A flexible approach for causal inference with multiple treatments and clustered survival outcomes

Liangyuan Hu; Jiayi Ji; Ronald D. Ennis; Joseph W. Hogan

doi:https://doi.org/10.1002/sim.9548

A flexible approach for causal inference with multiple treatments and clustered survival outcomes

Liangyuan Hu, Jiayi Ji, Ronald D. Ennis, Joseph W. Hogan

School of Public Health, Biostatistics

Rutgers, The State University

Research output: Contribution to journal › Article › peer-review

Abstract

When drawing causal inferences about the effects of multiple treatments on clustered survival outcomes using observational data, we need to address implications of the multilevel data structure, multiple treatments, censoring, and unmeasured confounding for causal analyses. Few off-the-shelf causal inference tools are available to simultaneously tackle these issues. We develop a flexible random-intercept accelerated failure time model, in which we use Bayesian additive regression trees to capture arbitrarily complex relationships between censored survival times and pre-treatment covariates and use the random intercepts to capture cluster-specific main effects. We develop an efficient Markov chain Monte Carlo algorithm to draw posterior inferences about the population survival effects of multiple treatments and examine the variability in cluster-level effects. We further propose an interpretable sensitivity analysis approach to evaluate the sensitivity of drawn causal inferences about treatment effect to the potential magnitude of departure from the causal assumption of no unmeasured confounding. Expansive simulations empirically validate and demonstrate good practical operating characteristics of our proposed methods. Applying the proposed methods to a dataset on older high-risk localized prostate cancer patients drawn from the National Cancer Database, we evaluate the comparative effects of three treatment approaches on patient survival, and assess the ramifications of potential unmeasured confounding. The methods developed in this work are readily available in the (Formula presented.) package (Formula presented.).

Original language	American English
Pages (from-to)	4982-4999
Number of pages	18
Journal	Statistics in Medicine
Volume	41
Issue number	25
DOIs	https://doi.org/10.1002/sim.9548
State	Published - Nov 10 2022

ASJC Scopus subject areas

Epidemiology
Statistics and Probability

Keywords

Bayesian machine learning
multilevel survival data
observational studies
sensitivity analysis

Access to Document

https://doi.org/10.1002/sim.9548

Cite this

@article{138de623fddb411c9fdd8c076da178c5,

title = "A flexible approach for causal inference with multiple treatments and clustered survival outcomes",

abstract = "When drawing causal inferences about the effects of multiple treatments on clustered survival outcomes using observational data, we need to address implications of the multilevel data structure, multiple treatments, censoring, and unmeasured confounding for causal analyses. Few off-the-shelf causal inference tools are available to simultaneously tackle these issues. We develop a flexible random-intercept accelerated failure time model, in which we use Bayesian additive regression trees to capture arbitrarily complex relationships between censored survival times and pre-treatment covariates and use the random intercepts to capture cluster-specific main effects. We develop an efficient Markov chain Monte Carlo algorithm to draw posterior inferences about the population survival effects of multiple treatments and examine the variability in cluster-level effects. We further propose an interpretable sensitivity analysis approach to evaluate the sensitivity of drawn causal inferences about treatment effect to the potential magnitude of departure from the causal assumption of no unmeasured confounding. Expansive simulations empirically validate and demonstrate good practical operating characteristics of our proposed methods. Applying the proposed methods to a dataset on older high-risk localized prostate cancer patients drawn from the National Cancer Database, we evaluate the comparative effects of three treatment approaches on patient survival, and assess the ramifications of potential unmeasured confounding. The methods developed in this work are readily available in the (Formula presented.) package (Formula presented.).",

keywords = "Bayesian machine learning, multilevel survival data, observational studies, sensitivity analysis",

author = "Liangyuan Hu and Jiayi Ji and Ennis, {Ronald D.} and Hogan, {Joseph W.}",

note = "Funding Information: This work was supported in part by the National Cancer Institute under Grant R21CA245855, and by the National Heart, Lung, and Blood Institute under Grant 1R01HL159077‐01A1, and by award ME_2017C3_9041 from the Patient‐Centered Outcomes Research Institute (PCORI). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or PCORI. Publisher Copyright: {\textcopyright} 2022 The Authors. Statistics in Medicine published by John Wiley & Sons Ltd.",

year = "2022",

month = nov,

day = "10",

doi = "https://doi.org/10.1002/sim.9548",

language = "American English",

volume = "41",

pages = "4982--4999",

journal = "Statistics in Medicine",

issn = "0277-6715",

publisher = "John Wiley and Sons Ltd",

number = "25",

}

TY - JOUR

T1 - A flexible approach for causal inference with multiple treatments and clustered survival outcomes

AU - Hu, Liangyuan

AU - Ji, Jiayi

AU - Ennis, Ronald D.

