| Unique ID issued by UMIN | UMIN000060250 |
|---|---|
| Receipt number | R000068879 |
| Scientific Title | The diagnostic accuracy of large language models in screening citations in orthopaedic literatures |
| Date of disclosure of the study information | 2026/01/06 |
| Last modified on | 2026/01/04 15:42:12 |
The diagnostic accuracy of large language models in screening citations in orthopaedic literatures
ORTHOSCREEN
The diagnostic accuracy of large language models in screening citations in orthopaedic literatures
ORTHOSCREEN
| Japan |
the musculoskeletal literature
| Orthopedics |
Others
NO
This study aims to elucidate the diagnostic accuracy, time expenditure, and associated costs of LLMs in screening citations for meta-analyses focused on hip fractures, in comparison to conventional manual screening methods.
Others
This study aims to elucidate time expenditure, and associated costs of LLMs in screening citations for meta-analyses focused on hip fractures, in comparison to conventional manual screening methods.
The primary endpoint of this study is the sensitivity and specificity of LLM screening for title and abstract screening as well as full-text screening.
The secondary endpoint of this study is the time and costs for LLM screening compared to manual screening.
Observational
| Not applicable |
| Not applicable |
Male and Female
Ten reviewers for five SRMAs (Two main reviewers for each SRMA) will be recruited by online meeting of Japanese orthopaedic trauma association systematic review committee.The inclusion criteria for participants are as follows: (1) being a member of the systematic review committee in JOTA; (2) availability to participate in citation screening; and (3) willingness to provide informed consent.
We will exclude committee members who cannot give informed consent and are unavailable for citation screening.
10
| 1st name | Natsumi |
| Middle name | |
| Last name | Saka |
Teikyo University, School of Medicine
Department of Orthopaedics
1738606
2-1-1 Kaga, Itabashi
0339641211
natsumi613@gmail.com
| 1st name | Natsumi |
| Middle name | |
| Last name | Saka |
Teikyo University, School of Medicine
Department of Orthopaedics
1738606
Kaga2-11-1, Itabashi
0339641211
natsumi613@gmail.com
Teikyo University
Ministry of Education, Culture, Sports, Science and Technology
Japanese Governmental office
Japan
Teikyo University Medical Research Ethics Committee
2-1-1 Kaga, Itabashi
0339647256
turb-office@teikyo-u.ac.jp
NO
東京都
| 2026 | Year | 01 | Month | 06 | Day |
Unpublished
Preinitiation
| 2025 | Year | 11 | Month | 14 | Day |
| 2025 | Year | 12 | Month | 01 | Day |
| 2026 | Year | 01 | Month | 15 | Day |
| 2027 | Year | 12 | Month | 31 | Day |
Search Strategy and Data Acquisition: Initially, search queries will be formulated for five Clinical Questions designated for new Systematic Reviews based on the Japanese Orthopedic Association Guidelines for Management of Hip Fractures. Literature will be retrieved from literature database as well as clinical trial registries.
Random Sampling: From the retrieved literature, 500 records will be randomly selected. This process will be conducted by the Principal Investigator using a dedicated software application.
Human Screening Protocol For the SR committee screening, two primary reviewers will independently assess the records using screening application to determine eligibility based on inclusion and exclusion criteria for each CQ. Discrepancies will be resolved by a third reviewer. Reviewers will remain blinded to each other's decisions and to the results of the Large Language Model screening described below.
LLM Screening and Time Tracking: Simultaneously, the LLM will be utilized to perform screening on the same dataset. Additionally, the time required for the human reviewers to complete the screening- automatically logged within screening application will be extracted for analysis.
| 2026 | Year | 01 | Month | 04 | Day |
| 2026 | Year | 01 | Month | 04 | Day |
Value
https://center6.umin.ac.jp/cgi-open-bin/ctr_e/ctr_view.cgi?recptno=R000068879