Multiple imputation using chained equations for missing data in survival models applied to multidrug-resistant tuberculosis and HIV data

Sizwe Vincent Mbona; Principal Ndlovu; Henry Mwambi; Shaun Ramroop

doi:10.4081/jphia.2023.2388

Authors

Sizwe Vincent Mbona
sizwem@dut.ac.za
Department of Statistics, Durban University of Technology, Durban, South Africa. https://orcid.org/0000-0002-8077-3833
Principal Ndlovu Department of Statistics, University of South Africa, Pretoria, South Africa. https://orcid.org/0000-0003-4254-0745
Henry Mwambi School of Mathematics, Statistics and Computer Science, University of KwaZulu-Natal, Pietermaritzburg, South Africa.
Shaun Ramroop School of Mathematics, Statistics and Computer Science, University of KwaZulu-Natal, Pietermaritzburg, South Africa. https://orcid.org/0000-0001-9305-6967

Background. Missing data are a prevalent problem in almost all types of data analyses, such as survival data analysis. Objective. To evaluate the performance of multivariable imputation via chained equations in determining the factors that affect the survival of multidrug-resistant-tuberculosis (MDR-TB) and HIV-coinfected patients in KwaZulu-Natal. Materials and Methods. Secondary data from 1542 multidrug-resistant tuberculosis patients were used in this study. First, data from patients with some missing observations were deleted from the original data set to obtain the complete case (CC) data set. Second, missing observations in the original data set were imputed 15 times to obtain complete data sets using a multivariable imputation case (MIC). The Cox regression model was fitted to both the CC and MIC data, and the results were compared using the model goodness of fit criteria [likelihood ratio tests, Akaike information criterion (AIC), and Bayesian Information Criterion (BIC)]. Results. The Cox regression model fitted the MIC data set better (likelihood ratio test statistic =76.88 on 10 df with P<0.01, AIC =1040.90, and BIC =1099.65) than the CC data set (likelihood ratio test statistic =42.68 on 10 df with P<0.01, AIC =1186.05 and BIC =1228.47). Variables that were insignificant when the model was fitted to the CC data set became significant when the model was fitted to the MIC data set. Conclusion. Correcting missing data using multiple imputation techniques for the MDR-TB problem is recommended. This approach led to better estimates and more power in the model.

Loveday M, Padayatchi N, Wallengren K, Roberts J, Brust JC, Ngozo J, Master I, Voce A. Association between health systems performance and treatment outcomes in patients co-infected with MDR-TB and HIV in KwaZulu-Natal, South Africa: implications for TB programmes. PLoS One. 2014 Apr 9;9(4):e94016. DOI: https://doi.org/10.1371/journal.pone.0094016

Rubin DB. Inference and missing data. Biometrika. 1976 Dec 1;63(3):581-92. DOI: https://doi.org/10.1093/biomet/63.3.581

Allison PD. Multiple imputation for missing data: A cautionary tale. Sociological methods & research. 2000 Feb;28(3):301-9. DOI: https://doi.org/10.1177/0049124100028003003

Schafer JL. Multiple imputation in multivariate problems when the imputation and analysis models differ. Statistica neerlandica. 2003 Feb;57(1):19-35. DOI: https://doi.org/10.1111/1467-9574.00218

Altman DG, Bland JM. Missing data. Bmj. 2007 Feb 22;334(7590):424-. DOI: https://doi.org/10.1136/bmj.38977.682025.2C

Baneshi MR, Talei AR. Impact of imputation of missing data on estimation of survival rates: an example in breast cancer. 2010.

