scibert_finetune_rct_consort

<!-- Provide a quick summary of what the model is/does. -->

Model Details

This is a scibert finetuned model on 50 randomized controlled trial articles annotated with fine-grained CONSORT checklist items at the sentence level.

Model Description

<!-- Provide a longer summary of what this model is. -->

This model classifies input RCT text into the following CONSORT item labels:

• LABEL_1: Any changes to trial outcomes after the trial commenced, with reasons
• LABEL_2: Completely defined pre-specified primary and secondary outcome measures, including how and when they were assessed
• LABEL_3: Description of trial design (such as parallel, factorial) including allocation ratio
• LABEL_4: Eligibility criteria for participants
• LABEL_5: How sample size was determined
• LABEL_6: If done, who was blinded after assignment to interventions (for example, participants, care providers, those assessing outcomes) and how
• LABEL_7: If relevant, description of the similarity of interventions
• LABEL_8: Important changes to methods after trial commencement (such as eligibility criteria), with reasons
• LABEL_9: Mechanism used to implement the random allocation sequence (such as sequentially numbered containers), describing any steps taken to conceal the sequence until interventions were assigned
• LABEL_10: Method used to generate the random allocation sequence
• LABEL_11: Methods for additional analyses, such as subgroup analyses and adjusted analyses
• LABEL_12: Settings and locations where the data were collected
• LABEL_13: Statistical methods used to compare groups for primary and secondary outcomes
• LABEL_14: The interventions for each group with sufficient details to allow replication, including how and when they were actually administered
• LABEL_15: Type of randomisation; details of any restriction (such as blocking and block size)
• LABEL_16: When applicable, explanation of any interim analyses and stopping guidelines
• LABEL_17: Who generated the random allocation sequence, who enrolled participants, and who assigned participants to interventions

Metrics

• accuracy: 0.731
• precision: 0.68
• recall: 0.73
• f1: 0.69