Pith. sign in

REVIEW 2 major objections 4 minor 131 references

The paper claims that combining a randomised trial with confounded real-world survival data yields per-patient treatment effects with trial-level validity and roughly half the uncertainty, provided the true effect transports across sources.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-03 09:47 UTC pith:I5JOPWPX

load-bearing objection A genuinely useful fusion framework for survival HTE, but the claimed efficiency gain is compromised by an un-matched prior comparison. the 2 major comments →

arxiv 2607.29295 v1 pith:I5JOPWPX submitted 2026-07-31 stat.ME stat.ML

Bayesian fusion forests for heterogeneous treatment effects on survival from randomised and real-world data

Tijn Jacobs , St\'ephanie L. van der Pas , Wessel N. van Wieringen This is my paper
classification stat.ME stat.ML MSC 62F1562G0862N0162P10
keywords Bayesian fusion forestheterogeneous treatment effectssurvival analysisdata fusionconfounding functionaccelerated failure timeBayesian additive regression treesinterval censoring
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper develops a Bayesian nonparametric model for estimating heterogeneous treatment effects on censored survival times by fusing a randomised controlled trial with real-world data. It argues that an accelerated-failure-time decomposition — baseline prognosis, source-specific deviation, treatment effect, and a confounding function — lets the real-world data contribute power and follow-up while the trial pins down the causal effect. The confounding function absorbs whatever bias the real-world treatment assignment carries, so the model never needs to assume the observational source is unconfounded. In simulations the fusion stays unbiased and roughly halves posterior variance compared with trial-only analysis. Applied to HIV antiretroviral therapy, it estimates an average acceleration factor of 1.65 and a posterior probability of benefit above 0.95 for 96.4% of patients, versus 38% from the trial alone.

Core claim

Assume consistency, RCT unconfoundedness, positivity, non-informative censoring, and cross-source transportability of the conditional treatment effect. Then the conditional mean log survival time decomposes as E[logT|A,X,S] = m0(X,S) + τ(X)A + (1-S)Ac(X), where τ is the CATE and c is a confounding function active only in the real-world data. Proposition 2 shows τ is identified from the RCT alone and c is identified given τ, while the RWD alone identifies only the sum τ+c. The Bayesian fusion forest places separate tree-ensemble priors on each component, with a hierarchical Dirichlet process mixture for the error distribution, and thereby estimates τ(x) with lower variance than a trial-only a

What carries the argument

The load-bearing object is the accelerated-failure-time decomposition (Eq. 3) with four separate Bayesian additive regression tree priors: a shared baseline prognosis, a source-specific deviation centred at zero, a depth-penalised treatment-effect forest, and a strongly regularised confounding function. The confounding function is the mechanism that absorbs RWD bias; the shared baseline with zero-centred deviation encodes borrowing between sources; the hierarchical Dirichlet process mixture lets the error law differ across sources while sharing mixture components. A blocked Gibbs sampler with data augmentation handles right- and interval-censored event times.

Load-bearing premise

The whole edifice rests on Assumption 4: that for every covariate profile, the average treatment effect is the same in the trial and the real-world population. This cannot be tested from the data, and if it is wrong, the real-world data drag the estimate toward a biased contrast.

What would settle it

Simulate two sources where the true CATE differs by covariates (e.g., τ_RWD = τ_RCT + g(X) with nonzero g), then fit the fusion forest: the posterior mean of τ will show bias that grows with the RWD sample size, while a trial-only fit stays unbiased. A real-data check would compare fusion estimates against an independent large randomised trial in a population where the RWD composition differs.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Fusion estimates remain unbiased under unmeasured confounding in the RWD and stay at nominal coverage.
  • Posterior variance is roughly half the trial-only variance across simulation settings; RMSE ratio stays below one up to p=500 covariates.
  • The method opens right- and interval-censored outcomes to data fusion.
  • In the HIV application, average acceleration factor 1.65 [1.43; 1.90], and 96.4% of patients have posterior probability of benefit >0.95 vs 38% trial-only.
  • A regression-tree projection of the treatment forest identifies CD8 and race as effect modifiers.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If transportability fails, the RWD contribution will pull the posterior of τ toward the confounded contrast; Assumption 4 is the fragile hinge, so sensitivity analyses comparing fusion with trial-only where the two disagree most would be valuable.
  • The time-invariance of the acceleration factor is an explicit limitation; a time-dependent variant would let delayed-onset or waning effects be captured.
  • The same decomposition could fuse multiple RWD sources, one confounding function per source, as long as τ transports to all of them.
  • The method's per-patient 'certainty to benefit' depends on the calibration established in simulations; in a single real cohort without ground truth, the 96.4% figure inherits that calibration assumption.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 4 minor

