Communications for Statistical Applications and Methods

The Korean Statistical Society (한국통계학회)
권호 표지
DOI QR코드

DOI QR Code

A case study of competing risk analysis in the presence of missing data

  • Limei Zhou (Institute for Clinical Evaluative Sciences (ICES)) ;
  • Peter C. Austin (Institute for Clinical Evaluative Sciences (ICES)) ;
  • Husam Abdel-Qadir (Institute for Clinical Evaluative Sciences (ICES))
  • Received : 2021.12.15
  • Accepted : 2022.09.22
  • Published : 2023.01.31
Download PDF
⟨ Previous Next ⟩

Abstract

Observational data with missing or incomplete data are common in biomedical research. Multiple imputation is an effective approach to handle missing data with the ability to decrease bias while increasing statistical power and efficiency. In recent years propensity score (PS) matching has been increasingly used in observational studies to estimate treatment effect as it can reduce confounding due to measured baseline covariates. In this paper, we describe in detail approaches to competing risk analysis in the setting of incomplete observational data when using PS matching. First, we used multiple imputation to impute several missing variables simultaneously, then conducted propensity-score matching to match statin-exposed patients with those unexposed. Afterwards, we assessed the effect of statin exposure on the risk of heart failure-related hospitalizations or emergency visits by estimating both relative and absolute effects. Collectively, we provided a general methodological framework to assess treatment effect in incomplete observational data. In addition, we presented a practical approach to produce overall cumulative incidence function (CIF) based on estimates from multiple imputed and PS-matched samples.

Keywords

References

  1. Abdel-Qadir H, Bobrowski D, Zhou L, Austin PC, Calvillo-Arguelles O, Amir E, Lee DS, and Thavendiranathan P (2021). Statin exposure and risk of heart failure after anthracycline-or trastuzumab-based chemotherapy for early breast cancer: A propensity score- matched cohort study, Journal of the American Heart Association, 10, e018393.
  2. Allison PD (2010). Survival Analysis Using SAS: A Practical Guide (2nd ed), SAS Press, Cary, NC. SAS Institute Inc.
  3. Austin PC (2011). An Introduction to propensity score methods for reducing the effects of confounding in observational studies, Multivariate Behavioral Research, 46, 399-424. https://doi.org/10.1080/00273171.2011.568786
  4. Austin PC, Lee DS, and Fine JP (2016). Introduction to the analysis of survival data in the presence of competing risks, Circulation, 133, 601-609. https://doi.org/10.1161/CIRCULATIONAHA.115.017719
  5. Austin PC and Fine JP (2019). Propensity-score matching with competing risks in survival analysis, Statistics in Medicine, 38, 751-777. https://doi.org/10.1002/sim.8008
  6. Austin PC, White IR, Lee DS, Buuren SV, Buuren LV, and Methodology and Statistics for the Behavioural and Social Sciences (2021). Missing data in clinical research: A tutorial on multiple imputation, Canadian Journal of Cardiology, 37, 1322-1331. https://doi.org/10.1016/j.cjca.2020.11.010
  7. Boyko EJ (2013). Observational research-opportunities and limitations, Journal of Diabetes and Its Complications, 27, 642-648. https://doi.org/10.1016/j.jdiacomp.2013.07.007
  8. Han S, Tsui KW, Zhang H, Kim GA, Lim YS, and Andrei AC (2021). Multiple imputation analysis for propensity score matching with missing causes of failure: An application tohepatocellular carcinoma data, Statistical Methods in Medical Research, 30, 2313-2328. https://doi.org/10.1177/09622802211037075
  9. He P, Eriksson F, Scheike TH, and Zhang MJ (2016). A proportional hazards regression model for the sub-distribution with covariates adjusted censoring weight for competing risks data, Scandinavian Journal of Statistics, 43, 103-122. https://doi.org/10.1111/sjos.12167
  10. Huque MH, Carlin JB, Simpson JA, and Lee KJ (2018). A comparison of multiple imputation methods for missing data in longitudinal studies, BMC Medical Research Methodology, 18, 168-184. https://doi.org/10.1186/s12874-018-0615-6
  11. Imbens G (2000). The role of the propensity score in estimating dose-response functions, Biometrika, 87, 706-710. https://doi.org/10.1093/biomet/87.3.706
  12. Jaeschke R, Guyatt G, Shannon H, Walter S, Cook D, and Heddle N (1995). Basic statistics for clinicians: 3.assessing the effects of treatment: Measures of association, Canadian Medical Association Journal, 152, 351-357.
  13. Laupacis A, Sackett DL, and Roberts RS (1988). An assessment of clinically useful measures of the consequences of treatment, New England Journal of Medicine, 318, 1728-1733. https://doi.org/10.1056/NEJM198806303182605
  14. Lee M, Dignam J, and Han J (2014). Multiple imputation methods for nonparametric inference on cumulative incidence with missing cause of failure, Statistics in Medicine, 33, 4605-4626. https://doi.org/10.1002/sim.6258
  15. Little RJ and Rubin DB (1987). Statistical Analysis with Missing Data, John Wiley & Sons, New York.
  16. Liu Y and De A (2015). Multiple imputation by fully conditional specification for dealing with missing data in a large epidemiologic study, International Journal of Statistics in Medical Research, 4, 287-295. https://doi.org/10.6000/1929-6029.2015.04.03.7
  17. Moreno-Betancur M and Latouche A (2013). Regression modeling of the cumulative incidence function with missing causes of failure using pseudo-values, Statistics in Medicine, 32, 3206-3223. https://doi.org/10.1002/sim.5755
  18. Morisot A, Bessaoud F, Landais P, Rebillard X, Tretarre B, and Daures JP (2015). Prostate cancer: Net survival and cause-specific survival rates after multiple imputation, BMC Medical Research Methodology, 15, 54-68. https://doi.org/10.1186/s12874-015-0048-4
  19. Moscovici JL and Ratitch B (2017). Combining Survival Analysis Results after Multiple Imputation of Censored Event Times, In Proceedings of PharmaSUG 2017 - Paper SP05, add city.
  20. Nguyen CD, Carlin JB, and Lee KJ (2017). Model checking in multiple imputation: An overview and case study, Emerging Themes in Epidemiology, 14, 8-20. https://doi.org/10.1186/s12982-017-0062-6
  21. Rathouz PJ (2007). Identifiability assumptions for missing covariate data in failure time regression models, Biostatistics, 8, 345-356. https://doi.org/10.1093/biostatistics/kxl014
  22. Rassen JA, Shelat AA, Franklin JM, Glynn RJ, Solomon DH, and Schneeweiss S (2013). Matching by propensity score in cohort studies with three treatment groups, Epidemiology, 24, 401-409. https://doi.org/10.1097/EDE.0b013e318289dedf
  23. Stuart EA, Azur M, Frangakis C, and Leaf P(2009). Multiple imputation with large data sets: A case study of the children's mental health initiative, American Journal of Epidemiology, 169, 1133-1139. https://doi.org/10.1093/aje/kwp026
  24. White IR, Royston P, and Wood AM (2011). Multiple imputation using chained equations: Issues and guidance for practice, Statistics in Medicine, 30, 377-399.   https://doi.org/10.1002/sim.4067