AU - Hogan, Joseph W.

N1 - Funding Information: This work was supported in part by the National Cancer Institute under Grant R21CA245855, and by the National Heart, Lung, and Blood Institute under Grant 1R01HL159077‐01A1, and by award ME_2017C3_9041 from the Patient‐Centered Outcomes Research Institute (PCORI). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or PCORI. Publisher Copyright: © 2022 The Authors. Statistics in Medicine published by John Wiley & Sons Ltd.

PY - 2022/11/10

Y1 - 2022/11/10

N2 - When drawing causal inferences about the effects of multiple treatments on clustered survival outcomes using observational data, we need to address implications of the multilevel data structure, multiple treatments, censoring, and unmeasured confounding for causal analyses. Few off-the-shelf causal inference tools are available to simultaneously tackle these issues. We develop a flexible random-intercept accelerated failure time model, in which we use Bayesian additive regression trees to capture arbitrarily complex relationships between censored survival times and pre-treatment covariates and use the random intercepts to capture cluster-specific main effects. We develop an efficient Markov chain Monte Carlo algorithm to draw posterior inferences about the population survival effects of multiple treatments and examine the variability in cluster-level effects. We further propose an interpretable sensitivity analysis approach to evaluate the sensitivity of drawn causal inferences about treatment effect to the potential magnitude of departure from the causal assumption of no unmeasured confounding. Expansive simulations empirically validate and demonstrate good practical operating characteristics of our proposed methods. Applying the proposed methods to a dataset on older high-risk localized prostate cancer patients drawn from the National Cancer Database, we evaluate the comparative effects of three treatment approaches on patient survival, and assess the ramifications of potential unmeasured confounding. The methods developed in this work are readily available in the (Formula presented.) package (Formula presented.).

AB - When drawing causal inferences about the effects of multiple treatments on clustered survival outcomes using observational data, we need to address implications of the multilevel data structure, multiple treatments, censoring, and unmeasured confounding for causal analyses. Few off-the-shelf causal inference tools are available to simultaneously tackle these issues. We develop a flexible random-intercept accelerated failure time model, in which we use Bayesian additive regression trees to capture arbitrarily complex relationships between censored survival times and pre-treatment covariates and use the random intercepts to capture cluster-specific main effects. We develop an efficient Markov chain Monte Carlo algorithm to draw posterior inferences about the population survival effects of multiple treatments and examine the variability in cluster-level effects. We further propose an interpretable sensitivity analysis approach to evaluate the sensitivity of drawn causal inferences about treatment effect to the potential magnitude of departure from the causal assumption of no unmeasured confounding. Expansive simulations empirically validate and demonstrate good practical operating characteristics of our proposed methods. Applying the proposed methods to a dataset on older high-risk localized prostate cancer patients drawn from the National Cancer Database, we evaluate the comparative effects of three treatment approaches on patient survival, and assess the ramifications of potential unmeasured confounding. The methods developed in this work are readily available in the (Formula presented.) package (Formula presented.).

KW - Bayesian machine learning

KW - multilevel survival data

KW - observational studies

KW - sensitivity analysis

UR - http://www.scopus.com/inward/record.url?scp=85135615925&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85135615925&partnerID=8YFLogxK

U2 - https://doi.org/10.1002/sim.9548

DO - https://doi.org/10.1002/sim.9548

M3 - Article

C2 - 35948011

SN - 0277-6715

VL - 41

SP - 4982

EP - 4999

JO - Statistics in Medicine

JF - Statistics in Medicine

IS - 25

ER -

A flexible approach for causal inference with multiple treatments and clustered survival outcomes

Abstract

ASJC Scopus subject areas

Keywords

Access to Document

Other files and links

Fingerprint

Cite this