Imputation approaches for potential outcomes in causal inference
Open accessThe DOI matches its Crossref record; 47 of the 47 references it cites were found in a registry or an index.
Abstract
BACKGROUND: The fundamental problem of causal inference is one of missing data, and specifically of missing potential outcomes: if potential outcomes were fully observed, then causal inference could be made trivially. Though often not discussed explicitly in the epidemiological literature, the connections between causal inference and missing data can provide additional intuition. METHODS: We demonstrate how we can approach causal inference in ways similar to how we address all problems of missing data, using multiple imputation and the parametric g-formula. RESULTS: We explain and demonstrate the use of these methods in example data, and discuss implications for more traditional approaches to causal inference. CONCLUSIONS: Though there are advantages and disadvantages to both multiple imputation and g-formula approaches, epidemiologists can benefit from thinking about their causal inference problems as problems of missing data, as such perspectives may lend new and clarifying insights to their analyses.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“While multiple imputation has mainly been applied to impute missing covariates, it can also be used to infer counterfactual responses (Westreich et al., 2015).”
From Customer price sensitivities in competitive insurance markets (2022) · cites this paper as Westreich et al., 2015
Citations
Cites
Multiple Imputation for Nonresponse in Surveys
Rubin, D. B. (1987) · Wiley Series in Probability and Statistics · doi:10.1002/9780470316696 · Cited by 20,701
Statistical Analysis With Missing Data
Lahiff, M., Little, R. J. A., & Rubin, D. B. (1989) · Journal of the American Statistical Association · doi:10.2307/2289883 · Cited by 17,265
RUBIN, D. B. (1976) · Biometrika · doi:10.1093/biomet/63.3.581 · Cited by 9,925
Multiple imputation using chained equations: Issues and guidance for practice
White, I. R., Royston, P., & Wood, A. M. (2011) · Statistics in Medicine · doi:10.1002/sim.4067 · Cited by 9,792
Estimating causal effects of treatments in randomized and nonrandomized studies
Rubin, D. B. (1974) · Journal of Educational Psychology · doi:10.1037/h0037350 · Cited by 9,436 · cited 1×
“The fundamental problem of causal inference, and thus of analytical epidemiology, is therefore one of missing data, or the inability to observe all but one potential outcome.(4,5) Yet, when epidemiologists analyse data for the purposes of causal inference, in the absence of explicitly missing data they typically do no…”
BackgroundStatistical Analysis with Missing Data
Little, R. J. A., & Rubin, D. B. (2002) · Wiley Series in Probability and Statistics · doi:10.1002/9781119013563 · Cited by 8,473
Marginal Structural Models and Causal Inference in Epidemiology
Robins, J. M., Hernán, M. Á., & Brumback, B. (2000) · Epidemiology · doi:10.1097/00001648-200009000-00011 · Cited by 5,581
A generalization of sampling without replacement from a finite universe
Horvitz, D. G., & Thompson, D. J. (1952) · Journal of the American Statistical Association · doi:10.1080/01621459.1952.10483446 · Cited by 4,293
Schafer, J. L. (1999) · Statistical Methods in Medical Research · doi:10.1177/096228029900800102 · Cited by 3,949
A Generalization of Sampling Without Replacement From a Finite Universe
Horvitz, D. G., & Thompson, D. J. (1952) · Journal of the American Statistical Association · doi:10.2307/2280784 · Cited by 3,011
Cited by
Stacked generalization: an introduction to super learning
Naimi, A. I., & Balzer, L. B. (2018) · European Journal of Epidemiology · doi:10.1007/s10654-018-0390-z · Cited by 430
“Indeed, some researchers have advocated for the use of data-adaptive methods, including Super Learner, for effect estimation via singly-robust methods, depending on estimation of either the conditional mean outcome or the propensity score [26–32].”
Challenges in Obtaining Valid Causal Effect Estimates With Machine Learning Algorithms
Naimi, A. I., Mishler, A., & Kennedy, E. H. (2021) · American Journal of Epidemiology · doi:10.1093/aje/kwab201 · Cited by 144
“3,4,42 But, as we have shown, for singly robust estimators this protection may not always be worth the price.”
Background“Truly free consent”? Clarifying the nature of police legitimacy using causal mediation analysis
Pósch, K., Jackson, J., Bradford, B., & MacQueen, S. (2020) · Journal of Experimental Criminology · doi:10.1007/s11292-020-09426-x · Cited by 124
Reflection on modern methods: when worlds collide—prediction, machine learning and causal inference
Blakely, T., Lynch, J., Simons, K., Bentley, R., & Rose, S. (2021) · International Journal of Epidemiology · doi:10.1093/ije/dyz132 · Cited by 98
“3,25,29 Given that, for this simple example at least, we could use a standard regression model to estimate an effect size for the exposure (cid:2) outcome association to undertake prediction of potential outcomes, why bother?”
