Pith. sign in

REVIEW 3 major objections 5 minor 2 cited by

Orthogonal Survival Learners for Estimating Heterogeneous Treatment Effects from Time-to-Event Data

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that one weighted loss family yields orthogonal, model-agnostic survival treatment effect learners whose minimizer equals the true heterogeneous treatment effect.

desk verdict A genuinely useful generalization of orthogonal learners to survival data, with a real practical gap: the advertised overlap robustness depends on an f-selection rule the paper doesn't provide. read the letter →

arxiv 2505.13072 v1 pith:F2JZRDXN submitted 2025-05-19 cs.LG stat.ML

Dennis Frauen , Maresa Schröder , Konstantin Hess , Stefan Feuerriegel This is my paper · ORCID
classification cs.LGstat.ML
keywords heterogeneoustreatmenteffectssurvivalanalysistime-to-eventdatacensoringNeymanorthogonalityoverlapviolationsmeta-learnerscausalinference
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that heterogeneous treatment effects (HTEs) from censored time-to-event data can be estimated by minimizing one orthogonal weighted loss, no matter which positive weighting function is used. The loss is built so that small errors in the nuisance estimates (propensity score, survival hazards, censoring hazards) do not shift the second-stage estimate, and its population minimizer equals the true conditional survival treatment effect whenever the model class contains it. The weighting function is the design choice: it can down-weight regions with low treatment, censoring, or survival overlap, yielding both existing survival DR- and R-learners as special cases and new learners targeted at overlap problems specific to survival data. If the claim is right, practitioners get a model-agnostic way to build survival HTE learners whose robustness to overlap violations is controlled by a single weighting function.

What carries the argument

The central object is the orthogonal loss $L_f(g,\eta_t)$ of Eq. (7), assembled from a positive weighting function $f(\tilde\eta_t(X))$ and two pseudo-outcome terms $\rho(Z,\eta_t)$ and $\varphi(Z,\eta_t)$ in Eqs. (8)-(9). The terms $\xi^S$ and $\xi^G$ in Eq. (10) are residual processes whose conditional means given $(X,A)$ are zero (Lemma H.2); those cancellations make all cross-derivatives $D_{\eta}D_g L_f$ vanish, which is exactly Neyman orthogonality. For the targeting claim, the $g$-dependent part of $L_f$ reduces to the weighted squared loss $\mathbb{E}[f(\tilde\eta_t(X))(\tau_t(X)-g(X))^2]$ up to terms independent of $g$, so the minimizer is $\tau_t$ whenever the model class contains it. The user controls $f$, choosing products of propensity, survival, or censoring functions to retarget the loss toward well-supported regions.

What would settle it

Take a synthetic dataset with known $\tau_t$, deliberately misspecify one nuisance function (for example, replace the true propensity score by a constant $0.5$) while keeping the others accurate, and evaluate Eq. (12) with several positive weighting functions $f$. If the PEHE changes at first order with the misspecification for some $f$, or if the empirical minimizer departs systematically from $\tau_t$ even though $\tau_t\in G$, the orthogonality claim fails; equivalently, compute a finite-sample estimate of the cross-derivative $D_{\eta}D_g L_f$ and check whether it is zero.

Watch

Extended reading notes

Core claim

The central claim is that the population loss $L_f(g,\eta_t)$ defined in Eq. (7), with $\rho$ and $\varphi$ given by Eqs. (8)-(10), is Neyman-orthogonal (that is, first-order insensitive to nuisance estimation errors) with respect to the nuisance functions $\eta_t=(\pi,\lambda^S_i(\cdot,1),\lambda^S_i(\cdot,0),\lambda^G_i(\cdot,1),\lambda^G_i(\cdot,0))_{i=0}^t$ for any positive weighting function $f$, and that its minimizer over a class $G$ coincides with the true conditional survival treatment effect $\tau_t(x)=P(T(1)>t\mid X=x)-P(T(0)>t\mid X=x)$ whenever $\tau_t\in G$ (Theorems 5.1 and 5.2). This means weighting changes only which population the loss emphasizes, not the target estimand. The paper instantiates the toolbox with $f=1$ (recovering the survival DR-learner), $f=\pi(1-\pi)$ (recovering the survival R-learner), $f=G_{t-1}(\cdot,1)G_{t-1}(\cdot,0)$ for censoring overlap, $f=S_{t-1}(\cdot,1)S_{t-1}(\cdot,0)$ for survival overlap, and combinations of these. In synthetic experiments, the learner whose weighting matches the induced overlap violation attains the lowest PEHE, and a twins birth-weight case study shows faster convergence for censoring-weighted learners.

