CONSULT is a modular decision-support system (DSS) intended to help patients suffering from chronic conditions self-manage their treatments.
The system combines input from commercial wellness sensors, a patient's electronic health record and clinical guidelines; interacts with these inputs using data science tools, and a branch of AI known as argumentation; and presents insights back to the user using a dashboard and interactive chatbot.
To illustrate the operation of the CONSULT system, let's consider the following scenario:
`A stroke patient has exacerbated blood pressure, how can they quickly and transparently be informed of this?'
Patient data
To begin, a patient first registers with our system, we grab their electronic health records from their provider, convert them to the FHIR standard, and store them in a FHIR server.
For this demo, we've just created a sample patient on your behalf, Joey, and collected their health records. Welcome to the CONSULT system, Joey.
Sensor data
In addition to static health data collected from health records, CONSULT also uses live patient data, not from expensive medical devices, but from consumer-grade devices, which a patient may already own, or can readily acquire. In this case, Joey has a Withing's Blood Pressure monitor, so lets connect this to our system on their behalf. Done!
Joey, a stroke patient, now experiences an exacerbation in their blood pressure. Click here to send some simulated readings through the system from this monitor to reflect this.
Data analysis
Right now, a set of blood pressure readings are being sent through our system, via a device integration service, they're also being converted to FHIR, and also being stored in our FHIR server, matched to the patient data that has already been acquired. They're also being analysed using our data miner, which checks for any exacerbations.
Dashboard
Our data miner has identified that Joey has a slight exacerbation in their blood pressure. Click here to open the CONSULT dashboard to view this information.
Chatbot
Similarly, we can talk to CONSULT's interactive chatbot, Connie, to hear about this exacerbation. Click here to see what she has to say.
By interacting with Connie about this alert, we can also ask for an explanation of why we received this alert. To do this, CONSULT uses a dialogue templating syntax and argumentation to analyse the results found by the data miner.
We also record the provenance of the decisions made, as well as other information in the system.
What else does CONSULT do?
- CONSULT also uses argumentation to interpret clinical guidelines to provide information on health symptoms. Try typing '/hello' in the chatbot interface to start a discusion with Connie about a symptom.
- Collects heart rate and ECG readings, as well as more detailed blood pressure readings, also matching this data to a patient's health record data. Click any of these links to send different types of data through the system.
- Engages patients in an exercise to collect information about their state of mind. Try typing '/hello' in the chatbot interface to start a discusion with Connie about this, or use the dashboard.
- Identifies interactions between clinical guidelines
