Sysrev

Connect Sysrev directly to PubMed and import documents using a targeted boolean search string.

Set up inclusion and exclusion boolean labels based on the research protocol (e.g., 'Human Subjects', 'Randomized Control Trial').

Assign a team of reviewers to independently read and label abstracts, utilizing Sysrev's double-blind screening.

Use the active learning model's predictions to automatically discard highly irrelevant papers, focusing human effort on borderline or highly relevant abstracts.

Export the final list of included articles as a PRISMA-compliant CSV.

Upload raw, unstructured text files or domain-specific PDFs via the Sysrev API.

Create custom annotation labels tailored to the entities needed for the ML model (e.g., 'Drug Dosage', 'Adverse Reaction').

Distribute annotation tasks to subject matter experts, utilizing concordance tracking to ensure high-quality, agreed-upon labels.

Fetch the labeled data via GraphQL API, directly feeding the structured JSON output into the model training pipeline.

Import target trial IDs directly from the ClinicalTrials.gov integration.

Design a project schema with categorical and string extraction labels targeting 'Phase', 'Patient Count', and 'Primary Endpoint'.

Have reviewers highlight the exact text in the trial documents that correspond to the data points for verifiable extraction.

Administrators resolve any conflicting extractions through the resolution dashboard.

Export a clean, structured dataset comparing all trials.

Upload bulk PDF reports into a secure, private Sysrev project.

Define strict extraction parameters and guidelines within the project description.

Reviewers navigate the PDFs within Sysrev's viewer, pinning annotations to specific chemical names and exposure limits.

Leverage Sysrev's ML to cluster similar documents, grouping unreviewed PDFs containing similar chemical terminologies.

Generate a standardized JSON report of all extracted chemical hazards for regulatory filing.

Establish a primary systematic review project with an active, automated PubMed integration.

Configure the integration to periodically fetch new publications matching the original search criteria.

As new documents arrive, Sysrev's pre-trained project ML model automatically scores and assigns a probability of inclusion.

Reviewers receive notifications for new high-probability documents and quickly verify the ML's prediction.

The project's exported dataset dynamically updates, maintaining a 'living' state of current research.