Load-bearing premise

The load-bearing premise is that the analyst can tell which overlap type (treatment, censoring, or survival) is actually violated and choose the matching weighting function $f$, because the paper's simulations pick $f$ with oracle knowledge of the violation and report that unsuitable weighting can harm performance, e.g., in the no-violation Scenario 2, the T+S learner gives PEHE 3.55 versus 1.64 for the unweighted learner.

Editorial extensions

If this is right

  • Any learner instantiated from the toolbox inherits orthogonality, so first-order errors in the propensity, survival-hazard, or censoring-hazard estimates do not bias the HTE estimate.
  • The loss minimizer is the true conditional survival treatment effect $\tau_t(x)$ for every positive $f$ whenever the second-stage model class is rich enough, so weighting affects robustness and variance but not the estimand.
  • The survival DR- and R-learners are recovered as the special cases $f=1$ and $f=\pi(1-\pi)$; the new C- and S-learners target censoring and survival overlap, with combined variants for multiple violations.
  • Empirically, the learner whose weighting matches the overlap violation present in the data achieves the lowest PEHE and faster convergence; mismatched weighting can increase variance.
  • Because the loss is model-agnostic, the same second-stage loss can be paired with any nuisance estimator or regression model for $g$, including neural networks, trees, or linear models.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A data-driven rule for choosing $f$ from estimated overlap diagnostics would remove the main practical gap, since the paper's experiments select the weighting that matches the simulated violation and mismatched weighting is shown to hurt.
  • The same orthogonal-loss construction should transfer to related survival estimands such as restricted mean survival times and treatment-specific survival curves using the appendix derivations; the main test would be whether the weighting choices remain equally beneficial there.
  • Because $f$ indexes a continuum, practitioners could interpolate between DR-type and R-type behavior or tune the retargeting strength by cross-validation on a PEHE-style criterion rather than committing to one overlap type.
  • For clinical use, pairing the orthogonal loss with an interpretable second-stage class (trees, linear models) would give directly readable HTE estimates that still carry the orthogonality guarantee.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes a general toolbox for constructing Neyman-orthogonal meta-learners that estimate conditional survival treatment effects τ_t(x) from censored time-to-event data. The core theoretical contribution is a family of losses L_f(g, η_t) (Eq. 7) parametrized by an arbitrary positive weighting function f; the paper proves that this loss is Neyman-orthogonal with respect to the nuisance functions (Theorem 5.1) and that its population minimizer coincides with the true τ_t whenever τ_t lies in the function class G (Theorem 5.2). The toolbox is instantiated with several weighting functions targeting treatment, censoring, and survival overlap, recovering the Survival DR-learner and Survival R-learner as special cases and yielding new learners for survival-specific overlap violations. The authors evaluate the learners on synthetic data with simulated overlap violations and on the Twins dataset, reporting PEHE improvements when the weighting function is matched to the violation type.

