TABLE OF CONTENTS

• Focus mode for data extraction-overview
• Workflow: Extract  and Connect
• Components of a Single Data Extraction Field
• Using AI Suggestions in Laser 

Focus mode for data extraction-overview

To enter data extraction focus mode you need to double-click on the selected record on the reference list or click the [Focus mode] icon 

The Focus mode view includes two main sections:

• Data extraction form: By default, you'll see an overview with all the tabs. To start data extraction, simply click on the tab of your choice.  Alternatively you can use the option find field. 

• PDF view (central part of the screen)

You can access supplementary materials if they have been attached by your librarian. You can also add a new attachment if necessary.Use the 'Next PDF' button to switch to those files.

To switch to the following study and postpone data extraction for this study, use the arrows in the upper side of focus mode.

If you need to search within the document, use the search button.

To export PDF, click on the export button.

In the left-side navigation bar, you will also find PDF navigation.

Workflow: Extract  and Connect

There are two possible actions when extracting data from studies:

• Extracting relevant values from PDF into extraction fields.

• Connecting the extracted data with other extracted values.

Let's examine an example of data extraction for outcomes. We aim to extract the number of patients with event for two intervention arms. Once the data extraction is complete, we want to connect these extracted values with other data, such as relevant subgroups. The panel for connection is located below each data extraction field (note that panel for data connection will be available if specified in the data extraction form.

To connect data:

• Click on the connection field (e.g., analysis population).

• Select the relevant value from the list (e.g., entire population).
   
  
  

In some cases, the connection will be done automatically (auto-connection). In the example below, Laser AI automatically connects all data related to study details (such as country, etc.). Your task is to extract individual fields, like the country where studies were conducted, and Laser will connect the country to other concepts that define study characteristics.

Components of a Single Data Extraction Field

Single data extraction field include 3 type of data:

• 'Extracted Value': This refers to a value provided by the extractor, which may be a vocabulary term.

• 'Author-Reported Value': Related value suggested by the AI model or selected text from a PDF file, also suggested by the model or added manually.

• ‘Comment’: your comment to extracted value.

Tip: The information on the types of data available for a single data extraction field can be found in the button near the field name. An arrow in the right upper corner indicates that a comment is available, while an arrow in the left down corner means that the author-reported value is available.

Tip: Detailed instructions (if provided by the project manager) on how to extract this specific type of data will be shown when you click on the 'information button' in the right corner of the field

Using AI Suggestions in Laser

In Laser AI, you'll find different models (for each data extraction field) that support extraction by providing suggestions directly within the field or by supporting table extraction. You should accept or reject these suggestions by clicking the "Accept" or "Reject" button next to the field. In some cases, you can use the batch acceptance button located near each section name.

To change the order of fields within each tab, use the dot menu.

RELATED ARTICLES 

1. How to connect extracted data 
2. How to extract data in a text field 
3. How to extract data in a vocabulary field 
4. How to extract data in nested tables with/without AI support 
5. How to extract data with the AI model suggestions? 
6. How to create data extraction form