EUVIMEDEUVIMED entdecken
Randomisierte kontrollierte StudieOpen Access · CC BY-NC-ND 4.0

Generalizing treatment effects with incomplete covariates: Identifying assumptions and multiple imputation algorithms

Imke Mayer iD, Julie Josse, Traumabase Group

Biometrical Journal · 2023-06 · 65 (5)

DOI: 10.1002/bimj.202100294 ↗

Kurz gefasst

Randomisierte kontrollierte Studie in der Zeitschrift Biometrical Journal, online erschienen am 12.03.2023. Verfasst von Imke Mayer und 2 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Public Health, Charité - Universitätsmedizin, Berlin, Germany. Hauptthemen laut MEDLINE: Models, Statistical, Algorithms. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 60 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Google, École des Hautes Etudes en Sciences Sociales.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

We focus on the problem of generalizing a causal effect estimated on a randomized controlled trial (RCT) to a target population described by a set of covariates from observational data. Available methods such as inverse propensity sampling weighting are not designed to handle missing values, which are however common in both data sources. In addition to coupling the assumptions for causal effect identifiability and for the missing values mechanism and to defining appropriate estimation strategies, one difficulty is to consider the specific structure of the data with two sources and treatment and outcome only available in the RCT. We propose three multiple imputation strategies to handle missing values when generalizing treatment effects, each handling the multisource structure of the problem differently (separate imputation, joint imputation with fixed effect, joint imputation ignoring source information). As an alternative to multiple imputation, we also propose a direct estimation approach that treats incomplete covariates as semidiscrete variables. The multiple imputation strategies and the latter alternative rely on different sets of assumptions concerning the impact of missing values on identifiability. We discuss these assumptions and assess the methods through an extensive simulation study. This work is motivated by the analysis of a large registry of over 20,000 major trauma patients and an RCT studying the effect of tranexamic acid administration on mortality in major trauma patients admitted to intensive care units. The analysis illustrates how the missing values handling can impact the conclusion about the effect generalized from the RCT to the target population.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Models, StatisticalAlgorithmsHumansComputer SimulationRegistriesData Interpretation, Statistical

Stichwörter der Autor:innen

causal effect transportabilitydata integrationexternal validitymissing incorporated in attributesmissing valuesmultiple imputationrandom forest

Autor:innen und Institutionen

  • Institute of Public Health, Charité - Universitätsmedizin, Berlin, Germany.
  • PreMeDICaL, Inria Sophia-Antipolis, Montpellier, France.
  • Department of Anesthesia and Intensive Care, Beaujon Hospital, AP-HP, Clichy, France.

ORCID-Profile: Imke Mayer

Förderung

  • Google
  • École des Hautes Etudes en Sciences Sociales

Zitieren

APA

Mayer, I., Josse, J., & Traumabase Group (2023). Generalizing treatment effects with incomplete covariates: Identifying assumptions and multiple imputation algorithms. Biometrical Journal, 65(5). https://doi.org/10.1002/bimj.202100294

Vancouver

Mayer I, Josse J, Traumabase Group. Generalizing treatment effects with incomplete covariates: Identifying assumptions and multiple imputation algorithms. Biometrical Journal. 2023;65(5). doi:10.1002/bimj.202100294

BibTeX
@article{Mayer2023,
  author = {Mayer, Imke and Josse, Julie and Traumabase Group, Traumabase Group},
  title = {Generalizing treatment effects with incomplete covariates: Identifying assumptions and multiple imputation algorithms},
  journal = {Biometrical Journal},
  year = {2023},
  volume = {65},
  number = {5},
  doi = {10.1002/bimj.202100294}
}
Kostenlos starten →