Product Overview

Convey Connect is your one-stop shop for integration needs in healthcare.

PACs

Here we overview what the PACs integration will do. We can offer two different tiers. First will be a lower priced tier where the customer can open up their local PACs to a mobile viewing solution for their users. This should be a simple price per user access and a single setup fee.

Second product would be a competitor to LaurelBridge that offers enterprise level support, connectivity to SSO options for the mobile viewer, encrypted DICOM server features etc. People can use the feature to perform auto-pushes and manual pushes to send images to third parties, or receive images from third parties to add to the local PACs.

Product runs as a docker image on a machine. The enterprise version needs significantly more local resources and drive space. All versions should have a cloud hosted web portal for administration, reporting, etc.

Data

Here we discuss our data integration tools. This is intended to be a Mirth replacement. I think we will need a base level where data is simply pushed to our cloud where is can be accessed or processed via a series of configurations. Examples would include:

Triggers / Webhooks along with third party integrations like "send a message via tool" or "push notification to XYZ", or send robocall to relatives when threshold met.
Allowing third party apps access to the data. We will need additional data structures outside the raw patient data to make this useful. Think things like user-based patient lists.
"Channels" would be similar to Mirth where incoming data is processed and pushed to outbound destinations which could be any array of things, including REST APIs, Files, etc.
All data should have some kind of "ownership" setup. This would then give a lever for owners to control access to their data. Owners could be hospitals or hospital systems, third party labs, imaging centers, or even individual patients. Need multiple owners technically... As an owner I would have access to an audit log and would possibly set a process to grant access. Example, a hospital would grant access to a third party system, while a patient might grant access to a relative or to a new hospital system for ADT etc.
Python would be the code stored that executes as part of the channel. We will need to add some serious wrappers on it though to prevent bad actors doing seriously bad things.
We should also add full support for FHIR in both directions.

Cortex

Cortex is the AI service from Convey. This system uses natural language as a means to query data and perform requested tasks.

Execute "agent workflows" that connect several different AI agents and standard APIs into complex chains. Enterprise only. Example: scheduling surgery (which may involve communications with doctors, access to a third part / EMR scheduling system)
Query patient information. Asking summary style questions or detailed questions about the data in a patient's record. "What's the INR?". "What did the last doctor say?". Natural language queries gather appropriate data and convert them into data "views" which are presented to the end user. Views may include graphs, data points, text, media like PDF files or images, and radiology studies.
Creating patient information. This would include writing notes, adding orders, signing dispositions, approving discharges, and logging data points. These interfaces require Connect to grant access to data pushing via either FHIR or HL7. Cortex can make recommendations to the user to increase billing outcomes or improve patient outcomes.

Network

The network product is a communication platform that allows text-based communications between healthcare providers, patients, vendors, etc in a HIPAA compliant manner. There should be three tiers:

Personal - An individual can log in using their email, google, or apple account and access a basic chat feature and a read-only view of patient data that is directly shared with them by other users. They can reach out to people outside the system via phone number or email, and can reply to anyone who sends them messages or is found in their contacts (in-app as well as phone). This will be free, won't save messages after a certain date, etc.
Professional - Allows the creation of an "organization" where multiple users can be added and managed from an access perspective. Additional add-on tools become available such as Cortex AI, and Connect for EMR data access. Professional users can create chat channels, and share patient data with any users of the system via links etc. Add-ons (integrations) cost extra, and there is a set monthly price per user. Messages are saved indefinitely. Cortex wrappers on patient data from feeds is included, but not more advanced "agent" features.
Enterprise - Same as professional but comes with increased support, integration with SSO solutions for users (such as Okta, Azure, LDAP, etc). Cortex supports "agent" features to perform elaborate tasks.

PreviousConvey Health NextGetting Started

Last updated 11 days ago

PACs

Data

Triggers / Webhooks along with third party integrations like "send a message via tool" or "push notification to XYZ", or send robocall to relatives when threshold met.

Allowing third party apps access to the data. We will need additional data structures outside the raw patient data to make this useful. Think things like user-based patient lists.

"Channels" would be similar to Mirth where incoming data is processed and pushed to outbound destinations which could be any array of things, including REST APIs, Files, etc.

All data should have some kind of "ownership" setup. This would then give a lever for owners to control access to their data. Owners could be hospitals or hospital systems, third party labs, imaging centers, or even individual patients. Need multiple owners technically... As an owner I would have access to an audit log and would possibly set a process to grant access. Example, a hospital would grant access to a third party system, while a patient might grant access to a relative or to a new hospital system for ADT etc.

Python would be the code stored that executes as part of the channel. We will need to add some serious wrappers on it though to prevent bad actors doing seriously bad things.

We should also add full support for FHIR in both directions.

Cortex

Cortex is the AI service from Convey. This system uses natural language as a means to query data and perform requested tasks.

Execute "agent workflows" that connect several different AI agents and standard APIs into complex chains. Enterprise only. Example: scheduling surgery (which may involve communications with doctors, access to a third part / EMR scheduling system)

Query patient information. Asking summary style questions or detailed questions about the data in a patient's record. "What's the INR?". "What did the last doctor say?". Natural language queries gather appropriate data and convert them into data "views" which are presented to the end user. Views may include graphs, data points, text, media like PDF files or images, and radiology studies.

Creating patient information. This would include writing notes, adding orders, signing dispositions, approving discharges, and logging data points. These interfaces require Connect to grant access to data pushing via either FHIR or HL7. Cortex can make recommendations to the user to increase billing outcomes or improve patient outcomes.

Network

The network product is a communication platform that allows text-based communications between healthcare providers, patients, vendors, etc in a HIPAA compliant manner. There should be three tiers:

Personal - An individual can log in using their email, google, or apple account and access a basic chat feature and a read-only view of patient data that is directly shared with them by other users. They can reach out to people outside the system via phone number or email, and can reply to anyone who sends them messages or is found in their contacts (in-app as well as phone). This will be free, won't save messages after a certain date, etc.

Professional - Allows the creation of an "organization" where multiple users can be added and managed from an access perspective. Additional add-on tools become available such as Cortex AI, and Connect for EMR data access. Professional users can create chat channels, and share patient data with any users of the system via links etc. Add-ons (integrations) cost extra, and there is a set monthly price per user. Messages are saved indefinitely. Cortex wrappers on patient data from feeds is included, but not more advanced "agent" features.

Enterprise - Same as professional but comes with increased support, integration with SSO solutions for users (such as Okta, Azure, LDAP, etc). Cortex supports "agent" features to perform elaborate tasks.