Microsoft Health Bot service, a solution that enhances healthcare organizations globally

Microsoft stood up to the challenge. The Microsoft Health Bot service has supported multiple health organizations in the face of the coronavirus outbreak. The good news is that, thanks to the integration with Infermedica API, the Health Bot also works with a wide range of other diseases.

Illustration by Tomek Tuz

Microsoft Health Bot is a flexible platform for creating virtual health assistants. During the recent distressing time of coronavirus, it has been quickly implemented by the U.S. Centers for Disease Control and Prevention (CDC), Providence, and Walgreens, as well as healthcare facilities across the world. This includes Emergency Medical Services Copenhagen, Rome Spallanzani Hospital, Helsinki University Hospital, and other international healthcare organizations.

CDC implemented the Coronavirus Self-Checker with support from Microsoft Health Bot (source)
CDC implemented the Coronavirus Self-Checker with support from Microsoft Health Bot (source

Since March 2020, health organizations have reached over 40 million individuals with Microsoft Health Bot. All patients were provided with reliable medical information and healthcare workers were supported in the triage processes.

A fast and simple implementation of Microsoft’s Health Bot helped multiple organizations to reduce the load of healthcare systems and focus on curing COVID-19. As Roni Gamzu, CEO of Tel Aviv Sourasky Medical Center, comments “other organizations should know that it does not take a tech expert to deploy the bot. It is very easy, fast, and intuitive”.

What is the Microsoft Health Bot service?

The Health Bot service is a cloud service equipped with Artificial Intelligence (AI) algorithms, Natural Language Processing (NLP), and medical content. It is dedicated to healthcare providers, healthcare companies, telemedicine providers, and health insurers. Any of them can use, customize, and deploy it as a part of their services and enable personalized and intelligent digital health services to their end-users. It is available in a SaaS model.

The bot prepared by Microsoft has several built-in scenarios. Organizations can use it to provide patients with triage and symptom checking tools, general information about conditions and symptoms, and ‘find a doctor’ features. They can also use it as the base for health plan inquiries, scheduling appointments, or finding providers of certain medical services. All this comes in a natural and conversational form that is pleasant for patients.

Thanks to these advanced scenarios, Aurora Healthcare, Quest Diagnostics, and Premera Blue Cross decided to set up their health bots with Microsoft's services and provided their customers with a quick way to get direct answers to their questions.

What’s most important, however, is that Microsoft Health Bot is compliant with over 50 security and privacy frameworks, including HIPAA and HiTrust standards.

Key technologies used in Microsoft Health Bot

The Health Bot service is built on top of Microsoft Azure. Organizations using it get access to a management portal that enables them to control the configurations and extensions of their bot instances. When it is ready, an individual instance of the health bot can be embedded in various digital touchpoints.

Scheme of information flow in Microsoft Health Bot (source)
Scheme of information flow in Microsoft Health Bot (source)

To set up this health bot, Microsoft leveraged its well-known framework for enterprise-grade conversational AI. In this case, the bot framework serves as a messaging and routing platform with end-users. It guides patients through built-in, healthcare-specific scenarios filled in with thousands of symptoms, risk factors, and conditions.

Additionally, it allows organizations to create custom scenarios. These can be developed with Microsoft's visual designer tool, and answer the most complex and intricate needs (like accessing patients’ records via FHIR protocols).

The role of Infermedica API in MHB

From the diagnostic point of view, the two components play a crucial role - patient triage and NLP. The first one is responsible for the analysis of the patient’s symptoms and suggesting the most probable conditions. The second understands the natural speech of users and connects it with the base of medical content. Both of them are delivered within Infermedica API.

Infermedica API is a cutting-edge triage engine that enables symptom checking based on advanced AI mechanisms and a vast database of medical terms. The reliability of the medical content is ensured by the multi-step validation processes used in Infermedica, as well as verification provided by Microsoft’s experts.

Symptoms, risk factors, and conditions available through Infermedica API allow for accurate patient triage in 17 languages. The key to the proper pre-diagnosis is understanding the end-users' intention and way of communicating. Translation between our every-day speech and medical terms is ensured with NLP. This technology understands the laymen's terms and connects them with medical vocabulary. It also learns over time. The accuracy of NLP is essential, which is why the Infermedica and Microsoft teams joined efforts in its development.

Learn more about Microsoft Health Bot service and Infermedica API.