Summary. The manuscript proposes the Bayesian fusion forest, a BART-based AFT model for heterogeneous treatment effects on survival that combines an RCT with real-world data. The key identifying device is a confounding function c(x) that absorbs RWD bias, while τ(x) is identified from the RCT under cross-source transportability (Assumptions 1–4). Each component of the decomposition Eq. (3) receives its own regularised tree ensemble, with a hierarchical Dirichlet process mixture for the error distribution. Simulation studies compare the method with a trial-only Bayesian causal forest and several machine-learning baselines, and claim large precision gains and robustness to confounding and covariate dimensionality. The method is applied to ACTG 175 plus MACS, giving an average acceleration factor of 1.65 [1.43;1.90] and a posterior probability of benefit above 0.95 for 96.4% of patients, versus 38% for the trial-only analysis.

Significance. The paper addresses an important problem: combining randomised and real-world evidence for heterogeneous treatment effects on censored survival outcomes without assuming the RWD is unconfounded. The causal decomposition is clean, Propositions 1 and 2 are proved transparently in the supplement, and the implementation is reproducible with public code and 1000-replication simulations. If the efficiency claim is established, this would be a useful contribution. However, the central quantitative claim—that fusion reduces posterior variance by about 88% and interval width by 62% relative to a trial-only analysis—rests on a comparator whose hyperparameters are not specified, so part or all of the apparent gain could be a prior-shrinkage artifact. The application's headline conclusion also depends on the untestable Assumption 4, and the reported certainty should be conditioned on that assumption or accompanied by a sensitivity analysis.

major comments (2)
  1. [§2.5 and §3; Table S4 in Supplement S.5.1] The comparator for the main precision claim is the 'single-source counterpart' AFT Bayesian causal forest (Jacobs, 2026), but its hyperparameters are never given. The fusion model deliberately uses a strongly shrunk treatment-effect forest: τ∼BART(100, 1/2, 0.95, 3) with kτ=1/2 (§2.5). If the trial-only causal forest uses standard BART defaults (e.g., J=200, k=1, α=0.95, β=2), the reported reductions in posterior variance (88%) and credible-interval width (62%) in Table S4, and the RMSE/variance ratios in Figures 1–2 and S2, may reflect prior shrinkage rather than information borrowed from the RWD. This is load-bearing for the paper's central claim that fusion estimates τ with lower variance than a trial-only analysis. Please re-run the trial-only causal forest with exactly the same τ-forest prior (J=100, k=1/2, α=0.95, β=3), and also, as a robustness check, the fusion with the trial-onl
  2. [§2.1 (Assumption 4) and §4 (Table 2, Figure 4)] The application's conclusion—'benefit for nearly every patient' and the precision gain at the individual level—is conditional on cross-source transportability, Assumption 4: E[logT(1)−logT(0)|X,S=1] = E[logT(1)−logT(0)|X,S=0]. Proposition 2(iii) shows the RWD identifies only the composite τ+c, so if Assumption 4 fails, the fusion's conditional benefit statements inherit the RWD bias. Restricting both sources to men with CD4 200–500 is a reasonable design choice but does not verify the assumption. The manuscript should either include a sensitivity analysis that perturbs the transportability assumption (e.g., by allowing a shift δ(x) in the CATE transport equation and re-examining the posterior probability of benefit) or temper the conclusion to explicitly state that the certainty holds only under Assumption 4. This is not a circularity concern—Proposition 2(i) identifies τ from the RCT in
minor comments (4)
  1. [§2.5] The statement that the hyperparameters are 'tuned further by cross-validation' is never operationalised. State what is tuned (e.g., kτ, kc, or tree-structure parameters), on which data, and with what criterion. The simulation and application both use 'default parameters', so the role of cross-validation is unclear.
  2. [§3; reference Jacobs (2026)] The trial-only comparator is described only via a software-package reference. Provide the exact model specification, including the BART prior hyperparameters and any treatment-effect shrinkage, so the comparison is reproducible.
  3. [Table 1 and Supplement S.5.2] The comparison with the four machine-learning methods is reasonable but the paper should be careful when saying it 'improves on all' of them: those methods are not designed for data fusion and are adapted in ways that may be suboptimal. The supplementary discussion appropriately notes the lack of ground truth in the application, but the main text should carry that caveat more explicitly.
  4. [§5] The limitation section correctly notes that the acceleration factor is assumed time-invariant. It would be helpful to also mention that the full model has no posterior-concentration or consistency theorem for the four-forest HDPM specification; the empirical calibration in the simulation is the current support. If the authors intend a theoretical claim, a proof or reference is needed.

