So Many Choices: A Guide to Selecting Among Methods to Adjust for Observed Confounders

Luke Keele; Richard Grieve

doi:10.1002/sim.10336

journal article Open Access Feb 13, 2025

So Many Choices: A Guide to Selecting Among Methods to Adjust for Observed Confounders

Luke Keele

Richard Grieve

Statistics in Medicine Vol. 44 No. 5 · Wiley

View at Publisher Save 10.1002/sim.10336

Abstract

ABSTRACT
Non‐randomised studies (NRS) typically assume that there are no differences in unobserved baseline characteristics between the treatment groups under comparison. Traditionally regression models have been deployed to estimate treatment effects adjusting for observed confounders but can lead to biased estimates if the model is missspecified, by making incorrect functional form assumptions. A multitude of alternative methods have been developed which can reduce the risk of bias due to model misspecification. Investigators can now choose between many forms of matching, weighting, doubly robust, and machine learning methods. We review key concepts related to functional form assumptions and how those can contribute to bias from model misspecification. We then categorize the three frameworks for modeling treatment effects and the wide variety of estimation methods that can be applied to each framework. We consider why machine learning methods have been widely proposed for estimation and review the strengths and weaknesses of these approaches. We apply a range of these methods in re‐analyzing a landmark case study. In the application, we examine how several widely used methods may be subject to bias from model misspecification. We conclude with a set of recommendations for practice.

Topics

No keywords indexed for this article. Browse by subject →

References

89

[1]

10.2105/ajph.2018.304337

[2]

Using Big Data to Emulate a Target Trial When a Randomized Trial Is Not Available: Table 1.

Miguel A. Hernán, James M. Robins

American Journal of Epidemiology 10.1093/aje/kwv254

[3]

10.2307/2344179

[4]

10.1037/h0037350

[5]

10.1162/rest_a_00318

[6]

10.1080/01621459.1989.10478868

[7]

Optimal Full Matching and Related Designs via Network Flows

Ben B Hansen, Stephanie Olsen Klopfer

Journal of Computational and Graphical Statistics 10.1198/106186006x137047

[8]

10.1080/01621459.2012.703874

[9]

10.1214/14-aoas770

[10]

10.1080/01621459.2014.997879

[11]

10.1198/jasa.2011.tm09599

[12]

10.1093/pan/mpr013

[13]

Matching As An Econometric Evaluation Estimator

James J. Heckman, Hidehiko Ichimura, Petra Todd

The Review of Economic Studies 10.1111/1467-937x.00044

[14]

Matching As An Econometric Evaluation Estimator: Evidence from Evaluating a Job Training Programme

J. J. Heckman, H. Ichimura, P. E. Todd

The Review of Economic Studies 10.2307/2971733

[15]

Methods for dealing with time‐dependent confounding

R.M. Daniel, S. N. Cousens, B. L. De Stavola et al.

Statistics in Medicine 10.1002/sim.5686

[16]

10.1002/sim.9234

[17]

The Effectiveness of Right Heart Catheterization in the Initial Care of Critically III Patients

Alfred F. Connors

JAMA 10.1001/jama.1996.03540110043030

[18]

10.1214/ss/1177012031

[19]

10.1177/0962280213507034

[20]

10.1097/ede.0000000000001644

[21]

Rubin D. B. "Which Ifs Have Causal Answers" Journal of the American Statistical Association (1986)

[22]

10.1097/ede.0b013e3181bd5638

[23]

10.1214/ss/1177010894

[24]

Causality P. J. (2009)

[25]

10.1002/sim.6128

[26]

10.1002/sim.9861

[27]

Dealing with limited overlap in estimation of average treatment effects

R. K. Crump, V. J. Hotz, G. W. Imbens et al.

Biometrika 10.1093/biomet/asn055

[28]

10.1198/jcgs.2011.09219

[29]

Balancing Covariates via Propensity Score Weighting

Fan Li, Kari Lock Morgan, Alan M. Zaslavsky

Journal of the American Statistical Association 10.1080/01621459.2016.1260466

[30]

10.1007/978-1-4419-1213-8

[31]

10.1016/j.csda.2018.05.003

[32]

L.Goff “The Bias From Misspecification of Control Variables as Linear ”Resources for the Future Discussion Paper(2014): 14–41. 10.2139/ssrn.2537916

[33]

10.1080/01621459.1999.10473862

[34]

A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect

James Robins

Mathematical Modelling 10.1016/0270-0255(86)90088-6

[35]

Hernán M. A. (2020)

[36]

10.2307/2529685

[37]

10.2307/2529684

[38]

10.1198/106186001317114938

[39]

Substantial Gains in Bias Reduction from Matching with a Variable Number of Controls

Kewei Ming, Paul R. Rosenbaum

Biometrics 10.1111/j.0006-341x.2000.00118.x

[40]

Full Matching in an Observational Study of Coaching for the SAT

Ben B Hansen

Journal of the American Statistical Association 10.1198/016214504000000647

[41]

Efficient Estimation of Average Treatment Effects Using the Estimated Propensity Score

Keisuke Hirano, Guido W. Imbens, Geert Ridder

Econometrica 10.1111/1468-0262.00442

[42]

Marginal Structural Models and Causal Inference in Epidemiology

James M. Robins, Miguel Ángel Hernán, Babette Brumback

Epidemiology 10.1097/00001648-200009000-00011

[43]

10.1023/a:1020371312283

[44]

Nonparametric Estimation of Average Treatment Effects Under Exogeneity: A Review

Guido W. Imbens

Review of Economics and Statistics 10.1162/003465304323023651

[45]

10.1080/01621459.1995.10476494

[46]

Entropy Balancing for Causal Effects: A Multivariate Reweighting Method to Produce Balanced Samples in Observational Studies

Jens Hainmueller

Political Analysis 10.1093/pan/mpr025

[47]

Stable Weights that Balance Covariates for Estimation With Incomplete Outcome Data

José R. Zubizarreta

Journal of the American Statistical Association 10.1080/01621459.2015.1023805

[48]

10.1515/jci‐2016‐0010

[49]

10.1214/18‐aos1698

[50]

10.1093/biomet/asz050

Showing 50 of 89 references

Metrics

3

Citations

89

References

Details

Published: Feb 13, 2025
Vol/Issue: 44(5)
License: View

Authors

L

Luke Keele

Dept. of Surgery University of Pennsylvania Pennsylvania USA

R

Richard Grieve

Department of Health Services Research and Policy London School of Hygiene and Tropical Medicine London UK

Cite This Article

Luke Keele, Richard Grieve (2025). So Many Choices: A Guide to Selecting Among Methods to Adjust for Observed Confounders. Statistics in Medicine, 44(5). https://doi.org/10.1002/sim.10336

So Many Choices: A Guide to Selecting Among Methods to Adjust for Observed Confounders

You May Also Like