Ranking of average treatment effects with generalized random forests for time-to-event outcomes

Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt

Standard

Ranking of average treatment effects with generalized random forests for time-to-event outcomes. / Rytgaard, Helene C.W.; Ekstrøm, Claus T.; Kessing, Lars V.; Gerds, Thomas A.

I: Statistics in Medicine, Bind 42, Nr. 10, 2023, s. 1542-1564.

Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt

Harvard

Rytgaard, HCW, Ekstrøm, CT, Kessing, LV & Gerds, TA 2023, 'Ranking of average treatment effects with generalized random forests for time-to-event outcomes', Statistics in Medicine, bind 42, nr. 10, s. 1542-1564. https://doi.org/10.1002/sim.9686

APA

Rytgaard, H. C. W., Ekstrøm, C. T., Kessing, L. V., & Gerds, T. A. (2023). Ranking of average treatment effects with generalized random forests for time-to-event outcomes. Statistics in Medicine, 42(10), 1542-1564. https://doi.org/10.1002/sim.9686

Vancouver

Rytgaard HCW, Ekstrøm CT, Kessing LV, Gerds TA. Ranking of average treatment effects with generalized random forests for time-to-event outcomes. Statistics in Medicine. 2023;42(10):1542-1564. https://doi.org/10.1002/sim.9686

Author

Rytgaard, Helene C.W. ; Ekstrøm, Claus T. ; Kessing, Lars V. ; Gerds, Thomas A. / Ranking of average treatment effects with generalized random forests for time-to-event outcomes. I: Statistics in Medicine. 2023 ; Bind 42, Nr. 10. s. 1542-1564.

Bibtex

@article{ea5f57a9fb4240f1a6deee676db0f6c9,
title = "Ranking of average treatment effects with generalized random forests for time-to-event outcomes",
abstract = "Linkage between drug claims data and clinical outcome allows a data-driven experimental approach to drug repurposing. We develop an estimation procedure based on generalized random forests for estimation of time-point specific average treatment effects in a time-to-event setting with competing risks. To handle right-censoring, we propose a two-step procedure for estimation, applying inverse probability weighting to construct time-point specific weighted outcomes as input for the generalized random forest. The generalized random forests adaptively handle covariate effects on the treatment assignment by applying a splitting rule that targets a causal parameter. Using simulated data we demonstrate that the method is effective for a causal search through a list of treatments to be ranked according to the magnitude of their effect on clinical outcome. We illustrate the method using the Danish national health registries where it is of interest to discover drugs with an unexpected protective effect against relapse of severe depression.",
keywords = "average treatment effect, competing risks, random forests, time-to-event",
author = "Rytgaard, {Helene C.W.} and Ekstr{\o}m, {Claus T.} and Kessing, {Lars V.} and Gerds, {Thomas A.}",
note = "Publisher Copyright: {\textcopyright} 2023 John Wiley & Sons Ltd.",
year = "2023",
doi = "10.1002/sim.9686",
language = "English",
volume = "42",
pages = "1542--1564",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "JohnWiley & Sons Ltd",
number = "10",

}

RIS

TY - JOUR

T1 - Ranking of average treatment effects with generalized random forests for time-to-event outcomes

AU - Rytgaard, Helene C.W.

AU - Ekstrøm, Claus T.

AU - Kessing, Lars V.

AU - Gerds, Thomas A.

N1 - Publisher Copyright: © 2023 John Wiley & Sons Ltd.

PY - 2023

Y1 - 2023

N2 - Linkage between drug claims data and clinical outcome allows a data-driven experimental approach to drug repurposing. We develop an estimation procedure based on generalized random forests for estimation of time-point specific average treatment effects in a time-to-event setting with competing risks. To handle right-censoring, we propose a two-step procedure for estimation, applying inverse probability weighting to construct time-point specific weighted outcomes as input for the generalized random forest. The generalized random forests adaptively handle covariate effects on the treatment assignment by applying a splitting rule that targets a causal parameter. Using simulated data we demonstrate that the method is effective for a causal search through a list of treatments to be ranked according to the magnitude of their effect on clinical outcome. We illustrate the method using the Danish national health registries where it is of interest to discover drugs with an unexpected protective effect against relapse of severe depression.

AB - Linkage between drug claims data and clinical outcome allows a data-driven experimental approach to drug repurposing. We develop an estimation procedure based on generalized random forests for estimation of time-point specific average treatment effects in a time-to-event setting with competing risks. To handle right-censoring, we propose a two-step procedure for estimation, applying inverse probability weighting to construct time-point specific weighted outcomes as input for the generalized random forest. The generalized random forests adaptively handle covariate effects on the treatment assignment by applying a splitting rule that targets a causal parameter. Using simulated data we demonstrate that the method is effective for a causal search through a list of treatments to be ranked according to the magnitude of their effect on clinical outcome. We illustrate the method using the Danish national health registries where it is of interest to discover drugs with an unexpected protective effect against relapse of severe depression.

KW - average treatment effect

KW - competing risks

KW - random forests

KW - time-to-event

U2 - 10.1002/sim.9686

DO - 10.1002/sim.9686

M3 - Journal article

C2 - 36815690

AN - SCOPUS:85148610032

VL - 42

SP - 1542

EP - 1564

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 10

ER -

ID: 339547252