Circularity Check

0 steps flagged

No significant circularity: τ is identified from the RCT contrast alone; the confounding-function decomposition is an explicit defining identity, and the empirical variance claims are simulation results, not derivations from inputs.

full rationale

Eq. (2) defines c(x) as the RWD treated-minus-control contrast minus τ(x), so Proposition 1 is the algebraic identity obtained by inserting that definition into the conditional mean; this is an explicit defining device, not a hidden circularity. The load-bearing identification is Proposition 2(i): τ(x) is the difference of the two RCT conditional means (Supplementary S.1.2, Eq. S11), which does not involve c or the RWD. Thus the central estimand is not defined in terms of the confounding function, and the RWD is used only as additional information under Assumption 4. Proposition 2(ii) is the same definition read backwards and is presented as such. The empirical claims (variance reduction, posterior benefit probabilities) come from simulations and an application, not from a derivation that equates inputs and outputs. There is a real benchmarking concern: the trial-only comparator is 'AFT Bayesian causal forest (Jacobs, 2026)' with unspecified hyperparameters, while the fusion's τ forest uses larger shrinkage (kτ=1/2, Jτ=100, βτ=3), so part of the reported precision gain may be prior-shrinkage rather than borrowed information. That is a correctness/comparability risk, not a circularity: no equation in the paper forces the variance ratio by construction, and the paper's identification of τ is independent of this comparison. Self-citations (Jacobs 2026, Jacobs et al. 2025) are method/software references, not load-bearing premises; no uniqueness theorem or ansatz is imported. Assumption 4 is untestable but explicitly acknowledged as the fragile substantive assumption.

Axiom & Free-Parameter Ledger

4 free parameters · 7 axioms · 0 invented entities

No new physical or structural entities are postulated; the confounding function, source deviation, and HDPM are model components, not independent empirical claims. The ledger is dominated by the explicit causal assumptions and by hand-chosen/calibrated hyperparameters that the results depend on.

free parameters (4)
  • Component-specific BART regularization ladder = J=(200,50,100,50); k=1 for sh,d,c and k_τ=1/2; α_c=0.25, β_τ=β_c=3
    Hand-chosen prior strengths that control the flexibility of baseline, deviation, treatment effect, and confounding function; the RWD separation of τ and c depends on this ladder.
  • HDPM error-model calibrations = ν=3; λ from empirical residual variance; σ²_θ from Yamato matching to a preliminary log-normal AFT fit with q=0.5
    Calibrated to the data before fitting; they set the scale and shape flexibility of the source-specific error distributions.
  • HDPM concentration hyperparameters = a_γ=a_M=2, b_γ=b_M=0.1 (prior mean 20, mode 10)
    Hand-specified; controls the number of shared atoms and the degree of error-distribution sharing across sources.
  • Target-population mixing weights α = α=(n0,n1) recommended; alternatives (1,1) or extremes
    Choice of target population changes the reported average treatment effect and its interval; the default is data-dependent.
axioms (7)
  • domain assumption Assumption 1: Consistency (T=T(a), C=C(a) under A=a)
    Section 2.1; standard in the potential-outcomes framework.
  • domain assumption Assumption 2: RCT unconfoundedness
    Section 2.1; holds by design in a randomized trial.
  • domain assumption Assumption 3: Positivity in both sources
    Section 2.1; ensures conditional expectations are well-defined.
  • domain assumption Assumption 4: Cross-source transportability of CATE
    Section 2.1; load-bearing and untestable; if false, τ and c are not separately identified from RWD (Prop. 2iii).
  • domain assumption Assumption 5: Conditionally non-informative censoring
    Section 2.1; allows the censored likelihood and data augmentation to recover conditional means.
  • domain assumption Time-invariant acceleration factor exp{τ(X)}
    Discussion; treatment effect is independent of time; authors acknowledge this may be inappropriate for delayed or waning effects.
  • domain assumption Causal identification of acceleration factors per Brathovde et al. (2026)
    Section 2.1; cited external theorem supplies the causal interpretation of exp{τ(X)}; its proof is not reproduced.