Eekhout I, de Boer RM, Twisk JW, de Vet HC, Heymans MW. Missing data: a systematic review of how they are reported and handled. Epidemiology. 2012 Sep 1;23(5):729-32. DOI: https://doi.org/10.1097/EDE.0b013e3182576cdb

Acock AC. Working with missing values. Journal of Marriage and family. 2005 Nov;67(4):1012-28. DOI: https://doi.org/10.1111/j.1741-3737.2005.00191.x

Little RJ, Rubin DB. Statistical inference with missing data. 2002. DOI: https://doi.org/10.1002/9781119013563

Peugh JL, Enders CK. Missing data in educational research: A review of reporting practices and suggestions for improvement. Review of educational research. 2004 Dec;74(4):525-56. DOI: https://doi.org/10.3102/00346543074004525

Rubin DB. Multiple Imputation for Non-response in Surveys John Wiley. New York. 1987. DOI: https://doi.org/10.1002/9780470316696

Dempster AP, Laird NM, Rubin DB. Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society: Series B (Methodological). 1977 Sep;39(1):1-22. DOI: https://doi.org/10.1111/j.2517-6161.1977.tb01600.x

Enders CK. A primer on maximum likelihood algorithms available for use with missing data. Structural Equation Modeling. 2001 Jan 1;8(1):128-41. DOI: https://doi.org/10.1207/S15328007SEM0801_7

Rubin DB. Multiple imputations in sample surveys-a phenomenological Bayesian approach to nonresponse. In Proceedings of the survey research methods section of the American Statistical Association 1978 Jan 2 (Vol. 1, pp. 20-34). Alexandria, VA, USA: American Statistical Association.

Rubin DB. Multiple imputation after 18+ years. Journal of the American statistical Association. 1996 Jun 1;91(434):473-89. DOI: https://doi.org/10.1080/01621459.1996.10476908

Schafer JL, Graham JW. Missing data: our view of the state of the art. Psychological methods. 2002 Jun;7(2):147. DOI: https://doi.org/10.1037/1082-989X.7.2.147

Patrician PA. Multiple imputation for missing data. Research in nursing & health. 2002 Feb;25(1):76-84. DOI: https://doi.org/10.1002/nur.10015

Van Buuren S, Brand JP, Groothuis-Oudshoorn CG, Rubin DB. Fully conditional specification in multivariate imputation. Journal of statistical computation and simulation. 2006 Dec 1;76(12):1049-64. DOI: https://doi.org/10.1080/10629360600810434

Van Buuren S, Groothuis-Oudshoorn K. mice: Multivariate imputation by chained equations in R. Journal of statistical software. 2011 Dec 12;45:1-67. DOI: https://doi.org/10.18637/jss.v045.i03

Lee KJ, Carlin JB. Multiple imputation for missing data: fully conditional specification versus multivariate normal imputation. American journal of epidemiology. 2010 Mar 1;171(5):624-32. DOI: https://doi.org/10.1093/aje/kwp425

Lee KJ, Carlin JB. Recovery of information from multiple imputation: a simulation study. Emerging themes in epidemiology. 2012 Dec;9(1):1-0. DOI: https://doi.org/10.1186/1742-7622-9-3

Stuart EA, Azur M, Frangakis C, Leaf P. Multiple imputation with large data sets: a case study of the Children's Mental Health Initiative. American journal of epidemiology. 2009 May 1;169(9):1133-9. DOI: https://doi.org/10.1093/aje/kwp026

He Y. Missing data analysis using multiple imputation: getting to the heart of the matter. Circulation: Cardiovascular Quality and Outcomes. 2010 Jan;3(1):98-105. DOI: https://doi.org/10.1161/CIRCOUTCOMES.109.875658

Schenker N, Raghunathan TE, Chiu PL, Makuc DM, Zhang G, Cohen AJ. Multiple imputation of missing income data in the National Health Interview Survey. Journal of the American Statistical Association. 2006 Sep 1;101(475):924-33. DOI: https://doi.org/10.1198/016214505000001375

Royston P, White IR. Multiple imputation by chained equations (MICE): implementation in Stata. Journal of statistical software. 2011 Dec 12;45:1-20. DOI: https://doi.org/10.18637/jss.v045.i04

White IR, Royston P, Wood AM. Multiple imputation using chained equations: issues and guidance for practice. Statistics in medicine. 2011 Feb 20;30(4):377-99. DOI: https://doi.org/10.1002/sim.4067

Marchenko Y. Chained equations and more in multiple imputation in Stata 12. In 2011 Italian Stata Users Group Meeting 2011 Sep 26.

