Risk Estimation. Am J Appl Math. 2022;10(5):176-204. doi: 10.11648/j.ajam.20221005.11

Copy | Download

@article{10.11648/j.ajam.20221005.11,
  author = {Michael Richard Crager},
  title = {Accounting for Propensity Score Variability in IPTW Weighted Cox Proportional Hazards Regression and Risk Estimation},
  journal = {American Journal of Applied Mathematics},
  volume = {10},
  number = {5},
  pages = {176-204},
  doi = {10.11648/j.ajam.20221005.11},
  url = {https://doi.org/10.11648/j.ajam.20221005.11},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajam.20221005.11},
  abstract = {Under the assumption of no unmeasured confounders, Cox proportional hazards regression with inverse probability of treatment (IPTW) weighting based on propensity scores can be used to produce approximately unbiased estimates of treatment effect hazard ratios and event risks using observational cohorts. Often the weights are treated as fixed even though they are random variables, typically derived from a logistic regression analysis applied to the same cohort with treatment use as the outcome. Bootstrapping the entire process of weight-derivation, Cox regression analysis and estimation produces valid confidence intervals that account for the variability in the weights, but this method may be time- and resource-intensive for large cohorts. Here the delta method is used to derive large sample interval estimates of treatment effects and event risks that account for variability in the weights analytically. External time-dependent covariates, left truncation, and cohort sampling study designs are accommodated. Simulation studies show that this method provides confidence interval coverage probabilities at or above nominal level for small and moderate sample sizes. Stabilization of the weights by multiplying them by the overall treatment rate noticeably improves confidence interval coverage probabilities. Software to perform the calculations is freely available.},
 year = {2022}
}

Copy | Download

TY  - JOUR
T1  - Accounting for Propensity Score Variability in IPTW Weighted Cox Proportional Hazards Regression and Risk Estimation
AU  - Michael Richard Crager
Y1  - 2022/10/17
PY  - 2022
N1  - https://doi.org/10.11648/j.ajam.20221005.11
DO  - 10.11648/j.ajam.20221005.11
T2  - American Journal of Applied Mathematics
JF  - American Journal of Applied Mathematics
JO  - American Journal of Applied Mathematics
SP  - 176
EP  - 204
PB  - Science Publishing Group
SN  - 2330-006X
UR  - https://doi.org/10.11648/j.ajam.20221005.11
AB  - Under the assumption of no unmeasured confounders, Cox proportional hazards regression with inverse probability of treatment (IPTW) weighting based on propensity scores can be used to produce approximately unbiased estimates of treatment effect hazard ratios and event risks using observational cohorts. Often the weights are treated as fixed even though they are random variables, typically derived from a logistic regression analysis applied to the same cohort with treatment use as the outcome. Bootstrapping the entire process of weight-derivation, Cox regression analysis and estimation produces valid confidence intervals that account for the variability in the weights, but this method may be time- and resource-intensive for large cohorts. Here the delta method is used to derive large sample interval estimates of treatment effects and event risks that account for variability in the weights analytically. External time-dependent covariates, left truncation, and cohort sampling study designs are accommodated. Simulation studies show that this method provides confidence interval coverage probabilities at or above nominal level for small and moderate sample sizes. Stabilization of the weights by multiplying them by the overall treatment rate noticeably improves confidence interval coverage probabilities. Software to perform the calculations is freely available.
VL  - 10
IS  - 5
ER  -

Copy | Download