Title and Abstract Screening and Evaluation in Systematic Reviews (TASER): a Pilot Randomised Controlled Trial of Title and Abstract Screening by Medical Students
Overview
Authors
Affiliations
Background: The production of high quality systematic reviews requires rigorous methods that are time-consuming and resource intensive. Citation screening is a key step in the systematic review process. An opportunity to improve the efficiency of systematic review production involves the use of non-expert groups and new technologies for citation screening. We performed a pilot study of citation screening by medical students using four screening methods and compared students' performance to experienced review authors.
Methods: The aims of this pilot randomised controlled trial were to provide preliminary data on the accuracy of title and abstract screening by medical students, and on the effect of screening modality on screening accuracy and efficiency. Medical students were randomly allocated to title and abstract screening using one of the four modalities and required to screen 650 citations from a single systematic review update. The four screening modalities were a reference management software program (EndNote), Paper, a web-based systematic review workflow platform (ReGroup) and a mobile screening application (Screen2Go). Screening sensitivity and specificity were analysed in a complete case analysis using a chi-squared test and Kruskal-Wallis rank sum test according to screening modality and compared to a final set of included citations selected by expert review authors.
Results: Sensitivity of medical students' screening decisions ranged from 46.7% to 66.7%, with students using the web-based platform performing significantly better than the paper-based group. Specificity ranged from 93.2% to 97.4% with the lowest specificity seen with the web-based platform. There was no significant difference in performance between the other three modalities.
Conclusions: Medical students are a feasible population to engage in citation screening. Future studies should investigate the effect of incentive systems, training and support and analytical methods on screening performance.
Systematic Review Registration: Cochrane Database CD001048.
Affengruber L, van der Maten M, Spiero I, Nussbaumer-Streit B, Mahmic-Kaknjo M, Ellen M BMC Med Res Methodol. 2024; 24(1):210.
PMID: 39294580 PMC: 11409535. DOI: 10.1186/s12874-024-02320-4.
Prioritization and sequential exclusion of articles in systematic reviews.
Saif-Ur-Rahman K, Hasan M, Hossain S, Anwar I, Hirakawa Y, Yatsuya H Campbell Syst Rev. 2023; 18(2):e1229.
PMID: 36911354 PMC: 8958185. DOI: 10.1002/cl2.1229.
Gopaul U, Charalambous M, Thilarajah S, Kwah L, Chapman S, Bayley M Syst Rev. 2022; 11(1):275.
PMID: 36529833 PMC: 9761956. DOI: 10.1186/s13643-022-02147-4.
OHearn K, Macdonald C, Tsampalieros A, Kadota L, Sandarage R, Jayawarden S BMC Med Res Methodol. 2021; 21(1):142.
PMID: 34238247 PMC: 8264476. DOI: 10.1186/s12874-021-01335-5.
Text mining to support abstract screening for knowledge syntheses: a semi-automated workflow.
Pham B, Jovanovic J, Bagheri E, Antony J, Ashoor H, Nguyen T Syst Rev. 2021; 10(1):156.
PMID: 34039433 PMC: 8152711. DOI: 10.1186/s13643-021-01700-x.