Significance. If the theoretical claims hold, this is a valuable unifying framework: it subsumes existing survival meta-learners, extends orthogonal estimation to censoring- and survival-overlap regimes, is model-agnostic, and comes with concrete extensions to marginalized effects, continuous time, and other estimands in the appendices. The proof strategy is a direct derivative computation and correctly reproduces the known special cases f=1 (Survival DR) and f=π(1−π) (Survival R), which is reassuring. The authors also provide code and the derivations are sufficiently detailed to be checked. The main weakness is that the advertised practical robustness depends on choosing f to match the unknown overlap violation, and the paper does not yet supply a data-driven selection rule; the experiments select f with oracle knowledge of the simulated violation. This gap is acknowledged in the limitations but remains load-bearing for the practical claim.

major comments (3)
  1. [Section 7, Table 3, Appendix J] The empirical evaluation of the toolbox's central practical advantage—robustness to unknown overlap violations—selects the targeted learner per simulated violation type (gray cells in Tables 2 and 3), i.e., with oracle knowledge of the true overlap problem. The paper's own Table 3 shows that a mismatched weighting can be substantially worse than no weighting (no-violation column: T+S PEHE 3.55±1.13 vs ∅ 1.64±0.19; survival-violation column: T 6.77±1.08 vs ∅ 4.54±0.36). Section 5's suggestion to 'inspect overlap' from estimated nuisances is not operationalized with a concrete diagnostic or a selection rule, and Appendix J concedes that inappropriate weighting 'can significantly slow down convergence.' Since real analysts do not know which overlap type is violated, the advertised robustness gain is not yet established; the manuscript should either supply a data-driven f-selection procedure with evidence that it recovers the oracle choice, or explicitly reframe the claim as robustness conditional on an appropriate choice of f.
  2. [Section 5, Theorem 5.1, Appendix H] The proof of Neyman orthogonality is the load-bearing theoretical result, but it is a long hand-derived computation that does not state the regularity conditions on f under which the required derivatives and the interchanges of differentiation and expectation are valid; notation such as ∂f/∂S_{t−1}(·,A) is also compressed. The computation appears to be correct—it reduces to the known Survival-DR (f=1) and Survival-R (f=π(1−π)) objective—but for a central theorem, the authors should either provide a more structural proof, state the smoothness/boundedness conditions on f and the hazards, or accompany the algebra with machine-checked verification.
  3. [Section 7, Tables 2 and 3] The empirical support for the practical claims is limited by the small number of runs (10) and the large, often overlapping standard deviations; in several settings the targeted learner is not the best (e.g., Table 2, Survival column: S PEHE 2.72±1.60 vs ∅ 2.70±1.63). The conclusion that 'targeted weighting generally improves performance' should be backed by more seeds and a formal comparison, or the claims should be softened to reflect the variability in the reported results.
minor comments (5)
  1. [Section 1] In the introduction, 'enable unbiased and robust estimation under both confounding end censoring' should read 'under both confounding and censoring.'
  2. [Appendix C] The sentence 'we can use parametric models λS_j(x_i,a_i,θ) parametrized by θ ... to minimize the resulting the resulting log-likelihood loss' contains a duplicated phrase 'the resulting the resulting.'
  3. [Section 6, S-learner] The phrase 'This results in weighting results in' should be corrected to 'This results in the following weighting' or similar.
  4. [Section 7, Tables 2 and 3] The text says 'we report the PEHE×10−4' but the table captions only say 'PEHE'; please clarify the scaling in both the text and the table captions.
  5. [Section 7, Figure 3] The caption 'Ratios of PEHE of the targeted learner wrt. the learner without the correct target' is confusing; please define the numerator and denominator explicitly and explain the color coding for the reader.

Circularity Check

0 steps flagged · score 1.0 of 10

No substantive circularity: the orthogonality and minimizer theorems are proved directly from the loss definitions, and the cited prior work is contextual rather than load-bearing.