BackgroundMethodLearning from data with structured missingness
Mitra, R., McGough, S. F., Chakraborti, T., Holmes, C. C., Copping, R., Hagenbuch, N., Biedermann, S., Noonan, J., Lehmann, B., Shenvi, A., et al. (2023) · Nature Machine Intelligence · doi:10.1038/s42256-022-00596-z · Cited by 139
“Extended to SM, there may be much to learn about how to estimate structures of missingness via causal inference methods such as inverse probability weighting and imputation (67; 68; 69).”
Parra, C. O., Daniel, R., & Bartlett, J. W. (2022) · Statistics in Biopharmaceutical Research · doi:10.1080/19466315.2022.2081599 · Cited by 63
“The equivalence of imputation approaches and G-formula was previously discussed by Westreich et al. (2015) and Qu et al. (2020).”
MethodMissing Outcome Data in Epidemiologic Studies
Cole, S. R., Zivich, P. N., Edwards, J. K., Ross, R. K., Shook‐Sa, B. E., Price, J. T., & Stringer, J. S. A. (2022) · American Journal of Epidemiology · doi:10.1093/aje/kwac179 · Cited by 45
Self-citationThe Causal Cookbook: Recipes for Propensity Scores, G-Computation, and Doubly Robust Standardization
Chatton, A., & Rohrer, J. M. (2024) · Advances in Methods and Practices in Psychological Science · doi:10.1177/25152459241236149 · Cited by 58
“…predicting both counterfactual worlds for all individuals in Step 3, we may instead predict only the unobserved outcome (e.g., the outcome without action for those individuals who did in fact receive the action) and keep the observed outcomes untouched to improve accuracy (Westreich et al., 2015).”
BackgroundResearch, application, and challenges of causal inference in industrial fault diagnosis: A survey
Li, B., Li, Q., Du, T., Liu, D., Yang, Q., Chen, T., Xiong, J., Peng, B., Ren, J., & Zhao, J. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.111376 · Cited by 43
Valcarce-Torrente, M., Sanchis, V. J., Gallardo, L., Fernández, J. G., & Planas, A. (2021) · International Journal of Environmental Research and Public Health · doi:10.3390/ijerph181910393 · Cited by 33
“comment on the importance of application use factors and how they can influence the user’s sports habits [29,30,37].”
BackgroundCustomer price sensitivities in competitive insurance markets
Verschuren, R. M. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2022.117133 · Cited by 19
“While multiple imputation has mainly been applied to impute missing covariates, it can also be used to infer counterfactual responses (Westreich et al., 2015).”
Mikkonen, J., Remes, H. M., Moustgaard, H., & Martikainen, P. (2020) · Social Forces · doi:10.1093/sf/soaa048 · Cited by 31
“G-computation can be considered a generalization of direct standardization and an approach to impute missing potential outcomes in the spirit of the counterfactual mediation framework (Westreich et al. 2015).”
BackgroundMethodMissingness in the Setting of Competing Risks: from Missing Values to Missing Potential Outcomes
Lau, B. M., & Lesko, C. R. (2018) · Current Epidemiology Reports · doi:10.1007/s40471-018-0142-3 · Cited by 16
Assessing heterogeneous effects and their determinants via estimation of potential outcomes
Nilsson, A., Bonander, C., Strömberg, U., & Björk, J. (2019) · European Journal of Epidemiology · doi:10.1007/s10654-019-00551-0 · Cited by 22
Long-term causal effects of far-right terrorism in New Zealand
Bulbulia, J. A., Afzali, M. U., Yogeeswaran, K., & Sibley, C. G. (2023) · PNAS Nexus · doi:10.1093/pnasnexus/pgad242 · Cited by 16
“We emphasize that missing values must be imputed separately for each exposure condition because the counterfactuals outcomes are not jointly observed (34).”
