$\texttt{AMEND++}$: Benchmarking Eligibility Criteria Amendments in Clinical Trials

Jacob Aptekar; Jimeng Sun; Mandis Beigi; Trisha Das

REVIEW 3 major objections 3 minor 19 references

Eligibility-criteria amendment risk is predictable from a trial's initial protocol text, and a revision-aware pretraining method consistently improves that prediction, reaching AUROC 0.714 on a 162,000-trial benchmark.

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 11:26 UTC pith:GIQ5MOJC

load-bearing objection Useful new benchmark and task, but the CAMLM improvement claim is not yet secure because the pretraining split is never defined; worth a serious referee. the 3 major comments →

arxiv 2601.06300 v2 pith:GIQ5MOJC submitted 2026-01-09 cs.CL cs.AIcs.LG

texttt{AMEND++}: Benchmarking Eligibility Criteria Amendments in Clinical Trials

Trisha Das , Mandis Beigi , Jacob Aptekar , Jimeng Sun This is my paper

classification cs.CL cs.AIcs.LG

keywords eligibility criteria amendmentsclinical trial protocolsamendment predictionchange-aware masked language modelingLLM label denoisingtrial designnatural language processingprotocol version histories

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 tries to establish that a clinical trial's future eligibility-criteria amendments can be forecast from the initial protocol, before the trial runs. It builds two datasets from public trial version histories — one large and raw, one denoised with an LLM to keep only substantive changes — and it proposes a pretraining strategy that focuses a masked-language model on the parts of eligibility text that historically get revised. Across encoders and classifiers, that strategy gives small but consistent gains, with the best fine-tuned model reaching AUROC 0.714 on the full dataset. The payoff, if true, is that sponsors and regulators could flag fragile protocols early and rework the criteria most likely to be amended, saving time and cost.

Core claim

The paper claims that the text of the initial eligibility criteria, plus basic trial metadata, carries predictive signal about whether those criteria will be amended in later versions. It introduces eligibility-criteria amendment prediction as a binary classification task and releases AMEND++: AMEND, a 161,970-trial benchmark built by scraping version histories from the public registry, and AMEND_LLM, a 64,641-trial subset whose training labels are produced by an LLM that classifies edits into added, removed, and modified criteria and ORs the three signals. The central empirical claim is that CAMLM, which masks tokens inside historically unstable spans at high probability and other tokens at

What carries the argument

The load-bearing object is the change-aware masking policy, CAMLM. It takes pairs of an earlier and a final version of a trial's eligibility criteria, computes token-level diffs, and trains a standard masked-language objective with a high masking probability (0.8) inside spans that were deleted or replaced and a low probability (0.05) elsewhere. This biases the encoder toward the kind of content that actually gets amended. The other machinery is the label-construction pipeline: human annotators review only the test split, and an LLM decomposes edits into added, removed, and modified criteria, yielding 99.34% agreement with humans on the test set versus 96.21% for raw registry labels.

Load-bearing premise

The gold-standard negative labels are never independently reviewed: whenever the registry records no eligibility amendment, the human annotators return label 0, so the entire test-set evaluation silently trusts the registry's completeness; if the registry under-reports amendments, every reported score is built on biased ground truth.

What would settle it

Sample several hundred trials that the registry labels as having no eligibility amendment, have two annotators apply the paper's own addition/removal/modification rules to the first and final eligibility versions, and compare with the registry labels; a mismatch rate well above the paper's assumed rarity would invalidate the benchmark's negative labels and the model scores.

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

Share X Bluesky LinkedIn Reddit HN

If this is right

Initial eligibility text can be scored at design time, so amendment-prone criteria can be targeted before submission.
LLM-denoised labels train better models than raw registry labels, improving AUROC from 0.672 to 0.681 on the same human-labeled test set.
CAMLM's gains are consistent across BERT-style, biomedical, and long-context encoders and across logistic-regression, random-forest, and fine-tuned classifiers.
The released benchmark includes all intermediate eligibility versions, not just first and final, so the history is available for future modeling.
The test set is shared across both datasets and independently annotated, giving a stable evaluation point.