full rationale

The central claims are Theorem 5.1 (orthogonality of L_f) and Theorem 5.2 (its minimizer coincides with the weighted target loss). Theorem 5.1 is proved in Appendix H by explicit directional-derivative calculations: the cross term D_eta D_g L_f is shown to vanish using Lemma H.2, which itself follows from the martingale structure of xi_S and xi_G, i.e., E[xi|X,A]=0. This is an internal proof rather than an appeal to a self-citation or to the target result. Theorem 5.2 is also proved by expanding L_f around tau_t(X); the cross term vanishes by E[xi|X,A]=0, and the identity E[rho|X]=f(eeta_t(X)) is derived from the definition of rho. Thus the minimizer result does not assume the conclusion. The weighting functions f are user-specified positive functions, not fitted parameters, so the learners are not 'predicting' a fitted quantity. The paper's self-citations [12,13] appear only in the related-work discussion and as an experimental best-practice reference, and they are not used to justify orthogonality or the minimizer property. The main practical weakness is that the correct weighting f must be chosen for a given overlap violation; the paper itself flags this in Section 5 ('we recommend using the estimated nuisance functions to inspect overlap') and in Appendix J, and the experiments select the targeted learner per simulated violation type. That is an evaluation-design limitation rather than a circular derivation. No step in the derivation reduces by definition to its own input, so no circularity is found.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The method adds no new particles, mediators, or latent variables; all quantities (propensity, hazards, survival functions) are already part of the standard survival analysis and causal inference setup. The only free choice is the weighting function f, which is not fitted to data.

free parameters (1)
  • Weighting function f(π, S_{t-1}, G_{t-1}) = f=1 (∅), f=π(1−π) (T), f=G_{t-1}(X,1)G_{t-1}(X,0) (C), f=S_{t-1}(X,1)S_{t-1}(X,0) (S), and products (T+C, T+S, C+S…
    User-selected target function defining the retargeted loss. The theory holds for any positive f, but finite-sample performance and the claim of robustness depend on choosing f that matches the actual overlap violation.
assumptions (4)
  • domain assumption Assumption 3.1: consistency, treatment overlap 0<π(x)<1, ignorability A ⊥ (T(1),T(0)) | X
    Identifiability of τ_t(x) as a causal contrast; standard in causal inference, invoked in Lemma H.1.
  • domain assumption Assumption 3.2: censoring overlap G_{t-1}(x,a)>0, survival overlap S_{t-1}(x,a)>0, non-informative censoring T ⊥ C | X,A
    Ensures the hazard functions are identified and the martingale residuals in Eq. (10) have conditional mean zero (Lemma H.2), which is the load-bearing property in the orthogonality proof.
  • domain assumption τ_t ∈ G, i.e., the second-stage function class contains the true CATE
    Theorem 5.2 proves g*_f = τ_t only under this capacity condition; otherwise the minimizer is a weighted projection.
  • domain assumption f is positive and differentiable in its arguments over the support
    The loss definition in Eq. (7) requires positivity, and the orthogonality proof in Appendix H requires first and second derivatives of f with respect to π, S_{t-1}, G_{t-1}.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Orthogonal Survival Learners for Estimating Heterogeneous Treatment Effects from Time-to-Event Data." pith.science (2026). https://pith.science/paper/F2JZRDXN

@misc{pith2026250513072,
  author       = {Pith},
  title        = {Pith review of: Orthogonal Survival Learners for Estimating Heterogeneous Treatment Effects from Time-to-Event Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F2JZRDXN}},
  note         = {Machine review of arXiv:2505.13072}
}
read the original abstract

Estimating heterogeneous treatment effects (HTEs) is crucial for personalized decision-making. However, this task is challenging in survival analysis, which includes time-to-event data with censored outcomes (e.g., due to study dropout). In this paper, we propose a toolbox of novel orthogonal survival learners to estimate HTEs from time-to-event data under censoring. Our learners have three main advantages: (i) we show that learners from our toolbox are guaranteed to be orthogonal and thus come with favorable theoretical properties; (ii) our toolbox allows for incorporating a custom weighting function, which can lead to robustness against different types of low overlap, and (iii) our learners are model-agnostic (i.e., they can be combined with arbitrary machine learning models). We instantiate the learners from our toolbox using several weighting functions and, as a result, propose various neural orthogonal survival learners. Some of these coincide with existing survival learners (including survival versions of the DR- and R-learner), while others are novel and further robust w.r.t. low overlap regimes specific to the survival setting (i.e., survival overlap and censoring overlap). We then empirically verify the effectiveness of our learners for HTE estimation in different low-overlap regimes through numerical experiments. In sum, we provide practitioners with a large toolbox of learners that can be used for randomized and observational studies with censored time-to-event data.