InfluentialKadengye, D. T., Iddi, S., Hunter, L. A., & McCoy, S. I. (2019) · Prevention Science · doi:10.1007/s11121-019-01010-8 · Cited by 14
Pósch, K. (2019) · Sociological Methods & Research · doi:10.1177/0049124119826159 · Cited by 23
Sensitivity analyses for the principal ignorability assumption using multiple imputation
Wang, C., Zhang, Y., Mealli, F., & Bornkamp, B. (2022) · Pharmaceutical Statistics · doi:10.1002/pst.2260 · Cited by 11
Causal Deep Learning: Encouraging Impact on Real-world Problems Through Causality
Berrevoets, J., Kacprzyk, K., Qian, Z., & van der Schaar, M. (2024) · Foundations and Trends® in Signal Processing · doi:10.1561/2000000123 · Cited by 54
Nonparametric Double RobustnessPreprint
Naimi, A. I., & Kennedy, E. H. (2017) · arXiv (Cornell University) · Cited by 10
Sensitivity Analyses for Misclassification of Cause of Death in the Parametric G-Formula
Edwards, J. K., Cole, S. R., Moore, R. D., Mathews, W. C., Kitahata, M. M., & Eron, J. J. (2018) · American Journal of Epidemiology · doi:10.1093/aje/kwy028 · Cited by 8
“The true potential outcomes are unobserved (12, 13).”
BackgroundSelf-citationGeneralization and Invariances in the Presence of Unobserved Confounding.Preprint
Bellot, A., & van der Schaar, M. (2020) · arXiv (Cornell University) · Cited by 13
Kawahara, T., Shiba, K., & Tsuchiya, A. (2022) · World Neurosurgery · doi:10.1016/j.wneu.2021.09.141 · Cited by 8
Gaber, C. E., Edwards, J. K., Lund, J. L., Peery, A. F., Richardson, D. B., & Kinlaw, A. C. (2023) · American Journal of Epidemiology · doi:10.1093/aje/kwad031 · Cited by 8
Self-citationStudy design approaches for future active-controlled HIV prevention trials
Donnell, D. J., Kansiime, S., Glidden, D. V., Luedtke, A. R., Gilbert, P. B., Gao, F., & Janes, H. (2023) · Statistical Communications in Infectious Diseases · doi:10.1515/scid-2023-0002 · Cited by 14
“Propensity scores and more sophisticated causal inference approaches based on a potential outcomes framework are also statistical methods appropriate for this setting [47, 58, 59].”
BackgroundInfluencia de una app en la adherencia a la práctica deportiva: protocolo de estudio
Valcarce-Torrente, M., & Rico, C. D. (2018) · e-Motion Revista de Educación Motricidad e Investigación · doi:10.33776/remo.v0i11.3416 · Cited by 12
Nabijiang, A. (2020) · Louisiana State University · doi:10.31390/gradschool_theses.5132 · Cited by 7
Takahashi, M. (2021) · Communications in Statistics - Simulation and Computation · doi:10.1080/03610918.2021.1960374 · Cited by 8
Multiply robust estimation of the average treatment effect with missing outcomes
Wei, K., Qin, G., Zhang, J., & Sui, X. (2022) · Journal of Statistical Computation and Simulation · doi:10.1080/00949655.2022.2143501 · Cited by 8
“Imputation can replace the potential outcomes [8,9] and the missing data [10,11] with the fitted values in regression, then the valid inference can be obtained through the complete-case analysis.”
BackgroundModel-based standardization using multiple imputation
Remiro‐Azócar, A., Heath, A., & Baio, G. (2024) · BMC Medical Research Methodology · doi:10.1186/s12874-024-02157-x · Cited by 10
“Despite the close relationships between the methodologies, these largely have been developed separately, with some exceptions [31].”
High breakdown point robust estimators with missing data
Statti, F., Sued, M., & Yohai, V. J. (2017) · Communication in Statistics- Theory and Methods · doi:10.1080/03610926.2017.1388396 · Cited by 5
“…E[y( j)], j = 0, 1 requires to usemissing data techniques, since y( j) is only available for those individuals with treatment level T = t j, and is missing for those with T = t j. Westreich et al. (2015) present an interesting discussion on the connections between causal inference and missing data.”
BackgroundBerrevoets, J., Kacprzyk, K., Qian, Z., & van der Schaar, M. (2023) · arXiv · doi:10.48550/arxiv.2303.02186 · Cited by 8
Watanabe, H., & Maruyama, T. (2024) · Journal of Choice Modelling · doi:10.1016/j.jocm.2024.100491 · Cited by 7
Kim, M., Wang, C., & Xue, X. (2018) · Statistics in Medicine · doi:10.1002/sim.7999 · Cited by 5
Nunes, A. P., Zhao, D., Jesdale, W. M., & Lapane, K. L. (2021) · BMC Medical Research Methodology · doi:10.1186/s12874-021-01327-5 · Cited by 3
Xuan, J., Mt‐Isa, S., Latimer, N. R., Bell‐Gorrod, H., Malbecq, W., Vandormael, K., Yorke-Edwards, V., & White, I. R. (2025) · Statistical Methods in Medical Research · doi:10.1177/09622802251387456 · Cited by 4
“Notably, under model mis-specification, IPCW with unstabilised weights yields slightly smaller RMSE compared to its stabilised counterparts and this is more evident in Scenarios 8, 9, 11 and 12 where the prevalence of the ICE is high.”
