Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 22 of the 22 references it cites were found in a registry or an index.

RobotReviewer: evaluation of a system for automatically assessing bias in clinical trials

Open access
Marshall, Iain J, Kuiper, Joël, Wallace, Byron C · 2016 · Journal of the American Medical Informatics Association

Abstract

Objective To develop and evaluate RobotReviewer, a machine learning (ML) system that automatically assesses bias in clinical trials. From a (PDF-formatted) trial report, the system should determine risks of bias for the domains defined by the Cochrane Risk of Bias (RoB) tool, and extract supporting text for these judgments. Methods We algorithmically annotated 12,808 trial PDFs using data from the Cochrane Database of Systematic Reviews (CDSR). Trials were labeled as being at low or high/unclear risk of bias for each domain, and sentences were labeled as being informative or not. This dataset was used to train a multi-task ML model. We estimated the accuracy of ML judgments versus humans by comparing trials with two or more independent RoB assessments in the CDSR. Twenty blinded experienced reviewers rated the relevance of supporting text, comparing ML output with equivalent (human-extracted) text from the CDSR. Results By retrieving the top 3 candidate sentences per document (top3 recall), the best ML text was rated more relevant than text from the CDSR, but not significantly (60.4% ML text rated ‘highly relevant' v 56.5% of text from reviews; difference +3.9%, [−3.2% to +10.9%]). Model RoB judgments were less accurate than those from published reviews, though the difference was <10% (overall accuracy 71.0% with ML v 78.3% with CDSR). Conclusion Risk of bias assessment may be automated with reasonable accuracy. Automatically identified text supporting bias assessment is of equal quality to the manually identified text in the CDSR. This technology could substantially reduce reviewer workload and expedite evidence syntheses.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1093/jamia/ocv044

Cites

22 of 22 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1:. Algorithmic annotation of clinical trial PDFs using data from the CDSR.
Figure 1:. Algorithmic annotation of clinical trial PDFs using data from the CDSR.CC BY-NC · PubMed Central
Figure 2: Schematic of corpus construction, and outline of the distant supervision process.
Figure 2: Schematic of corpus construction, and outline of the distant supervision process.CC BY-NC · PubMed Central
Figure 3: A schematic depiction of our multi-task learning approach. We define a joint classification model across domains. To achieve this, we include k representations of each instance (e.g., document) d in the design matrix, one per risk of bias domain. We construct the target vector y with the corresponding per-domain labels for d (derived from the CDSR). Each of the k entries for d comprises k+1 concatenated vectors of length equal to the vocabulary size (V), where for the document model this vocabulary is the union of unique uni- and bi-grams appearing in at least two articles, and for the sentence model V is the union of uni- and bi-grams appearing in at least two supporting sentences. The first copy of V in each row is shared across all entries representing the corresponding article; the remaining are domain specific copies. Thus in any given row, all but two of these sub-vectors will be zero vectors. Specifically, each row i will contain two non-zero vectors that are copies of the bag-of-words representation (binary indicators for V) for instance i (x i ): one shared and the other domain specific. The shared component allows borrowing of strength across domains, while the domain-specific components enable the model to learn words that signal low risk of bias (or the likelihood of supporting RoB assessment, for the sentence prediction task) only in specific domains.
Figure 3: A schematic depiction of our multi-task learning approach. We define a joint classification model across domains. To achieve this, we include k representations of each instance (e.g., document) d in the design matrix, one per risk of bias domain. We construct the target vector y with the corresponding per-domain labels for d (derived from the CDSR). Each of the k entries for d comprises k+1 concatenated vectors of length equal to the vocabulary size (V), where for the document model this vocabulary is the union of unique uni- and bi-grams appearing in at least two articles, and for the sentence model V is the union of uni- and bi-grams appearing in at least two supporting sentences. The first copy of V in each row is shared across all entries representing the corresponding article; the remaining are domain specific copies. Thus in any given row, all but two of these sub-vectors will be zero vectors. Specifically, each row i will contain two non-zero vectors that are copies of the bag-of-words representation (binary indicators for V) for instance i (x i ): one shared and the other domain specific. The shared component allows borrowing of strength across domains, while the domain-specific components enable the model to learn words that signal low risk of bias (or the likelihood of supporting RoB assessment, for the sentence prediction task) only in specific domains.CC BY-NC · PubMed Central
Figure 4: Test/training set partition for the document-level judgments, and use of the second Risk of Bias assessment as a surrogate for human performance.
Figure 4: Test/training set partition for the document-level judgments, and use of the second Risk of Bias assessment as a surrogate for human performance.CC BY-NC · PubMed Central
Figure 5. Example of our prototype system showing the bias assessment for random sequence generation and a supporting sentence.
Figure 5. Example of our prototype system showing the bias assessment for random sequence generation and a supporting sentence.CC BY-NC · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

10 more claims about this paper

Create a free account to see every claim other papers make about this one.

See all claims — free account

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 22 references

Cited by

See all 320

Metadata

Authors
Marshall, I. J., Kuiper, J., & Wallace, B. C.
Affiliations
  • King's College London (GB)
  • University Medical Center Groningen (NL)
  • The University of Texas at Austin (US)
Corresponding author
Marshall, Iain J
Venue
Journal of the American Medical Informatics Association
Published
2015-06-22
Volume
23
Issue
1
Pages
193-201
ISSN
  • 1067-5027 (ISSN-L)
  • 1527-974X
OpenAlex
W2184378182
Cited by
317
References
22
Citations by year
2026: 34 · 2025: 39 · 2024: 39 · 2023: 32 · 2022: 23 · 2021: 41 · 2020: 49 · 2019: 13 · 2018: 23 · 2017: 16 · 2016: 5 · 2015: 2
FWCI
13.97
Type
journal article
Language
en
Open access
Locations
  • Journal of the American Medical Informatics Association (publishedVersion, cc-by-nc)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by-nc)
  • PubMed Central (publishedVersion, cc-by-nc)
  • PubMed (publishedVersion)
  • University of Groningen research database (University of Groningen / Centre for Information Technology) (publishedVersion, cc-by-nc)
  • Research Portal (King's College London) (submittedVersion, cc-by-nc)
  • HighWire Press Open Archive (submittedVersion)
  • University of Groningen research database (University of Groningen / Centre for Information Technology) (submittedVersion, cc-by-nc)
Funders
  • Brown University
  • King's College Hospital NHS Foundation Trust
  • National Institute for Health and Care Research
  • King's College London
  • Monash University
  • NIHR Collaboration for Leadership in Applied Health Research and Care South London
  • Medical Research Council (MR/N015185/1)
  • Institute of Education, University of London
Topics
  • Social Sciences › Decision Sciences › Statistics, Probability and Uncertainty › Meta-analysis and systematic reviews
  • Social Sciences › General Social Sciences › Computational and Text Analysis Methods
  • Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
Keywords
risk of bias assessment, clinical trials, Cochrane risk of bias tool, multi-task learning, Cochrane Database of Systematic Reviews, evidence synthesis
MeSH terms
Machine Learning, Algorithms, Clinical Trials as Topic, Natural Language Processing, Review Literature as Topic, Bias, Peer Review, Research, Databases as Topic, Data Mining

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: RobotReviewer: evaluation of a system for automatically assessing bias in clinical trials. https://app.cochat.ai/papers/10.1093/jamia/ocv044
The paper
Marshall, I. J., Kuiper, J., & Wallace, B. C. (2016). RobotReviewer: evaluation of a system for automatically assessing bias in clinical trials. Journal of the American Medical Informatics Association. https://doi.org/10.1093/jamia/ocv044

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.