Figures

Figures reproduced from arXiv: 2505.13072 by the authors.

Figure 1
Figure 1. Causal graph for censored time-to-event data. Yellow variables are observed while blue variables are unob￾served. Intuitively, our goal is to recover the red arrow from A to T based on the observed variables. Data: We consider the standard setting for estimating HTEs from time-to-event data [8, 7]. That is, we consider a population (X, A, T, C) ∼ P, where X ∈ X ⊆ R p are observed covariates, A ∈ {0, 1} is a binary t… view at source ↗
Figure 2
Figure 2. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Benefit of targeted weighting over time. Ratios of PEHE of the targeted learner wrt. the learner without the correct target (data scenario I across 10 runs). Blue: Low censoring overlap scenario. Green: Low survival overlap scenario. Red: Low treatment overlap scenario. Results: We evaluate the performance of our various orthogonal learners based on the precision in the estima￾tion of heterogeneous effects (PEHE) wi… view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: Validation loss for 10-day survival: Fastest con￾vergence for the C- and T+C￾learners, indicating the presence of censoring and treatment over￾lap violations. 0 1000 2000 3000 4000 5000 6000 Epochs 0 1 2 3 4 Loss Learners 30 T 30 C 30 S 30 T+C 30 T+S 30 C+S 30 [PITH_F…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Debiased Machine Learning for Partially Linear Accelerated Failure Time Models

    stat.ME 2026-08 conditional novelty 7.0 of 10

    A debiased machine learning estimator for partially linear accelerated failure time models achieves valid inference on a target exposure under right censoring via an orthogonalized rank-based U-statistic and block-pai...

  2. Debiased maximum-likelihood estimators for hazard ratios under kernel-based machine-learning adjustment

    stat.ML 2025-07 conditional novelty 6.0 of 10

    A debiased maximum-likelihood estimator for hazard ratios in a baseline-hazard-free exponential model with kernel ML adjustment recovers true causal hazard ratios in simulations, but a key convergence assumption is le...

Reference graph

Works this paper leans on

53 extracted references · 32 canonical work pages · cited by 2 Pith papers

  1. [1]

    Conformal meta-learners for predictive inference of individual treatment effects

    Ahmed Alaa, Zaid Ahmad, and Mark van der Laan. “Conformal meta-learners for predictive inference of individual treatment effects”. In: NeurIPS. 2023

    work page 2023

  2. [2]

    Policy learning with observational data

    Susan Athey and Stefan Wager. “Policy learning with observational data”. In:Econometrica 89.1 (2021), pp. 133–161

    2021

  3. [3]

    Doubly robust estimation in missing data and causal inference models

    Heejung Bang and James M. Robins. “Doubly robust estimation in missing data and causal inference models”. In: Biometrics 61.4 (2005), pp. 962–973

    work page 2005

  4. [4]

    One-step targeted maximum likelihood for time-to- event outcomes

    Weixin Cai and Mark J. van der Laan. “One-step targeted maximum likelihood for time-to- event outcomes”. In: Biometrics 76.3 (2020)

    work page 2020

  5. [5]

    Double/debiased machine learning for treatment and structural parameters

    Victor Chernozhukov et al. “Double/debiased machine learning for treatment and structural parameters”. In: The Econometrics Journal 21.1 (2018), pp. C1–C68

    2018

  6. [6]

    Counterfactual predic- tions under runtime confounding

    Amanda Coston, Edward H. Kennedy, and Alexandra Chouldechova. “Counterfactual predic- tions under runtime confounding”. In: NeurIPS. 2020

    work page 2020

  7. [7]

    Estimating heterogeneous treatment effects with right-censored data via causal survival forests

    Yifan Cui et al. “Estimating heterogeneous treatment effects with right-censored data via causal survival forests”. In: Journal of the Royal Statistical Society Series B: Statistical Methodology 85.2 (2023), pp. 179–211

    work page 2023

  8. [8]