Where Pith is reading between the lines

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

Editorial inference: A practical deployment would need to decide whether a 0.70–0.71 AUROC is enough to act on; the utility is likely as a triage screen that selects protocols for human review, not as an automated veto.
Editorial inference: Because the dataset records the full version history, the binary task could be extended to predict how many amendments will occur, when they will happen, or which specific criteria will change — the paper's add/remove/modify decomposition already defines those substeps.
Editorial inference: The change-aware masking recipe is not specific to eligibility criteria; any document type with version histories, such as consent forms or statistical analysis plans, could receive the same treatment.
Editorial inference: The cheapest way to harden the benchmark is an independent audit of trials the registry marks as un-amended; if those negative labels contain hidden errors, all reported scores shift.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

Useful new benchmark and task, but the CAMLM improvement claim is not yet secure because the pretraining split is never defined; worth a serious referee.

read the letter

Plainly: this is a useful paper. It is the first to frame eligibility-criteria amendment forecasting as a supervised NLP problem, and it will release a large-scale versioned dataset from ClinicalTrials.gov. That alone is a real contribution. The LLM denoising decomposition into added/removed/modified criteria is sensible, and the evaluation against a human-labeled test set with bootstrap and DeLong tests is more careful than most benchmark papers.

The soft spots, in order of importance. First and most load-bearing: the paper never states whether the CAMLM pretraining corpus includes the test trials. CAMLM builds version pairs (initial, final) for every trial with a recorded history and masks tokens in unstable spans. If any test trial appears in that corpus, the encoder has effectively been told which spans in the initial EC were later changed—that is the label. The reported improvements are small (AUROC +1.4–1.5%, AUPRC +2.1%), so even a handful of leaked trials could explain the entire gain. The authors need to state explicitly that test trials were excluded from pretraining and, ideally, re-run the comparison with a clean split. This is an easy fix but a required one.

Second, Algorithm 1 inherits the registry's label-0 as ground truth without independent verification. The authors acknowledge this in the Limitations, so it is not hidden, but it means the gold standard is only as good as ClinicalTrials.gov's amendment flags. That is a reasonable starting point, but the paper should discuss the direction of the bias.

Third, the LLM/human agreement of 99.34% is partly circular because both follow the same rule set. It measures instruction-following more than label validity. That is fine for scaling, but it should not be presented as independent confirmation.

The absolute performance is moderate (AUROC ~0.71), so this is a triage tool, not a replacement for expert review. Still, the task is clearly valuable to trial designers, and the benchmark fills a real gap.

This is for researchers in clinical NLP and trial operations. It deserves a serious referee; I would send it out but require the pretraining split to be clarified and the experiments redone without the leakage possibility before acceptance. I would also ask for the data and code to be released as part of the review.

Referee Report

3 major / 3 minor

Summary. The paper introduces eligibility criteria amendment prediction as a binary classification task: given the initial eligibility criteria (EC) text and metadata from ClinicalTrials.gov, predict whether the EC will be amended in later protocol versions. The authors release two datasets: AMEND (161,970 trials with raw registry-derived labels) and AMEND_LLM (64,641 trials with LLM-denoised labels). They also propose CAMLM, a change-aware masked-language-modeling pretraining strategy that up-weights masking of tokens in historically unstable spans. Evaluation with BioBERT/BERT/Longformer backbones and logistic-regression/random-forest/fine-tuning classifiers shows moderate AUROC (best 0.714 on AMEND, 0.697 on AMEND_LLM) and consistent CAMLM gains. The paper reports bootstrap confidence intervals and DeLong tests, and includes ablations over pretraining strategies.

