Matching patients to clinical trials is a huge challenge. Based on the World Health Organization, in 2021 there were more than 82,000 clinical trials actively recruiting participants worldwide[1], containing increasingly complicated trial eligibility criteria. Given that a patient’s clinical information can be composed of hundreds of clinical notes, requiring review and analysis by a healthcare professional, matching patients to clinical trials becomes an unsustainable process.
The Trial Matcher capability in Microsoft Project Health Insights addresses this challenge – by providing a tool to effectively match patients with diverse conditions to clinical trials for which they may be eligible. This is done through analysis of patient data and the complex eligibility criteria of the relevant trials.
Project Health Insights is an AI service that provides built-in AI models for key scenarios in the health domain, including clinical trial recruitment.
Available in public preview, the Trial Matcher model is already deployed by several organizations worldwide to enable matching of patients to clinical trials.
Open-source blueprint for deployment of patient-centric chatbot
Today, we are introducing a new open-source blueprint to quickly deploy Trial Matcher chatbot experience with one click of a button.
The Trial Matcher chatbot is an Azure Health Bot instance with Clinical Trial Matching template integration. The chatbot enables patients to provide relevant patient information and checks their eligibility for potential trials, and after finding matches, provides a list of clinical trials the patient may be eligible for. Using this blueprint organizations can now expedite the onboarding process and start realizing the time saving benefits of the trial matcher capability right away.
Deployment instructions
The Azure resources that will be created and deployed when using this open-source blueprint include:
Follow the instructions of the Readme file to deploy the Azure Health Bot with Trial Matcher integration based on the clinical trial matching template.
Figure 1 List of scenarios enabling Trial Matcher from Project Health Insights
After a successful deployment, open your generated instance of Azure Health Bot, and go to the management portal. Under Scenario Management, you will see several scenarios that power the clinical trials matching from Project Health Insights that were automatically ingested by the blueprint into your Azure Health Bot instance. To engage with the matching experience, start a chat conversation by clicking on the chat icon in the blue task bar on top to the right, and ask for clinical trials.
For example, you can ask the chatbot: ‘find clinical trials in Arizona for a 24-year-old female with lung cancer’.
The chatbot will use Azure AI Language Understanding that was ingested in your LU resource by the blueprint to recognize the pre-built intent in the words "find clinical trials", and extract the relevant clinical trial characteristics to look for from your request.
The data provided in this case includes patient age (24), patient sex (female), patient medical condition (lung cancer), and preferred trial location (Arizona).
Figure 2 Example chatbot conversation for clinical trial matching
To improve the patient qualification process, additional details can be captured by asking the patient for further information. Your chatbot does that by generating a question and waiting for user input. After receiving the needed response, the Azure Health Bot sends a request to the Trial Matcher model, receives a trials list, and uses the response to start asking additional questions about the patient's condition to enable the Trial Matcher model to perform a more accurate qualification and narrow down the list of clinical trials the patient could be eligible for.
After a few iterations of questions and additional information provided by the user, you’ll be provided with a list of clinical trials that are relevant according to the clinical data and preferences that were provided. The model also provides a list of ineligible trials with reasons (one or more) for ineligibility so that the patient may assess this information independently.
Important
Trial Matcher is a capability provided “AS IS” and “WITH ALL FAULTS.” Trial Matcher isn't intended or made available for use as a medical device, clinical support, diagnostic tool, or other technology intended to be used in the diagnosis, cure, mitigation, treatment, or prevention of disease or other conditions, and no license or right is granted by Microsoft to use this capability for such purposes. This capability isn't designed or intended to be implemented or deployed as a substitute for professional medical advice or healthcare opinion, diagnosis, treatment, or the clinical judgment of a healthcare professional, and should not be used as such. The customer is solely responsible for any use of Trial Matcher.
Get started today.
Follow the guidelines in the Readme file to deploy the Azure Health Bot with Trial Matcher integration and then to start a chat conversation using the bot to experience patient centric workflow microsoft/ClinicalTrialsBlueprint: Setup the resources for provisioning Clinical Trials search (gith...
Learn more about Project Health Insights and Trial Matcher
Project Health Insights | Microsoft Learn
Trial Matcher documentation - Project Health Insights | Microsoft Learn
[1] Percent of trials with enrollment year preceding registration year (1999-2022) (who.int)