pith-pipeline@v1.3.0-daily-deepseek · 32044 in / 20905 out tokens · 226515 ms · 2026-08-03T09:47:42.516019+00:00 · methodology

0 comments
read the original abstract

We develop the Bayesian fusion forest, a nonparametric framework to estimate heterogeneous treatment effects on survival outcomes by combining a randomised controlled trial and real-world data. The framework relaxes the unconfoundedness assumption on the real-world data by assuming instead that the treatment effect transports across the two sources. Our method opens up right- and interval-censored outcomes to data fusion. We model the survival time with an accelerated failure time decomposition into a shared baseline prognosis, a source-specific deviation, a treatment effect, and a confounding function. The confounding function absorbs the confounding bias in the real-world data. Each component receives a Bayesian tree ensemble prior. The shared baseline prognosis borrows strength across sources, while the deviation captures between-source heterogeneity. A hierarchical Dirichlet process mixture models the error distribution nonparametrically. A simulation study shows efficiency gains over a trial-only analysis across varying levels of confounding and between-source heterogeneity. We combine the ACTG 175 trial with the Multicenter AIDS Cohort Study to estimate the effect of combination antiretroviral therapy for HIV. The fusion identifies a benefit for nearly every patient whereas the trial alone is inconclusive.

Figures

Figures reproduced from arXiv: 2607.29295 by St\'ephanie L. van der Pas, Tijn Jacobs, Wessel N. van Wieringen.

Figure 1
Figure 1. Figure 1: CATE metrics under varying unmeasured confounding strength [PITH_FULL_IMAGE:figures/full_fig_p012_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: CATE metrics as the number of covariates [PITH_FULL_IMAGE:figures/full_fig_p014_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Kaplan–Meier curves of event-free survival, by data source (RCT, RWD) and treatment [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Subject-level acceleration factor for every patient in the trial-aligned cohort, from [PITH_FULL_IMAGE:figures/full_fig_p016_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Regression-tree summary of the acceleration factor [PITH_FULL_IMAGE:figures/full_fig_p017_5.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

131 extracted references · 2 canonical work pages · 1 internal anchor

  1. [1]

    Biometrical Journal , year =

    Kong, Jianfen and Zhang, Shuhong , title =. Biometrical Journal , year =

  2. [2]

    Journal of Computational and Graphical Statistics , year =

    Barnwal, Avinash and Cho, Hyunsu and Hocking, Toby , title =. Journal of Computational and Graphical Statistics , year =

  3. [3]

    , title =

    Antoniak, Charles E. , title =. The Annals of Statistics , year =

  4. [4]

    and George, Edward I

    Chipman, Hugh A. and George, Edward I. and McCulloch, Robert E. , title =. The Annals of Applied Statistics , year =

  5. [5]

    and West, Mike , title =

    Escobar, Michael D. and West, Mike , title =. Journal of the American Statistical Association , year =

  6. [6]

    Richard and Murray, Jared S

    Hahn, P. Richard and Murray, Jared S. and Carvalho, Carlos M. , title =. Bayesian Analysis , year =

  7. [7]

    Statistical Science , year =

    Hastie, Trevor and Tibshirani, Robert , title =. Statistical Science , year =. doi:10.1214/ss/1009212815 , publisher =

    arXiv

  8. [8]

    and Katzenstein, David A

    Hammer, Scott M. and Katzenstein, David A. and Hughes, Michael D. and Gundacker, H. and Schooley, R. T. and Haubrich, R.H. and Henry, W.K. and Lederman, M.M. and Phair, J.P. and Niu, M. and Hirsch, M.S. and Merigan, T.C. , title =. New England Journal of Medicine , year =