Significance. If the labels are trustworthy and the pretraining procedure is free of leakage, the paper makes a useful contribution: it is the first ML-ready benchmark for amendment prediction, releases longitudinal EC version histories at scale, and proposes a revision-aware pretraining objective that transfers across encoders. The evaluation protocol is standard, with non-parametric bootstrap errors and DeLong significance tests, and the ablation study (Table 5) is a positive feature. However, the benchmark's validity rests on two unverified assumptions: that ClinicalTrials.gov label-0 entries are true negatives, and that CAMLM pretraining does not include the test trials' future EC versions. Because the reported gains are small (+1.4–1.5% AUROC, +4% relative accuracy), even modest label noise or leakage could change the conclusions.

major comments (3)

[§3.3, Algorithm 1; §5 Limitations] Algorithm 1 returns 0 whenever ctgov_label=0, so human annotators never review non-amended trials. The gold-standard test set (shared between AMEND and AMEND_LLM) therefore inherits every raw registry label-0 without verification. The Limitations section acknowledges this assumption but only says false negatives 'are expected to be rare' without supporting evidence. Because all reported AUROC/accuracy numbers depend on these unverified negatives, the benchmark's central validity claim is unsecured. The authors should either manually audit a random sample of label-0 test trials, or provide a sensitivity analysis under plausible false-negative rates.
[§3.4, Eq. (4); Tables 3–5] The paper never states that the 6,067 test trials are excluded from the CAMLM pretraining corpus. Eq. (4) constructs pairs (E^(t)_i, E^(T)_i) from recorded EC histories; if a test trial's initial and final ECs both appear as a pretraining pair, the MLM objective directly trains the encoder to predict tokens whose later change is the exact label signal. The ablation in Table 5 compares CAMLM to MLM and Span MLM on the same corpus, so all variants could benefit equally from any leakage, but CAMLM is uniquely vulnerable because it selectively masks changed spans. The reported gains are small; a held-out pretraining split must be specified and the experiments re-run on a non-leaking corpus before the CAMLM improvement claim is credible.
[§4.5, Table 6; §3.3] The 99.34% agreement between LLM-generated labels and human annotations is partly circular. The LLM prompts are explicitly instructed to 'abide by the rules in Algorithm 1', and the human annotators apply the same Algorithm 1. Thus Table 6 largely measures prompt adherence to a shared rule set, not independent label fidelity. The problem is compounded by the early-return in Algorithm 1: both the LLM (through the ctgov_label shortcut? — at least through the same rules) and the humans are constrained by the registry label. An independent evaluation, ideally with clinicians blinded to ctgov_label and to the LLM outputs, is needed to establish that AMEND_LLM labels are truly denoised.

minor comments (3)

[§4.2] The text says 'The exact prompts used for denoising are provided in Appendix 5,' but the appendix is lettered 'Appendix A' and the prompts are Figures 4–6. Please fix the cross-reference.
[Algorithm 1] There is a formatting artifact in Rule 3: 'if any criterion is modified such that:then' has a stray colon. Please clean up the pseudo-code.
[Table 2] The header 'Train trials (1 / 0)' is ambiguous. It seems to indicate positive/negative counts, but this should be stated explicitly in the caption.

Circularity Check

2 steps flagged

CAMLM pretraining pairs initial and final EC versions for every trial and never states that test trials are excluded, so reported gains may be inflated by future-version leakage; the LLM-denoiser validation is also circular because both human and LLM labels implement the same Algorithm 1 shortcut.

specific steps

fitted input called prediction [Section 3.4 (Eq. 4; pretraining pair construction) and Sections 3.2/4.3 (task definition and reported gains)]
"Given a version pair ( E(t) i , E(T) i ), we compute a token-level diff ΔE(t)i = Diff(E(t)i, E(T)i), which identifies EC content that is deleted or replaced in the final version."

