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

The DOI matches its Crossref record; 30 of the 30 references it cites were found in a registry or an index; 5 of the 13 sources its claims cite were checked for support.

Faster title and abstract screening? Evaluating Abstrackr, a semi-automated online screening program for systematic reviewers

Open access
Rathbone, John, Hoffmann, Tammy, Glasziou, Paul · 2015 · Systematic Reviews

Abstract

BACKGROUND: Citation screening is time consuming and inefficient. We sought to evaluate the performance of Abstrackr, a semi-automated online tool for predictive title and abstract screening. METHODS: Four systematic reviews (aHUS, dietary fibre, ECHO, rituximab) were used to evaluate Abstrackr... RESULTS: Of the unscreened citations, Abstrackr's prediction algorithm correctly identified all relevant citations for the rituximab and dietary fibre reviews...

Overview

CoChat Summary

## TL;DR

This study evaluated Abstrackr, a semi-automated online title-and-abstract screening tool, on four existing systematic reviews. It found substantial workload savings (9% to 80%) with very few missed relevant citations, but performance varied by review complexity and dataset balance.

## Key findings

  • Across four datasets, Abstrackr’s workload saving ranged from 9% in rituximab to 80% in the larger ECHO sensitivity analysis.
  • In the main analyses, Abstrackr identified all relevant citations for the rituximab and dietary fibre reviews, while missing one relevant citation each in the aHUS and ECHO reviews.
  • Precision varied by dataset: 16.8% (aHUS), 24.7% (dietary fibre), 29.2% (ECHO), and 45.5% (rituximab).
  • The false negative rate ranged from 2.4% to 21.7%, while the actual proportion missed was 0% to 0.21%; average workload saving across the four datasets was 41%.

## Methods

  • The authors evaluated Abstrackr using four existing systematic reviews drawn from biomedical database searches: aHUS (n = 1415), dietary fibre (n = 517), ECHO (n = 1735), and rituximab (n = 1042).
  • Titles and abstracts were screened in Abstrackr until the algorithm generated predictions; the remaining unscreened citations were then classified as relevant/irrelevant and compared against the original review decisions.
  • Performance was assessed using precision, false negative rate, proportion missed, and workload saving; post hoc sensitivity analyses were done for the aHUS dataset (including case reports) and a larger ECHO dataset (15,920 citations).

## Limitations

  • The paper states that findings may be limited by the four datasets used, and that other clinical specialties could yield different precision, workload saving, and missed-study rates.
  • The authors note that the datasets came from recently published reviews with trials mostly from 1995 onward, so the results may not generalize to older studies with less detailed titles/abstracts.
  • Evident from the text: Abstrackr was tested on existing review datasets, not prospectively within live review workflows, so real-world performance may differ.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1186/s13643-015-0067-6

Cites

30 of 30 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Percentage of citations predicted by Abstrackr that were relevant for further full text inspection. *Raw numbers of the proportion of citations selected for inspection
Fig. 1 Percentage of citations predicted by Abstrackr that were relevant for further full text inspection. *Raw numbers of the proportion of citations selected for inspectionCC BY · PubMed Central
Fig. 2 False negative rate. *Raw numbers of the proportion of citations incorrectly predicted by Abstrackr to be irrelevant for further inspection
Fig. 2 False negative rate. *Raw numbers of the proportion of citations incorrectly predicted by Abstrackr to be irrelevant for further inspectionCC BY · PubMed Central
Fig. 3 Percentage of studies missed by Abstrackr—but were included in the reviews. *Raw numbers of the proportion of citations missed (predicted not relevant)
Fig. 3 Percentage of studies missed by Abstrackr—but were included in the reviews. *Raw numbers of the proportion of citations missed (predicted not relevant)CC BY · PubMed Central
Fig. 4 Workload saving (%) when using Abstrackr in each of the four datasets. *Raw numbers of the proportion of citations predicted not relevant from the total
Fig. 4 Workload saving (%) when using Abstrackr in each of the four datasets. *Raw numbers of the proportion of citations predicted not relevant from the totalCC BY · PubMed Central

Verification

Claims the paper makes

4 Valid · 0 Not valid · 6 Not checked · 36 Own results and interpretations

22 more claims

Create a free account to see all 46 claims and to verify the 2 references still open.

See all claims — free account
  • Probably not a substantial claim. 6
  • It may state two separate claims as one. 16

Create a free account to see these claims.

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?

3 more claims about this paper

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

See all claims — free account

Citations

Cites

Show all 30 references

Research works printed but not linked: 1

  1. Frequency of updating Cochrane Reviews [ http://www.cochrane.org/editorial-and-publishing-policy-resource/cochrane-review-updates ]

Cited by

See all 184

Metadata

Authors
Rathbone, J., Hoffmann, T., & Glasziou, P.
Affiliations
Bond University (AU)
Corresponding author
Rathbone, John
Venue
Systematic Reviews
Published
2015-06-15
Volume
4
Issue
1
Pages
80-80
ISSN
2046-4053 (ISSN-L)
OpenAlex
W1760234209
Cited by
184
References
30
Citations by year
2026: 10 · 2025: 25 · 2024: 16 · 2023: 25 · 2022: 17 · 2021: 21 · 2020: 23 · 2019: 22 · 2018: 12 · 2017: 7 · 2016: 5 · 2015: 1
FWCI
1.37
Type
journal article
Language
en
Open access
Locations
  • Systematic Reviews (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Bond University Research Portal (Bond University) (submittedVersion, cc-by)
  • e-publications@bond (Bond University) (submittedVersion)
Topics
  • Social Sciences › Decision Sciences › Statistics, Probability and Uncertainty › Meta-analysis and systematic reviews
  • Physical Sciences › Computer Science › Information Systems › Information Retrieval and Search Behavior
  • Social Sciences › Library and Information Sciences › scientometrics and bibliometrics research
Keywords
title and abstract screening, semi-automated screening, citation screening, systematic reviews, predictive classification, false negative rate, research waste, aHUS, rituximab, dietary fiber, ECHO trial
MeSH terms
Abstracting and Indexing, Algorithms, Humans, Review Literature as Topic, Sensitivity and Specificity, Reproducibility of Results

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.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

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: Faster title and abstract screening? Evaluating Abstrackr, a semi-automated online screening program for systematic reviewers. https://app.cochat.ai/papers/10.1186/s13643-015-0067-6
The paper
Rathbone, J., Hoffmann, T., & Glasziou, P. (2015). Faster title and abstract screening? Evaluating Abstrackr, a semi-automated online screening program for systematic reviewers. Systematic Reviews. https://doi.org/10.1186/s13643-015-0067-6

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.