  9. [9]

    and Louis, Thomas A

    Henderson, Nicholas C. and Louis, Thomas A. and Rosner, Gary L. and Varadhan, Ravi , title =. Biostatistics , year =

  10. [10]

    The hazards of hazard ratios , journal =

    Hern. The hazards of hazard ratios , journal =. 2010 , volume =

    2010

  11. [11]

    Statistics in Medicine , year =

    Hupf, Bradley and Bunn, Veronica and Lin, Jianchang and Dong, Cheng , title =. Statistics in Medicine , year =

  12. [12]

    , title =

    Ishwaran, Hemant and James, Lancelot F. , title =. Journal of the American Statistical Association , year =

  13. [13]

    ShrinkageTrees: An R Package for Bayesian Tree Ensembles for Survival Analysis and Causal Inference

    Jacobs, Tijn , title =. 2026 , eprint =. doi:10.48550/arXiv.2606.12317 , url =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2606.12317 2026

  14. [14]

    and van der Pas, St

    Jacobs, Tijn and van Wieringen, Wessel N. and van der Pas, St. Horseshoe Forests for High-Dimensional Causal Survival Analysis , year =. doi:10.48550/arXiv.2507.22004 , url =. 2507.22004 , archivePrefix =

    Pith/arXiv arXiv doi:10.48550/arxiv.2507.22004

  15. [15]

    Journal of Statistical Software , year =

    Kapelner, Adam and Bleich, Justin , title =. Journal of Statistical Software , year =

  16. [16]

    and Lu, Xiaomin and Zhang, Min and Davidian, Marie and Tsiatis, Anastasios A

    Juraska, Michal and Gilbert, Peter B. and Lu, Xiaomin and Zhang, Min and Davidian, Marie and Tsiatis, Anastasios A. , year =

  17. [17]

    and Ostrow, David G

    Kaslow, Richard A. and Ostrow, David G. and Detels, Roger and Phair, John P. and Polk, B. Frank and Rinaldo, Charles R. and. The. American Journal of Epidemiology , year =

  18. [18]

    and Darby, Graham and Richman, Douglas D

    Larder, Brendan A. and Darby, Graham and Richman, Douglas D. , title =. Science , year =

  19. [19]

    , title =

    MacEachern, Steven N. , title =. ASA Proceedings of the Section on Bayesian Statistical Science , year =

  20. [20]

    The nested

    Rodr. The nested. Journal of the American Statistical Association , year =

  21. [21]

    Statistica Sinica , year =

    Sethuraman, Jayaram , title =. Statistica Sinica , year =

  22. [22]

    and Logan, Brent R

    Sparapani, Rodney A. and Logan, Brent R. and Maiers, Martin J. and Laud, Purushottam W. and McCulloch, Robert E. , title =. Biometrics , year =

  23. [23]

    and Jordan, Michael I

    Teh, Yee W. and Jordan, Michael I. and Beal, Matthew J. and Blei, David M. , title =. Journal of the American Statistical Association , year =

  24. [24]

    , title =

    Wang, Chong and Paisley, John and Blei, David M. , title =. Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics , year =

  25. [25]

    The Annals of Probability , year =

    Yamato, Hajime , title =. The Annals of Probability , year =

  26. [26]

    and Baird, Donna , title =

    Yang, Mingan and Dunson, David B. and Baird, Donna , title =. Computational Statistics and Data Analysis , year =

  27. [27]

    Bayesian Analysis , year =

    Antonelli, Joseph and Parmigiani, Giovanni and Dominici, Francesca , title =. Bayesian Analysis , year =

  28. [28]

    Medical Decision Making , year =

    Bagust, Adrian and Beale, Sophie , title =. Medical Decision Making , year =

  29. [29]

    Medical Decision Making , year =

    Bell Gorrod, Helen and Kearns, Benjamin and Stevens, John and Thokala, Praveen and Labeit, Alexander and Latimer, Nicholas and Tyas, David and Sowdani, Ahmed , title =. Medical Decision Making , year =

  30. [30]

    Guyot, Patricia and Ades, A. E. and Ouwens, Mario J. N. M. and Welton, Nicky J. , title =. BMC Medical Research Methodology , year =

  31. [31]