Pan, Z., Qin, Y., Bai, W., He, Q., Yin, X., & He, J. (2024) · BMC Medical Research Methodology · doi:10.1186/s12874-024-02321-3 · Cited by 3
“Furthermore, the estimate of MI-MRMC corresponds to the randomized/enrolled population, aligning with the ICH E9(R1) framework [49] and principles of causal inference [50].”
MethodApplying the Principal Stratum Strategy in Equivalence Trials: A Case Study
Sepin, J., Debray, T. P. A., Wei, W., Ebbers, H. C., Fernández‐Mendívil, C., & Mitroiu, M. (2025) · Pharmaceutical Statistics · doi:10.1002/pst.70008 · Cited by 4
“It therefore bears many similarities with the missing at random (MAR) assumption in the missing data literature [27, 28], or the ignorability or no unmea-sured confounder assumption in observational settings [15], chap.”
Causal Inference, Causal Effect Estimation, and Systematic Error
Westreich, D. J. (2019) · Oxford University Press eBooks · doi:10.1093/oso/9780190665760.003.0004 · Cited by 5
Self-citationLi, Z., Li, X., Yi, X., Li, T. X., Huang, X., Ren, X., Ma, T., Li, K., Guo, H., Chen, S., et al. (2022) · Frontiers in Oncology · doi:10.3389/fonc.2022.838840 · Cited by 2
“Although the causal inference method could explicitly unveil the effect modification of different types of CMM on patient outcomes and identify vulnerable groups at the population level (32), comprehensive nomograms could provide estimation of the cumulative incidence of death at an individual patient level (33).”
MethodWatanabe, H., & Maruyama, T. (2022) · SSRN Electronic Journal · doi:10.2139/ssrn.4086629 · Cited by 2
Kawahara, T., Shinozaki, T., & Matsuyama, Y. (2020) · BMC Medical Research Methodology · doi:10.1186/s12874-020-01087-8 · Cited by 1
Multiple Imputation of a Continuous Outcome with Fully Observed Predictors Using TabPFN
Sepin, J. (2026) · Stats · doi:10.3390/stats9020038 · Cited by 2
“We follow the approach outlined in Westreich et al. [47], leveraging MI approaches to estimate the average causal effect (ACE) of smoking cessation on weight gain, where Y a = 1 and Y a = 0 denote the potential weight gains had all individuals quit or not quit smoking, respectively.”
Model-based inference on average causal effect in observational clustered data
Wu, M., & Yucel, R. M. (2019) · Health Services and Outcomes Research Methodology · doi:10.1007/s10742-019-00196-2 · Cited by 2
Using multiple imputation to classify potential outcomes subgroups
Li, Y., Bondarenko, I., Elliott, M. R., Hofer, T. P., & Taylor, J. M. G. (2021) · Statistical Methods in Medical Research · doi:10.1177/09622802211002866 · Cited by 1
Silva, G. C., & Gutman, R. (2021) · Statistics in Medicine · doi:10.1002/sim.9231 · Cited by 2
Pooling methods for likelihood ratio tests in multiply imputed data setsPreprint
Grund, S., Lüdtke, O., & Robitzsch, A. (2021) · PsyArXiv (OSF Preprints) · doi:10.31234/osf.io/d459g · Cited by 2
Meisner, A., Xia, F., Chan, K. C. G., Mayer, K. H., Wheeler, D. P., Zangeneh, S. Z., & Donnell, D. J. (2024) · medRxiv · doi:10.1101/2024.01.10.24301113 · Cited by 1
“This approach requires assuming conditional exchangeability, i.e., no unmeasured confounding, and positivity, that is, a positive probability of [34, 35].”
