How to set up and Extract Data in EasySLR?

How to set up and Extract Data in EasySLR?

Overview
Data extraction in EasySLR is designed to be flexible and easy to set up. You can configure your data extraction framework either:
  • During project setup, or
  • After completing Title–Abstract (TiAb) and Full-Text screening

EasySLR provides predefined templates to help you get started quickly. These templates can be fully customised to match your project requirements.
To access and customise templates:
  1. Click on “Add New Sheet”
  2. Select a suitable template
  3. Modify



Creating Custom Data Extraction Fields

EasySLR allows you to create a structured extraction framework using three types of fields:
  • Static Fields
  • Dimensions
  • Measure Groups

Field Creation Guidelines
When creating a field:
  • Provide a clear field name
  • Add a brief description specifying what the AI should extract
  • Ensure the field name:
    • Begins with a letter or underscore (_)
    • Uses only allowed special characters at the end:
             _ ( ) % [ ] . ?
This ensures consistency and improves AI extraction accuracy.

Understanding Field Types

1. Static Fields
Static fields capture information that remains constant throughout the study.
Examples:
  • Authors
  • Title
  • Country
  • Sponsor
  • Study Design
Key Characteristics:
  • One value per study
  • Does not vary across groups, timepoints, or sections

2. Dimensions
Dimensions act as reference points or categories that help organize and segment the extracted data. They define where or how the data is being reported, and each dimension can have multiple values.

Examples:
  • Treatment Arms (e.g., Drug A, Placebo)
  • Subgroups (e.g., Male, Female, Age > 60)
  • Timepoints (e.g., Baseline, Week 12, Month 6)
  • Interventions
Key Characteristics:
  • Can have multiple values
  • Help “slice” the data for comparison

3. Measure Groups
Measure groups capture the actual quantitative or descriptive outcomes reported in the study, often in relation to dimensions.
Example:
  • Age
    • Mean
    • Median
    • Range
Key Characteristics:
  • Typically linked to dimensions
  • Can include multiple related metrics

How They Work Together
  • Static Fields → Provide overall study-level information
  • Dimensions → Define how the data is grouped or segmented
  • Measure Groups → Capture the actual values being reported
This structure ensures your data extraction aligns with how studies present data.



Structuring Your Extraction Form
Reorder Fields
Use the Reorder function to adjust the sequence of fields
Helps improve organisation and usability

Full Screen Mode
Use Full Screen mode for a smoother setup experience.

Multiple Sheets
  • Add multiple sheets for different outcomes
  • Name each sheet based on the data it captures

You can also define the type of Static Data Extraction field based on the nature of the data being captured. Supported field types include:

Text – for free-text inputs
Single Select – for choosing one option from a predefined list
Multi-Select – for selecting multiple applicable options


Refer to the video for a set by step guide on Data Extraction setup: https://youtu.be/lINcXXSgw6Q
You also have the option to upload your own data extraction template. For guidance, we provide a sample template as a reference to help you create your own. When you select Upload Excel, you’ll be prompted to either Add to existing sheets or Replace all existing sheets.


If you choose Replace all existing sheets, all current fields will be overwritten, and any previously extracted data will be lost.

We have also added (14) QA templates to help teams easily conduct critical appraisal of the included studies. You can access predefined QA templates and customise them further to fit your review needs. These templates are built directly into the extraction workflow, ensuring reviewers complete QA consistently.

After updating the data extraction fields, activate AI functionality to enable automated extraction of all available information.

To validate AI extraction, go to the “Data Extraction” section under Review stages.

Data Extraction is divided into three sections to categorise studies based on their extraction stage. Initially, all articles will be categorised as "Articles to Extract," signifying that the extraction process has not begun.

Click “Start Extracting” to:
  • View AI-extracted data
  • Validate accuracy and completeness
  • Access source references directly from the PDF
You'll find the extracted data alongside each field in the right-hand side pane. This panel also provides links to the specific sources (one or multiple) within the PDF where the AI sourced that information. This feature facilitates quicker quality control (QC) and validation, allowing you to easily verify the accuracy of the extracted data.


Additionally, you can edit the extracted data, to make necessary changes and updates as required. The linked sources associated with the extracted data are also editable allowing you to add or delete sources as needed. 

To add sources, click on the 'Add Source' button. Select the relevant data from the PDF, click '+', then 'Done'. The new source will be linked.

You can also apply filters on dimensions to view specific information or directly add precise values in the table.

Use the “Add Excel Row” button to add new values to dimensions when additional data is identified during extraction.

Once you've completed the extraction and made necessary updates, select ‘Extraction done’ available at the bottom of the screen and click ‘Save’. This action moves the article to the "Extraction Done" section, where it will be available for another reviewer to begin the Quality Control (QC) process.

Note: Using the "Save & Next" button allows reviewers to proceed to the next article while keeping the current article in its respective extraction stage.

In the “Extraction Done” section, the reviewer will have access to the finalised extracted data, along with detailed information about the individuals who contributed to the data extraction at various stages. This ensures transparency and accountability for each step of the process.

Once the QC process is completed, the reviewer selects "QC Done" and clicks "Save." This action moves the article to the "QC Done" section, indicating that the work on that study is complete.


If rework is needed, the reviewer can flag the relevant fields with a comment, then set the study status to 'Pending' and click 'Save' to send it back for rework.

Once the extraction process is complete, you can easily export all the data into an Excel file for further analysis and reporting.
 
The downloaded Excel file will include all comments added by reviewers for reference and record-keeping and the data can be directly used for Meta Analysis.



    • Related Articles

    • Direct PDF Upload to Data Extraction

      EasySLR allows you to start Data Extraction directly by uploading PDFs, without first completing Title–Abstract or Full-Text screening. This feature is designed for teams that already have finalised studies or want to begin extraction immediately, ...

    • Mini Data Extraction in EasySLR

      The Mini Data Extraction feature in EasySLR allows teams to extract a focused set of predefined fields during the screening process. It's ideal when only high-level or limited data is required before full-scale data extraction begins. How to Set Up ...

    • How do you identify duplicates? Can I get a list of the duplicates identified?

      Duplicate detection and resolution are essential for ensuring a clean and accurate dataset in systematic reviews. EasySLR provides a robust de-duplication feature that helps identify, review, and manage duplicate records efficiently. 1. Setting Up ...

    • Guide to tracking AI and data processing tasks in EasySLR

      Tasks triggered by user actions, such as AI processing or Data Extraction runs, continue to operate in the background. Depending on the size of data, these tasks may take some time to complete. Tracking the progress of these tasks gives you full ...

    • How to set up a Protocol in easySLR?

      Setting up a protocol in EasySLR is simple and user-friendly. A pre-set template is available to guide you, so you don’t have to create everything from scratch. Your protocol defines the study criteria, guiding both human reviewers and AI in ...