The downstream label y_i is defined by whether E_i^(T) substantively differs from E_i^(0) (Sec. 3.2; Algorithm 1). CAMLM pretraining on the same trial's pair (E_i^(0), E_i^(T)) trains the encoder to predict exactly the tokens in the diff spans, i.e., to reconstruct the final-version signal that defines the amendment label, from text that is not available at trial initiation. The paper never states that the 6,067 test trials are excluded from this pretraining corpus, so the reported +1.4-1.5% AUROC / +2.1% AUPRC gains may measure exposure to each test trial's final EC version rather than a forecast from the initial protocol.
self definitional [Section 3.3 (Human Annotation; Label Denoising using LLM), Algorithm 1, Table 6]
"if ctgov_label = 0 then return 0 // Use ctgov label as ground truth ... Instructions provided to the LLM also abide by the rules in Algorithm 1."

Human 'gold-standard' annotations never independently review non-amended trials: Algorithm 1 outputs 0 whenever ClinicalTrials.gov does. The LLM denoiser is explicitly instructed to follow the same Algorithm 1. Consequently, the 99.34% agreement in Table 6 largely measures that both pipelines hardwire the same registry shortcut on the 3,281 negative trials, not independent evidence of label fidelity. The claimed 'high-quality amendment labels' (contribution 2) are validated against a label constructed by the same rule, making the agreement partly definitional.

full rationale

Two load-bearing steps reduce, as written, to their own inputs. First, CAMLM's pretraining signal is the diff between an earlier and the final EC version - the same comparison that defines the amendment label (Eq. 3 and Algorithm 1). Because the paper constructs pretraining pairs for 'each clinical trial' and never excludes the 6,067 test trials, the reported CAMLM improvements over plain MLM may be driven by having seen the final EC version of each test trial during pretraining; the task definition (Sec. 3.2) requires input available at initiation, so this is a target-information leak rather than a clean forecast. Second, the human test labels and the LLM denoiser both implement Algorithm 1, including the 'ctgov_label = 0 -> return 0' shortcut, so Table 6's agreement is partly by construction and cannot validate the denoising claim. The paper's Limitations acknowledges the ctgov_label=0 assumption but does not connect it to the denoiser agreement or to the CAMLM pretraining split. The prediction experiments are otherwise anchored in a real benchmark and are not circular via self-citation; the cited prior work is not load-bearing. If the authors later document a time-based pretraining split excluding test trials and independent human review of label-0 trials, the central claims would become non-circular; as written, they are not fully secured.

Axiom & Free-Parameter Ledger

2 free parameters · 5 axioms · 0 invented entities

The central claim rests on the correctness of raw CTGov label-0 values, the completeness of scraped version histories, the informativeness of initial EC text, the fidelity of LLM denoising, and an unspecified pretraining/test split. The first and last of these are the most fragile: the first is acknowledged as a limitation but is load-bearing for the gold test set, and the last is unstated and could invalidate the CAMLM comparison if false.

free parameters (2)

CAMLM masking probabilities pspan=0.8, plow=0.05 = pspan=0.8, plow=0.05
Chosen by hand (Appendix A) with no sensitivity analysis; the reported CAMLM gains depend on this imbalance between unstable-span and background masking.
LLM generation temperature = 0.1
Set to 0.1 to make Qwen2.5-7B outputs deterministic; no ablation on temperature.

axioms (5)

domain assumption ClinicalTrials.gov label-0 (no EC amendment) is correct.
Algorithm 1 returns 0 immediately whenever ctgov_label=0; human annotators and LLM denoisers never verify non-amended trials.
domain assumption Scraped Record History versions reconstruct true EC amendment chronology.
Data collection relies on the availability and completeness of ClinicalTrials.gov Record History pages (Section 3.1).
domain assumption Initial EC text and trial metadata contain predictive signal for future amendment.
The supervised task assumes the input features are informative; experiment results support only weak signal.
domain assumption LLM-generated labels are faithful enough to train on.
AMEND_LLM training/validation labels come from Qwen2.5-7B prompts; validation against human labels is partly circular because both follow Algorithm 1.
domain assumption CAMLM pretraining does not leak test-trial future information.
The paper never states whether the paired EC version histories used for CAMLM pretraining include the test trials; if included, the model sees the future versions of exactly the trials it later classifies.