    SurvITE: Learning heterogeneous treatment effects from time-to-event data

    Alicia Curth, Changhee Lee, and Mihaela van der Schaar. “SurvITE: Learning heterogeneous treatment effects from time-to-event data”. In: NeurIPS. 2021

    work page 2021

Show all 53 references
  1. [9]

    Nonparametric estimation of heterogeneous treat- ment effects: From theory to learning algorithms

    Alicia Curth and Mihaela van der Schaar. “Nonparametric estimation of heterogeneous treat- ment effects: From theory to learning algorithms”. In: AISTATS. 2021

    2021

  2. [10]

    Causal machine learning for predicting treatment outcomes

    Stefan Feuerriegel et al. “Causal machine learning for predicting treatment outcomes”. In: Nature Medicine (2024)

    2024

  3. [11]

    Orthogonal statistical learning

    Dylan J. Foster and Vasilis Syrgkanis. “Orthogonal statistical learning”. In: The Annals of Statistics 53.3 (2023), pp. 879–908

    2023

  4. [12]

    Estimating individual treatment effects under unob- served confounding using binary instruments

    Dennis Frauen and Stefan Feuerriegel. “Estimating individual treatment effects under unob- served confounding using binary instruments”. In: ICLR. 2023

    2023

  5. [13]

    Model-agnostic meta-learners for estimating heterogeneous treatment effects over time

    Dennis Frauen, Konstantin Hess, and Stefan Feuerriegel. “Model-agnostic meta-learners for estimating heterogeneous treatment effects over time”. In: ICLR. 2025

    2025

  6. [14]

    Estimating Heterogeneous Treatment Effects for General Responses

    Zijun Gao and Trevor Hastie. “Estimating Heterogeneous Treatment Effects for General Responses”. In: arXiv preprint arXiv::2103.04277 (2022)

    2022 arXiv

  7. [15]

    Causal inference in public health

    Thomas A. Glass et al. “Causal inference in public health”. In:Annual Review of Public Health 34 (2013), pp. 61–75

    2013

  8. [16]

    Individualized treatment effects with censored data via fully nonparametric Bayesian accelerated failure time models

    Nicholas C. Henderson et al. “Individualized treatment effects with censored data via fully nonparametric Bayesian accelerated failure time models”. In: Biostatistics 21.1 (2020), pp. 50– 68

    2020

  9. [17]

    Bayesian nonparametric modeling for causal inference

    Jennifer L. Hill. “Bayesian nonparametric modeling for causal inference”. In: Journal of Computational and Graphical Statistics 20.1 (2011), pp. 2017–2040

    2011

  10. [18]

    Estimating heterogeneous survival treatment effect in observational data using machine learning

    Liangyuan Hu, Jiayi Ji, and Fan Li. “Estimating heterogeneous survival treatment effect in observational data using machine learning”. In: Statistics in Medicine 40.21 (2021), pp. 4691– 4713

    2021

  11. [19]

    Identification and estimation of local average treatment effects

    Guido W. Imbens and Joshua D. Angrist. “Identification and estimation of local average treatment effects”. In: Econometrica 62.2 (1994), pp. 467–475

    1994

  12. [20]

    More efficient policy learning via optimal retargeting

    Nathan Kallus. “More efficient policy learning via optimal retargeting”. In: Journal of the American Statistical Association 116.534 (2021), pp. 646–658

    2021

  13. [21]

    DeepSurv: Personalized treatment recommender system using a Cox proportional hazards deep neural network

    Jared L. Katzman et al. “DeepSurv: Personalized treatment recommender system using a Cox proportional hazards deep neural network”. In: BMC medical research methodology 18.1 (2018), pp. 1–12

    2018

  14. [22]

    Semiparametric doubly robust targeted double machine learning: A review

    Edward H. Kennedy. “Semiparametric doubly robust targeted double machine learning: A review”. In: arXiv preprint (2022)

    2022

  15. [23]