BackgroundUsing Multiple Imputation to Classify Potential Outcomes SubgroupsPreprint
Li, Y., Bondarenko, I., Elliott, M. R., Hofer, T. P., & Taylor, J. M. G. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.04257
Causal Inference with Missingness in Confounders
Bagmar, M. S. H. (2019) · doi:10.11575/prism/36819
Causal Inference and Prediction on Observational Data with Survival Outcomes
Chen, X. (2020) · SMU Scholar (Southern Methodist University)
Estimating the optimal timing of surgery by imputing potential outcomes
Chen, X., Heitjan, D. F., Greil, G. F., & Jeon‐Slaughter, H. (2021) · Statistics in Medicine · doi:10.1002/sim.9217 · Cited by 1
A survey on temporal network dynamics with incomplete data
Wu, D. D., Mao, S., Xiong, L., & Tang, Y. S. (2022) · Electronic Research Archive · doi:10.3934/era.2022193 · Cited by 1
Carpenter, J. R., Kenward, M. G., Bartlett, J. W., Morris, T. P., Quartagno, M., & Wood, A. (2023) · Wiley eBooks · doi:10.1002/9781119756118.ref
Cole, S. R., Zivich, P. N., Edwards, J. K., Ross, R. K., & Shook‐Sa, B. E. (2023) · American Journal of Epidemiology · doi:10.1093/aje/kwad216
Self-citationSelection Bias Understood Through Sequential Causal DiagramsPreprint
Bulbulia, J. A. (2023) · PsyArXiv (OSF Preprints) · doi:10.31234/osf.io/cjgey
An Invitation to Causal Inference in Environmental PsychologyPreprint
Bulbulia, J. A., & Hine, D. W. (2024) · PsyArXiv (OSF Preprints) · doi:10.31234/osf.io/tbjx8
Estimating the effect of pre-exposure prophylaxis in Black men who have sex with men
Meisner, A., Xia, F., Chan, K. C. G., Mayer, K. H., Wheeler, D. P., Zangeneh, S. Z., & Donnell, D. J. (2024) · International Journal of Epidemiology · doi:10.1093/ije/dyae170 · Cited by 1
A comparison of propensity score-based causal estimators for analyzing partially missing confounder
Bagmar, M. S. H., & Shen, H. (2025) · Communications for Statistical Applications and Methods · doi:10.29220/csam.2025.32.3.259
Estimating average treatment effect using multiple imputation for handling confounder missingness
Bagmar, M. S. H., & Shen, H. (2025) · Communications in Statistics - Simulation and Computation · doi:10.1080/03610918.2025.2534890
Causal Inference for Latent Markov Models Using the Parametric G-Formula
Clouth, F. J., Bijlsma, M. J., Pauws, S. C., & Vermunt, J. K. (2025) · Sociological Methods & Research · doi:10.1177/00491241251377068
“In fact, multiple imputation and the g-formula are closely related (Westreich et al. 2015).”
Kawahara, T., Shinozaki, T., & Matsuyama, Y. (2020) · Figshare · doi:10.6084/m9.figshare.12750090
Nunes, A. P., Zhao, D., Jesdale, W., & Lapane, K. L. (2021) · Figshare · doi:10.6084/m9.figshare.14853498
Additional file 1 of Model-based standardization using multiple imputation
Remiro‐Azócar, A., Heath, A., & Baio, G. (2024) · Figshare · doi:10.6084/m9.figshare.25201939
Causal Inference in Plant Disease Contexts
Shah, D. A., Sanjel, S., & Esker, P. D. (2026) · Annual Review of Phytopathology · doi:10.1146/annurev-phyto-011325-094447
Metadata
- Authors
- Westreich, D., Edwards, J. K., Cole, S. R., Platt, R. W., Mumford, S. L., & Schisterman, E. F.
- Affiliations
- University of North Carolina at Chapel Hill (US)
- McGill University (CA)
- National Institutes of Health (US)
- Corresponding author
- Westreich, Daniel
- ORCID
- Venue
- International Journal of Epidemiology
- Publisher
- Oxford University Press (OUP)
- Published
- 2015-07-25
- Volume
- 44
- Issue
- 5
- Pages
- 1731-1737
- ISSN
- 0300-5771 (ISSN-L)
- 1464-3685
- PMC
- PMC4707196
- OpenAlex
- W2278756228
- Cited by
- 64
- References
- 45
- Citations by year
- 2026: 2 · 2025: 6 · 2024: 9 · 2023: 7 · 2022: 9 · 2021: 9 · 2020: 8 · 2019: 7 · 2018: 5 · 2017: 2
- FWCI
- 2.08
- Type
- journal article
- Language
- en
- Open access
- green
- www.ncbi.nlm.nih.gov
- Locations
- International Journal of Epidemiology (publishedVersion)
- PubMed Central (publishedVersion)
- PubMed (publishedVersion)
- Funders
- National Institutes of Health
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Bayesian Inference
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
- Keywords
- potential outcomes, causal inference, missing data, multiple imputation, g-formula
- MeSH terms
- Epidemiology, Humans, Research Design, Models, Statistical, Causality
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.