Intelligent document processing (IDP) transforms the administrative burden by cutting the average fax handling time from 3-8 minutes to just a fraction of that. It will automatically classify documents, extract patient data, and send them straight to your EHR with little manual work. IDP actions will be available within Workspaces.
This article will teach you how to use our IDP tool.
Table of contents:
IDP Glossary
Workspaces and IDP Overview
Workspace Settings
Processing a Document
Document Processing via API
We currently integrate with the following EHRs:
OpenEMR
NextGen® Software NextGen Enterprise
ModMed
PointClickCare
IDP Glossary
Active Revision: The specific revision model that is currently in use for document classification.
Classification: A machine learning engine that consumes unstructured documents and returns a document type and confidence score for that classification. Documo offers two methods for using a model: a Standard Model of common medical documents, and custom classification where the user can train a model on a set of documents they provide.
Confidence Threshold: The minimum confidence level, expressed as a percentage, that the classification tool must reach when classifying a document type. This threshold is manually configurable in the workspace settings, allowing you to define the level of certainty required before a classification is accepted.
Data Source: Origin of where an incoming document comes from. A workspace can have a variety of data sources that route documents into a specific workspace. These are configured to each workspace independently. Example: a Data Source is a fax number. When an admin configures 281-111-1222 to their workspace, all inbound faxes to that number will be added to that workspace.
NOTE: A fax number can only be routed to ONE workspace at a time, but a workspace can have multiple numbers assigned to it.
Document Class: The label that our classification engine assigns to a document based on the training model. Example document types: referral, prior authorization, medical records, etc.
Document Field: Document fields are the term for key-value pairs. Our engine takes specified field profiles as input and produce extracted field data as output.
Example: Patient First Name is the field profile, John is the field data.
Document Processor: An advanced document viewer where users can review documents, data & metadata, and conduct ‘processing’ workflows such as adding a document to an patient’s chart.
Document Retention: The length of time a document remains stored in your workspace before it is automatically deleted.
Document Type: Document classes and manually defined types of documents that can be used to automatically or manually set the document type in the processor.
EHR Integration: Provides API connectivity to an EHR to support patient lookup, encounter lookup, uploading files, and a variety of other connection endpoints through Documo’s portal.
Engine: A system of machine learning and AI technologies that processes documents and returns results back to the workspaces component of the Documo application.
Extraction: Information automatically identified and captured from a document, then populated into designated data fields.
Optical Character Recognition (OCR): Technology that converts images or scanned documents containing text into machine-readable text format, generating a searchable PDF with the captured information.
Revision: A trained AI model designed to classify and categorize document types based on the provided document types.
Source Fax: Copies of faxes will be converted into a PDF and added to the workspace. The original fax (source fax) will still be available in the user's fax history.
Standard Model: Documo’s out-of-the-box model for document classification. Our standard model is trained on around 20 common healthcare document types.
Workspace: Collaborative inbox for users to process, review, and manage incoming documents from a variety of data sources.
Workspaces and IDP Overview
Workspace Settings
First, navigate to Workspaces, a dedicated hub for documents awaiting processing.
Next, click Workspace Settings in the Workspaces menu. Here, you'll set the Active Revision and your Confidence Threshold:
Document types with a lock next to them have been created by the Active Revision. Document types with a pencil next to them have been created manually by clicking Add New. The names of manually created document types can be changed by clicking the pencil:
Adjust individual workspace settings by clicking the 3 dots on the workspace:
Manage Members: Add or delete members from that specific workspace:
Edit Workspace:
Automation: Enable or disable AI automations for that specific workspace. All documents in the workspace will automatically have OCR, Classification, or Extraction run if they're enabled.
Set document retention to the amount of days you would like documents to remain in your workspace before they're automatically deleted.
Members: View members that can access that workspace:
Data sources: Select the fax number(s) on your account for any inbound faxes you would like to appear in your workspace. Note that a fax number can only be added to one workspace at a time, but you can have multiple fax numbers assigned to a workspace:
Processing a Document
Upon clicking a Workspace, all documents in that Workspace will display. Below you will find what each column means:
Date: The date that the document was uploaded or faxed into the workspace
Patient: This is updated once the patient name is selected in the document processor
Document Name:
Uploads: The name of the original document
Inbound faxes: inbound_date and time of fax record_messageID of the fax
These can be changed in the processor window and the new name will display in the workspace table
Type: Classified document type
Source: Will display Upload for manually uploaded faxes and will display the fax number if it was an inbound fax
IDP: All IDP actions on the workspace have run
Status: Status of the document
Possible statuses:
New
Opened
In Progress
Sent to EHR
Clicking the 3 dots next to a document opens up the options for that document:
If OCR has already been ran, OCR Data will be provided next to the document for easy copying of the text by clicking OCR Data:
If classification and extraction have already been ran, processing the document populates the document type and the data fields in the document processor. Document type and data fields can be added manually if classification and extraction have not been ran. Click Process with your EHR (choose the EHR your account is integrated with):
You're able to edit the document name on the top of the processor window:
Ensure that the document type is correct, and the extracted patient data is also correct, then click Find Patient. Match the patient, then click Next:
A document type can manually be changed by clicking the dropdown, and selecting a new type:
Choose the location, then click Send To (Your EHR):
Navigate to the selected location in your EHR and view the document there:
Info & History gives an audit trail of actions that users and the system took:
NOTE: If an inbound fax fails at any point during transmission, the document will NOT be uploaded to your workspace.
Document Processing via API
View the entire list of Workspaces API endpoints (still in beta).
Users are able to create and configure workspaces, classify documents, extract information from documents, and get OCR data from documents. Currently, users are unable to send a document to their EHR via API. This will have to be done in the portal.
IDP webhooks:
-Receive a webhook when the document class/type has been assigned the first time or updated
-Receive a webhook when the field values have been assigned during extraction
There are 8 data field types that are pulled via API:
- First name
- Last name
- Date of birth
- Middle name
- Zip code
- MRN (medical record number)
- Date of medical encounter
- Patient identifiers
If you have any other questions, please reach out to support@documo.com.