    Towards optimal doubly robust estimation of heterogeneous causal effects

    Edward H. Kennedy. “Towards optimal doubly robust estimation of heterogeneous causal effects”. In: Electronic Journal of Statistics 17.2 (2023), pp. 3008–3049

    2023

  16. [24]

    Real-world progression, treatment, and survival outcomes during rapid adoption of immunotherapy for advanced non–small cell lung cancer

    Sean Khozin et al. “Real-world progression, treatment, and survival outcomes during rapid adoption of immunotherapy for advanced non–small cell lung cancer”. In: Cancer 125.22 (2019), pp. 4019–4032. 11

    2019

  17. [25]

    Klein and Melvin L

    John P. Klein and Melvin L. Moeschberger.Survival Analysis: Techniques for censored and truncated data. 2nd ed. New York: Springer New York, 2003

    2003

  18. [26]

    Prediction from randomly right censored data

    Michael Kohler and Kinga Mathe. “Prediction from randomly right censored data”. In:Journal of Multivariate Analysis 80.1 (2002), pp. 73–100

    2002

  19. [27]

    Metalearners for estimating heterogeneous treatment effects using machine learning

    Sören R. Künzel et al. “Metalearners for estimating heterogeneous treatment effects using machine learning”. In: Proceedings of the National Academy of Sciences (PNAS) 116.10 (2019), pp. 4156–4165

    2019

  20. [28]

    A meta-learner for heterogeneous effects in difference-in-differences

    Hui Lan et al. “A meta-learner for heterogeneous effects in difference-in-differences”. In: ICML. V ol. arXiv:2502.04699. 2025

    2025 arXiv

  21. [29]

    Causal effect inference with deep latent-variable models

    Christos Louizos et al. “Causal effect inference with deep latent-variable models”. In:NeurIPS. 2017

    2017

  22. [30]

    On a general class of orthogonal learners for the estimation of heterogeneous treatment effects

    Pawel Morzywolek, Johan Decruyenaere, and Stijn Vansteelandt. “On a general class of orthogonal learners for the estimation of heterogeneous treatment effects”. In: arXiv preprint arXiv:2303.12687 (2023)

    2023 arXiv

  23. [31]

    Quasi-oracle estimation of heterogeneous treatment effects

    Xinkun Nie and Stefan Wager. “Quasi-oracle estimation of heterogeneous treatment effects”. In: Biometrika 108.2 (2021), pp. 299–319

    2021

  24. [32]

    B-learner: Quasi-oracle bounds on heterogeneous causal effects under hidden confounding

    Miruna Oprescu et al. “B-learner: Quasi-oracle bounds on heterogeneous causal effects under hidden confounding”. In: ICML. 2023

    2023

  25. [33]

    Clinical trials in older people

    Kaisu H. Pitkala and Time E. Strandberg. “Clinical trials in older people”. In: Age and Ageing 51.5 (2022), afab282

    2022

  26. [34]

    Robust estimation in sequentially ignorable missing data and causal inference models

    James M. Robins. “Robust estimation in sequentially ignorable missing data and causal inference models”. In: Proceedings of the American Statistical Association on Bayesian Statistical Science (1999), pp. 6–10

    1999

  27. [35]

    Estimation of reression coefficients when some regressors are not always observed

    James M. Robins, Andrea Rotnitzky, and Lue Ping Zhao. “Estimation of reression coefficients when some regressors are not always observed”. In: Journal of the American Statistical Association 89.427 (1994), pp. 846–688

    1994

  28. [36]

    Doubly robust censoring unbiased transformations

    Daniel Rubin and Mark J. van der Laan. “Doubly robust censoring unbiased transformations”. In: The International Journal of Biostatistics 3.1 (2007)

    2007

  29. [37]

    Estimating causal effects of treatments in randomized and nonrandomized studies

    Donald B. Rubin. “Estimating causal effects of treatments in randomized and nonrandomized studies”. In: Journal of Educational Psychology 66.5 (1974), pp. 688–701

    1974

  30. [38]