    Guyot, Patricia and Ades, A. E. and Beasley, Matthew and Lueza, B. Extrapolation of survival curves from cancer trials using external information , journal =. 2017 , volume =

    2017

  32. [32]

    , title =

    Jackson, Christopher H. , title =. BMC Medical Research Methodology , year =

  33. [33]

    Proceedings of the 32nd International Conference on Neural Information Processing Systems , pages =

    Kallus, Nathan and Puli, Aahlad Manas and Shalit, Uri , title =. Proceedings of the 32nd International Conference on Neural Information Processing Systems , pages =. 2018 , publisher =

    2018

  34. [34]

    , title =

    Latimer, Nicholas R. , title =. Medical Decision Making , year =

  35. [35]

    and Evans, Robin J

    Lin, Xi and Tarp, Jens M. and Evans, Robin J. , title =. 2024 , note =

    2024

  36. [36]

    , title =

    Rubin, Donald B. , title =. The Annals of Statistics , year =

  37. [37]

    and Jackson, Christopher H

    Timmins, Iain R. and Jackson, Christopher H. and colleagues , title =. 2025 , note =

    2025

  38. [38]

    and Qiu, Sky and van der Laan, Lars , title =

    van der Laan, Mark J. and Qiu, Sky and van der Laan, Lars , title =. 2024 , note =

    2024

  39. [39]

    Advances in Therapy , year =

    Vyas, Manish and Hettle, Robert and Rana, Mehul and Stevenson, Katie and others , title =. Advances in Therapy , year =

  40. [40]

    and Murray, Jared S

    Woody, Spencer and Carvalho, Carlos M. and Murray, Jared S. , title =. Journal of Computational and Graphical Statistics , year =

  41. [41]

    Bernoulli , year =

    Yang, Shu and Liu, Siyi and Zeng, Donglin and Wang, Xiaofei , title =. Bernoulli , year =

  42. [42]

    Proceedings of the 33rd ACM International Conference on Information and Knowledge Management , pages =

    Yao, Dong and Tang, Caizhi and Cui, Qing and Li, Longfei , title =. Proceedings of the 33rd ACM International Conference on Information and Knowledge Management , pages =. 2024 , isbn =. doi:10.1145/3627673.3679593 , abstract =

    arXiv 2024

  43. [43]

    Lifetime Data Analysis , year =

    Ye, Xin and Yang, Shu and Wang, Xiaofei and Liu, Yanyan , title =. Lifetime Data Analysis , year =

  44. [44]

    and Cook, Richard J

    Aalen, Odd O. and Cook, Richard J. and R. Does. Lifetime Data Analysis , year =. doi:10.1007/s10985-015-9335-y , publisher =

    work page doi:10.1007/s10985-015-9335-y

  45. [45]

    Brathovde, Mari and Putter, Hein and Valberg, Morten and Post, Richard A. J. , title =. 2026 , note =

    2026

  46. [46]

    and Sackett, David L

    Cook, Richard J. and Sackett, David L. , title =. BMJ , year =

  47. [47]

    Cox, D. R. and Oakes, D. , title =

  48. [48]

    and Royston, Patrick and Clements, Mark , title =

    Crowther, Michael J. and Royston, Patrick and Clements, Mark , title =. Biostatistics , year =

  49. [49]

    Structural accelerated failure time models for survival analysis in studies with time-varying treatments , journal =

    Hern. Structural accelerated failure time models for survival analysis in studies with time-varying treatments , journal =. 2005 , volume =

    2005

  50. [50]

    Biometrics , year =

    Hougaard, Philip , title =. Biometrics , year =

  51. [51]

    and Prentice, Ross L

    Kalbfleisch, John D. and Prentice, Ross L. , title =. 2002 , doi =

    2002

  52. [52]

    and Klein, John P

    Keiding, Niels and Andersen, Per K. and Klein, John P. , title =. Statistics in Medicine , year =

  53. [53]

    , title =

    Martinussen, Torben and Vansteelandt, Stijn and Andersen, Per K. , title =. Lifetime Data Analysis , year =

  54. [54]

    Post, Richard A. J. and van den Heuvel, Edwin R. and Putter, Hein , title =. Lifetime Data Analysis , year =

  55. [55]

    Royston, Patrick and Parmar, Mahesh K. B. , title =. Statistics in Medicine , year =

