Title/Abstract Screening Module:
Leveraging GPT-4o mini’s deep reasoning capabilities, this module processes thousands of titles and abstracts in minutes, flagging and filtering studies with industry-leading precision. The module has undergone dozens of revisions and testing procedures using previously completed published systematic reviews and is currently under publication in a peer-reviewed journal.
Full-Text Screening Module:
Powered by Claude 3.7 Sonnet, our system dives into full text articles evaluating different sections of the manuscripts for eligibility as per PICO criteria at hand to ensure only the most pertinent papers advance to the next stage. The module has undergone dozens of revisions and testing procedures using previously completed published systematic reviews and is currently under publication in a peer-reviewed journal.
Data Extraction Module:
Utilizing Claude Sonnet 3.7, this component automatically captures all required study details needed for data synthesis including but not limited to study design, interventions, participant demographics, as well as efficacy and safety outcome characteristics from text, tables, and figures ensuring comprehensive, error-free data collection. The module has undergone dozens of revisions and testing procedures using previously completed published systematic reviews and is currently under publication in a peer-reviewed journal.
Risk of Bias Assessor Module:
Currently under development, the module is powered by Claude Sonnet 3.7 and evaluates risk of bias in studies based on the Cochrane risk of bias for randomized trials, as well as Newcastle Ottawa Scale for observational studies.