    BITES: Balanced individual treatment effect for survival data

    S. Schrod et al. “BITES: Balanced individual treatment effect for survival data”. In: Bioinfor- matics 38.1 (2022), pp. 60–67

    2022

  31. [39]

    Meta-learners for partially-identified treatment effects across multiple environments

    Jonas Schweisthal et al. “Meta-learners for partially-identified treatment effects across multiple environments”. In: ICML. 2024

    2024

  32. [40]

    Estimating individual treatment effect: Generalization bounds and algorithms

    Uri Shalit, Fredrik D. Johansson, and David Sontag. “Estimating individual treatment effect: Generalization bounds and algorithms”. In: ICML. 2017

    2017

  33. [41]

    Association of Patient Characteristics and Tumor Genomics with clinical outcomes among patients with non–small cell lung cancer using a clinicogenomic database

    Gaurav Singal et al. “Association of Patient Characteristics and Tumor Genomics with clinical outcomes among patients with non–small cell lung cancer using a clinicogenomic database”. In: Jama 321.14 (2019), pp. 1391–1399

    2019

  34. [42]

    Machine learning estimation of heterogeneous treatment effects with instruments

    Vasilis Syrgkanis et al. “Machine learning estimation of heterogeneous treatment effects with instruments”. In: NeurIPS. 2019

    2019

  35. [43]

    Non-parametric individual treatment effect estimation for survival data with random forests

    Sami Tabib and Denis Larocque. “Non-parametric individual treatment effect estimation for survival data with random forests”. In: Bioinformatics 36.2 (2020), pp. 629–636

    2020

  36. [44]

    Statistical inference for variable importance

    Mark J. van der Laan. “Statistical inference for variable importance”. In: The International Journal of Biostatistics 2.1 (2006), pp. 1–31

    2006

  37. [45]

    van der Laan and James M

    Mark J. van der Laan and James M. Robins. Unified methods for censored longitudinal data and causality. Springer New York, 2003

    2003

  38. [46]

    Targeted maximum likelihood learning

    Mark J. van der Laan and Donald B. Rubin. “Targeted maximum likelihood learning”. In: The International Journal of Biostatistics 2.1 (2006)

    2006

  39. [47]

    Asymptotic statistics

    Aart van der Vaart. Asymptotic statistics. Cambridge: Cambridge University Press, 1998

    1998

  40. [48]

    Inference for treatment-specific survival curves using machine learning

    Ted Westling et al. “Inference for treatment-specific survival curves using machine learning”. In: Journal of the American Statistical Association 119.546 (2023), pp. 1542–1553

    2023

  41. [49]

    Deep Learning for Survival Analysis: A Review

    Simon Wiegrebe et al. “Deep Learning for Survival Analysis: A Review”. In: Artificial Intelligence Review 57.3 (2024), pp. 46–99. 12

    2024

  42. [50]

    Estimating heterogeneous treatment effects on survival outcomes using counterfactual censoring unbiased transformations

    Shenbo Xu et al. “Estimating heterogeneous treatment effects on survival outcomes using counterfactual censoring unbiased transformations”. In: arXiv preprint arXiv:2401.11263 (2024)

    2024 arXiv

  43. [51]

    Treatment heterogeneity for survival outcomes

    Yizhe Xu et al. “Treatment heterogeneity for survival outcomes”. In:Handbook of Matching and Weighting Adjustments for Causal Inference(2023), pp. 445–482

    2023

  44. [52]

    Mining heterogeneous causal effects for personalized cancer treatment

    Weijia Zhang et al. “Mining heterogeneous causal effects for personalized cancer treatment”. In: Bioinformatics 33.15 (2017), pp. 2372–2378

    2017

  45. [53]

    Targeted estimation of heterogeneous treatment effect in ob- servational survival analysis

    Jie Zhu and Blanca Gallego. “Targeted estimation of heterogeneous treatment effect in ob- servational survival analysis”. In: Journal of Biomedical Informatics 107.8 (2020), pp. 103– 474. 13 A Extended related work Semiparametric inference and orthogonal learning: The concept ...

    2020

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.