  56. [56]

    Royston, Patrick and Parmar, Mahesh K. B. , title =. BMC Medical Research Methodology , year =

  57. [57]

    Uno, Hajime and Claggett, Brian and Tian, Lu and Inoue, Eisuke and Gallo, Paul and Miyata, Toshio and Schrag, Deborah and Takeuchi, Masahiro and Uyama, Yoshiaki and Zhao, Lihui and Skali, Hicham and Solomon, Scott and Jacobus, Susanna and Hughes, Michael and Packer, Milton and Wei, L. J. , title =. Journal of Clinical Oncology , year =

  58. [58]

    Wei, L. J. , title =. Statistics in Medicine , year =

  59. [59]

    Bayesian Analysis , year =

    Evans, Michael and Moshonov, Hadas , title =. Bayesian Analysis , year =

  60. [60]

    Statistica Sinica , year =

    Gelman, Andrew and Meng, Xiao-Li and Stern, Hal , title =. Statistica Sinica , year =

  61. [61]

    Variable importance measures for heterogeneous treatment effects , year =

    Hines, Oliver and D. Variable importance measures for heterogeneous treatment effects , year =

  62. [62]

    and Carlin, Bradley P

    Hobbs, Brian P. and Carlin, Bradley P. and Mandrekar, Sumithra J. and Sargent, Daniel J. , title =. Biometrics , year =

  63. [63]

    Journal of Causal Inference , year =

    Levy, Jonathan and van der Laan, Mark and Hubbard, Alan and Pirracchio, Romain , title =. Journal of Causal Inference , year =

  64. [64]

    2023 , note =

    Li, Haodong and Hubbard, Alan and van der Laan, Mark , title =. 2023 , note =

    2023

  65. [65]

    , title =

    Neuenschwander, Beat and Capkun-Niggli, Gorana and Branson, Michael and Spiegelhalter, David J. , title =. Clinical Trials , year =

  66. [66]

    2024 , note =

    Shin, Heejun and Linero, Antonio and Audirac, Michelle and Irene, Kezia and Braun, Danielle and Antonelli, Joseph , title =. 2024 , note =

    2024

  67. [67]

    and others , title =

    Ziersen, Simon C. and others , title =. 2024 , note =

    2024

  68. [68]

    and George, Edward I

    Chipman, Hugh A. and George, Edward I. and McCulloch, Robert E. , title =. Journal of the American Statistical Association , year =

  69. [69]

    , title =

    Freund, Yoav and Schapire, Robert E. , title =. Journal of Computer and System Sciences , year =

  70. [70]

    , title =

    Friedman, Jerome H. , title =. The Annals of Statistics , year =

  71. [71]

    , title =

    Hill, Jennifer L. , title =. Journal of Computational and Graphical Statistics , year =

  72. [72]

    Annual Review of Statistics and Its Application , year =

    Hill, Jennifer and Linero, Antonio and Murray, Jared , title =. Annual Review of Statistics and Its Application , year =

  73. [73]

    , title =

    Linero, Antonio R. , title =. Journal of the American Statistical Association , year =

  74. [74]

    and Logan, Brent R

    Sparapani, Rodney A. and Logan, Brent R. and McCulloch, Robert E. and Laud, Purushottam W. , title =. Statistics in Medicine , year =

  75. [75]

    On Theory for

    Ro. On Theory for. Proceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics , pages =. 2019 , editor =

    2019

  76. [76]

    Posterior concentration for

    Ro. Posterior concentration for. The Annals of Statistics , year =

  77. [77]

    , title =

    Rubin, Donald B. , title =. Journal of Educational Psychology , year =. doi:10.1037/h0037350 , publisher =

    doi:10.1037/h0037350

  78. [78]

    , title =

    Rubin, Donald B. , title =. Annals of Statistics , year =. doi:10.1214/aos/1176344064 , publisher =

    arXiv

  79. [79]

    , title =

    Swindell, William R. , title =. Experimental Gerontology , volume =. 2009 , doi =

    2009

  80. [80]

    Biostatistics , volume =

    Shyr, Cathy and Ren, Boyu and Patil, Prasad and Parmigiani, Giovanni , title =. Biostatistics , volume =. 2025 , publisher =

    2025

Showing first 80 references.