Klein JP, Moeschberger ML. Survival analysis: techniques for censored and truncated data. New York: Springer; 2003 Feb. DOI: https://doi.org/10.1007/b97377

Etikan I, Babatope G. Survival Analysis: A Major Decision Technique in Healthcare Practices. International Journal of Science and Research Methodology. 2018;8(4):121-35.

Bengura P. Identification of factors affecting the survival lifetime of HIV+ terminal patients in Albert Luthuli municipality of South Africa. University of South Africa (South Africa); 2020. DOI: https://doi.org/10.21203/rs.3.rs-15949/v1

Kleinbaum DG, Klein M. Survival analysis: a self-learning text. New York: Springer; 2012. DOI: https://doi.org/10.1007/978-1-4419-6646-9

Lemeshow S, May S, Hosmer Jr DW. Applied survival analysis: regression modeling of time-to-event data. John Wiley & Sons; 2011 Sep 23.

Cleves MA, Gould WW, Gutierrez RG, Marchenko YV. Competing risks. An Introduction to Survival Analysis Using Stata. 2010.

Zhang HH. Checking proportionality for Cox's regression model (Master's thesis). 2015. Retrieved May 23, 2018 from https://www.duo.uio.no/bitstream/handle/10852/45324/HuiHongZhang_thesis.pdf?sequence=1&isAllowed=y

Xu R, Vaida F, Harrington DP. Using profile likelihood for semiparametric model selection with application to proportional hazards mixed models. Statistica Sinica. 2009 Apr;19(2):819.

Van Buuren S. Flexible imputation of missing data. CRC press; 2018 Jul 17. DOI: https://doi.org/10.1201/9780429492259

Yang S. Flexible Imputation of Missing Data: Boca Raton, FL: Chapman & Hall/CRC Press, 2018, xxvii+ 415 pp., $91.95 (H), ISBN: 978-1-13-858831-8. Journal of the American Statistical Association. 2019 Jul;114(527).

Little RJ, Wang Y. Pattern-mixture models for multivariate incomplete data with covariates. Biometrics. 1996 Mar 1:98-111. DOI: https://doi.org/10.2307/2533148

Azur MJ, Stuart EA, Frangakis C, Leaf PJ. Multiple imputation by chained equations: what is it and how does it work?. International journal of methods in psychiatric research. 2011 Mar;20(1):40-9. DOI: https://doi.org/10.1002/mpr.329

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 Jul 1;160(1):34-45. DOI: https://doi.org/10.1093/aje/kwh175

Chen C, Zhu T, Wang Z, Peng H, Kong W, Zhou Y, Shao Y, Zhu L, Lu W. High latent TB infection rate and associated risk factors in the eastern China of low TB incidence. PLOS one. 2015 Oct 27;10(10):e0141511. DOI: https://doi.org/10.1371/journal.pone.0141511

Ambler G, Omar RZ, Royston P. A comparison of imputation techniques for handling missing predictor values in a risk model with a binary outcome. Statistical methods in medical research. 2007 Jun;16(3):277-98. DOI: https://doi.org/10.1177/0962280206074466

Musil CM, Warner CB, Yobas PK, Jones SL. A comparison of imputation techniques for handling missing data. Western journal of nursing research. 2002 Nov;24(7):815-29. DOI: https://doi.org/10.1177/019394502762477004

Freund Y, Schapire RE. Experiments with a new boosting algorithm. Inicml 1996 Jul 3 (Vol. 96, pp. 148-156).

Friedman J, Hastie T, Tibshirani R. Additive logistic regression: a statistical view of boosting (with discussion and a rejoinder by the authors). The annals of statistics. 2000 Apr;28(2):337-407. DOI: https://doi.org/10.1214/aos/1016218223

Chatterjee S, Price B. Selection of variables in a regression equation. Regression analysis by example. 1977:201-3.

Mbona, S. V., Ndlovu, P., Mwambi, H., & Ramroop, S. (2023). Multiple imputation using chained equations for missing data in survival models applied to multidrug-resistant tuberculosis and HIV data. Journal of Public Health in Africa, 14(8). https://doi.org/10.4081/jphia.2023.2388