pith-pipeline@v1.3.0-alltime-deepseek · 11349 in / 13163 out tokens · 133250 ms · 2026-08-03T11:26:35.729731+00:00 · methodology

0 comments

Cite this review

Pith. "Pith review of $\texttt{AMEND++}$: Benchmarking Eligibility Criteria Amendments in Clinical Trials." pith.science (2026). https://pith.science/paper/GIQ5MOJC

@misc{pith2026260106300,
  author       = {Pith},
  title        = {Pith review of: $\textttAMEND++$: Benchmarking Eligibility Criteria Amendments in Clinical Trials},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GIQ5MOJC}},
  note         = {Machine review of arXiv:2601.06300}
}

Share X Bluesky LinkedIn Reddit HN

read the original abstract

Clinical trial amendments frequently introduce delays, increased costs, and administrative burden, with eligibility criteria being the most commonly amended component. We introduce \textit{eligibility criteria amendment prediction}, a novel NLP task that aims to forecast whether the eligibility criteria of an initial trial protocol will undergo future amendments. To support this task, we release $\texttt{AMEND++}$, a benchmark suite comprising two datasets: $\texttt{AMEND}$, which captures eligibility-criteria version histories and amendment labels from public clinical trials, and $\verb|AMEND_LLM|$, a refined subset curated using an LLM-based denoising pipeline to isolate substantive changes. We further propose $\textit{Change-Aware Masked Language Modeling}$ (CAMLM), a revision-aware pretraining strategy that leverages historical edits to learn amendment-sensitive representations. Experiments across diverse baselines show that CAMLM consistently improves amendment prediction, enabling more robust and cost-effective clinical trial design.

Figures

Figures reproduced from arXiv: 2601.06300 by Jacob Aptekar, Jimeng Sun, Mandis Beigi, Trisha Das.

**Figure 2.** Figure 2: Clinical trial eligibility criteria amend [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Figure 3: Clinical trial eligibility criteria amend [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

**Figure 4.** Figure 4: Prompt given to the LLM for the “Criteria [PITH_FULL_IMAGE:figures/full_fig_p010_4.png] view at source ↗

**Figure 5.** Figure 5: Prompt given to the LLM for the “Criteria Modified” task. [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗

**Figure 6.** Figure 6: Prompt given to the LLM for the “Criteria Added” task. [PITH_FULL_IMAGE:figures/full_fig_p012_6.png] view at source ↗

discussion (0)

Reference graph

Works this paper leans on

19 extracted references · 1 canonical work pages

[1]

Emily Botto, Zachary Smith, and Kenneth Getz. 2024. New benchmarks on protocol amendment experience in oncology clinical trials. Therapeutic Innovation & Regulatory Science, 58(4):645--654

2024
[2]

Jintai Chen, Yaojun Hu, Mingchen Cai, Yingzhou Lu, Yue Wang, Xu Cao, Miao Lin, Hongxia Xu, Jian Wu, Cao Xiao, and 1 others. 2024. Trialbench: Multi-modal artificial intelligence-ready clinical trial datasets. arXiv preprint arXiv:2407.00631

Pith/arXiv arXiv 2024
[3]

Ton J Cleophas, Aeilko H Zwinderman, Toine F Cleophas, and Eugene P Cleophas. 2009. Statistics applied to clinical trials. Springer

2009
[4]

Trisha Das, Afrah Shafquat, Mandis Beigi, Jacob Aptekar, Jason Mezey, and Jimeng Sun. 2025 a . Seqtrial: Utility preserving sequential clinical trial data generator. In AMIA Annual Symposium Proceedings, volume 2024, page 329

2025
[5]

Trisha Das, Afrah Shafquat, Mandis Beigi, Jacob Aptekar, and Jimeng Sun. 2025 b . https://doi.org/10.18653/v1/2025.acl-long.264 secret : Semi-supervised clinical trial document similarity search . In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 5278--5291, Vienna, Austria. Associati...

work page doi:10.18653/v1/2025.acl-long.264 2025
[6]

Trisha Das, Zifeng Wang, and Jimeng Sun. 2023. Twin: Personalized clinical trial digital twin generation. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 402--413

2023
[7]

Python Software Foundation. 2025. The `difflib` module — `difflib` — python 3.12.0 documentation. https://docs.python.org/3/library/difflib.html. Accessed: 2025-11-25

2025
[8]

Tianfan Fu, Kexin Huang, Cao Xiao, Lucas M Glass, and Jimeng Sun. 2022. Hint: Hierarchical interaction network for clinical-trial-outcome predictions. Patterns, 3(4)

2022
[9]

Chufan Gao, Jathurshan Pradeepkumar, Trisha Das, Shivashankar Thati, and Jimeng Sun. 2025. https://arxiv.org/abs/2406.10292 Automatically labeling clinical trial outcomes: A large-scale benchmark for drug development . Preprint, arXiv:2406.10292

Pith/arXiv arXiv 2025
[10]

Kenneth Getz, Zachary Smith, Emily Botto, Elisabeth Murphy, and Arnaud Dauchy. 2024. New benchmarks on protocol amendment practices, trends and their impact on clinical trial performance. Therapeutic Innovation & Regulatory Science, 58(3):539--548

2024
[11]

Kenneth A Getz, Stella Stergiopoulos, Mary Short, Leon Surgeon, Randy Krauss, Sybrand Pretorius, Julian Desmond, and Derek Dunn. 2016. The impact of protocol amendments on clinical trial performance and cost. Therapeutic innovation & regulatory science, 50(4):436--441

2016
[12]

Shivam Joshi. 2023. Common clinical trial amendments, why they are submitted and how they can be avoided: a mixed methods study on nhs uk sponsored research (amendments assemble). Trials, 24(1):10

2023
[13]

John A. Lewis. 1999. https://www.ncbi.nlm.nih.gov/pubmed/10440877 Statistical principles for clinical trials (ich e9): an introductory note on an international guideline. Statistics in medicine, 18 15:1903--42

arXiv 1999
[14]

o sch and Markus Neuh \

Christian L \"o sch and Markus Neuh \"a user. 2008. The statistical analysis of a clinical trial when a protocol amendment changed the inclusion criteria. BMC medical research methodology, 8(1):16

2008
[15]

Junyu Luo, Cheng Qian, Lucas Glass, and Fenglong Ma. 2024. Clinical trial retrieval via multi-grained similarity learning. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 2950--2954

2024
[16]

Yitong Tseo, MI Salkola, Ahmed Mohamed, Anuj Kumar, and Freddy Abnousi. 2020. Information extraction of clinical trial eligibility criteria. arXiv preprint arXiv:2006.07296

Pith/arXiv arXiv 2020
[17]

Chi Yuan, Patrick B Ryan, Casey Ta, Yixuan Guo, Ziran Li, Jill Hardin, Rupa Makadia, Peng Jin, Ning Shang, Tian Kang, and 1 others. 2019. Criteria2query: a natural language interface to clinical databases for cohort definition. Journal of the American Medical Informatics Association, 26(4):294--305

2019
[18]

online" 'onlinestring :=

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRING...
[19]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

[1] [1]

Emily Botto, Zachary Smith, and Kenneth Getz. 2024. New benchmarks on protocol amendment experience in oncology clinical trials. Therapeutic Innovation & Regulatory Science, 58(4):645--654

2024

[2] [2]

Jintai Chen, Yaojun Hu, Mingchen Cai, Yingzhou Lu, Yue Wang, Xu Cao, Miao Lin, Hongxia Xu, Jian Wu, Cao Xiao, and 1 others. 2024. Trialbench: Multi-modal artificial intelligence-ready clinical trial datasets. arXiv preprint arXiv:2407.00631

Pith/arXiv arXiv 2024

[3] [3]

Ton J Cleophas, Aeilko H Zwinderman, Toine F Cleophas, and Eugene P Cleophas. 2009. Statistics applied to clinical trials. Springer

2009

[4] [4]

Trisha Das, Afrah Shafquat, Mandis Beigi, Jacob Aptekar, Jason Mezey, and Jimeng Sun. 2025 a . Seqtrial: Utility preserving sequential clinical trial data generator. In AMIA Annual Symposium Proceedings, volume 2024, page 329

2025

[5] [5]

Trisha Das, Afrah Shafquat, Mandis Beigi, Jacob Aptekar, and Jimeng Sun. 2025 b . https://doi.org/10.18653/v1/2025.acl-long.264 secret : Semi-supervised clinical trial document similarity search . In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 5278--5291, Vienna, Austria. Associati...

work page doi:10.18653/v1/2025.acl-long.264 2025

[6] [6]

Trisha Das, Zifeng Wang, and Jimeng Sun. 2023. Twin: Personalized clinical trial digital twin generation. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 402--413

2023

[7] [7]

Python Software Foundation. 2025. The `difflib` module — `difflib` — python 3.12.0 documentation. https://docs.python.org/3/library/difflib.html. Accessed: 2025-11-25

2025

[8] [8]

Tianfan Fu, Kexin Huang, Cao Xiao, Lucas M Glass, and Jimeng Sun. 2022. Hint: Hierarchical interaction network for clinical-trial-outcome predictions. Patterns, 3(4)

2022

[9] [9]

Chufan Gao, Jathurshan Pradeepkumar, Trisha Das, Shivashankar Thati, and Jimeng Sun. 2025. https://arxiv.org/abs/2406.10292 Automatically labeling clinical trial outcomes: A large-scale benchmark for drug development . Preprint, arXiv:2406.10292

Pith/arXiv arXiv 2025

[10] [10]

Kenneth Getz, Zachary Smith, Emily Botto, Elisabeth Murphy, and Arnaud Dauchy. 2024. New benchmarks on protocol amendment practices, trends and their impact on clinical trial performance. Therapeutic Innovation & Regulatory Science, 58(3):539--548

2024

[11] [11]

Kenneth A Getz, Stella Stergiopoulos, Mary Short, Leon Surgeon, Randy Krauss, Sybrand Pretorius, Julian Desmond, and Derek Dunn. 2016. The impact of protocol amendments on clinical trial performance and cost. Therapeutic innovation & regulatory science, 50(4):436--441

2016

[12] [12]

Shivam Joshi. 2023. Common clinical trial amendments, why they are submitted and how they can be avoided: a mixed methods study on nhs uk sponsored research (amendments assemble). Trials, 24(1):10

2023

[13] [13]

John A. Lewis. 1999. https://www.ncbi.nlm.nih.gov/pubmed/10440877 Statistical principles for clinical trials (ich e9): an introductory note on an international guideline. Statistics in medicine, 18 15:1903--42

arXiv 1999

[14] [14]

o sch and Markus Neuh \

Christian L \"o sch and Markus Neuh \"a user. 2008. The statistical analysis of a clinical trial when a protocol amendment changed the inclusion criteria. BMC medical research methodology, 8(1):16

2008

[15] [15]

Junyu Luo, Cheng Qian, Lucas Glass, and Fenglong Ma. 2024. Clinical trial retrieval via multi-grained similarity learning. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 2950--2954

2024

[16] [16]

Yitong Tseo, MI Salkola, Ahmed Mohamed, Anuj Kumar, and Freddy Abnousi. 2020. Information extraction of clinical trial eligibility criteria. arXiv preprint arXiv:2006.07296

Pith/arXiv arXiv 2020

[17] [17]

Chi Yuan, Patrick B Ryan, Casey Ta, Yixuan Guo, Ziran Li, Jill Hardin, Rupa Makadia, Peng Jin, Ning Shang, Tian Kang, and 1 others. 2019. Criteria2query: a natural language interface to clinical databases for cohort definition. Journal of the American Medical Informatics Association, 26(4):294--305

2019

[18] [18]

online" 'onlinestring :=

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRING...

[19] [19]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...