Microsoft software already powers a lot of healthcare, from PCs in GP surgeries to embedded systems controlling medical equipment in hospitals. But the increasing focus on digital health, personalised care and the use of AI to improve diagnoses and outcomes means the same tools and services that businesses use are useful for health providers too. Today that’s about Office 365 (and in particular Microsoft Teams) for collaboration and communication, virtual assistants and chatbots, and security, compliance and privacy for storage and governance of healthcare data on Azure. Looking to the immediate future, there’s Azure AI and Cognitive Services for building diagnostic AI tools, and confidential cloud computing that lets different health organisations share patient data for research and machine learning without exposing private information to each other.
Medicine is a very collaborative field. It’s not just passing health records around, important as that is. Nurses and physiotherapists and staff from care homes who are treating the same patient in different places can be much more productive if they can chat with each other, and nurses on a ward often use messaging apps to arrange their shifts and schedules – but doing that in a consumer messaging app isn’t secure or compliant. The boards that decide on who to prioritise for transplants or which treatments to follow in complicated cancer cases are made up of senior doctors who aren’t just busy; they often work in different locations and the time it takes to travel to a board meeting can dictate how often medical boards can meet.
The health version of Microsoft Teams supports secure compliant chat and video meetings, along with workflow and shift management. Staff can share and annotate images from their mobile devices – something that’s very important for understanding medical issues – but the images aren’t saved even on the device where the picture is taken, so there are no worries about staff accidentally leaking confidential photos. The background blur feature is particularly useful if someone is calling from a ward or office where patients or confidential health records might be visible behind them. There are templates designed for wards and hospitals to make it easier to set Teams up consistently with the structure of channels that will be useful for collaborating within and across departments.
In particular, there are templates for the ‘huddle’ meetings that health teams regularly hold, making it easier to centralise notes, follow checklists and see patterns of symptoms using Teams with SharePoint lists, tasks and boards in Planner, Power BI tabs to visualise metrics and what impact they’re having, and using the Bot framework to capture new ideas in the right format to put into a list, plan or visualisation. If you want to try that out, there’s a sample on GitHub that sets up a complete environment so you can see it as an option or adapt it for your own environment.
Health records get faster
Teams continues to get healthcare-specific features, including priority alerts for urgent messages that repeat until the message is answered and delegation to automatically redirect important messages when a doctor can’t be disturbed because they’re in surgery or with a patient, and a secure hub for coordinating care for multiple patients (currently in private preview). It already integrates with many patient care and record systems as part of Microsoft’s health data platform.
The aim is to improve interoperability of health data by having all three Microsoft clouds – Azure, Office 365 and Dynamics 365 – understand the ‘language’ of healthcare in terms of data standards (as well as compliance and security). So whether you’re building an appointment scheduling system, analysing the impact of a different cleaning schedule, tracking re-admission statistics, working with patients or selecting patient data for research, information about symptoms, treatments, medications and health issues can be referenced with the same terms and viewed by anyone who’s supposed to have access, whatever system they’re using.
Twenty years ago, most healthcare records were on paper; now nearly all of them are digital. Changes to US healthcare rules mean that shared electronic health records are finally becoming a reality but being able to move patient data between healthcare providers is important in the UK too. Digital records are easier to file, easier to share – so your GP can explain the results of your hospital tests – and easier to search, whether that’s to look back and see when you last had a tetanus jab or to use in machine learning systems that try to understand broader health patterns.
Microsoft Azure is the first cloud to have native support for the FHIR (Fast Healthcare Interoperability Resource) Format that uses REST APIs to normalise health data into standard representations that are easier to exchange between different health record systems, with unique identifiers so you don’t accidentally connect records from a different patient who had the same reference number in another system. The Azure FHIR API handles ingesting, normalising and storing the data securely, with auditing and role-based access controls.
Great Ormand Street is using FHIR in Azure to connect to Aridhia who provides their health research environment, so they can scale out machine learning workloads on Azure to work on patient data in the system their paediatricians are already familiar with.
You can also view FHIR data natively in Teams: the Patient App solution in Teams saves the time hospital doctors would otherwise spend copying patient records into a spreadsheet they can take on their rounds with them.
FHIR isn’t the only standard, so Microsoft is supporting HL7 – the Health Layer 7 international standards for transferring clinical and administrative data between applications, DICOM – the Digital Imaging and Communications in Medicine standard for storing, processing and displaying medical imaging, and other standards like the BAM format for genome sequence data, in Azure, Office 365 and Dynamics 365.
Tell the doctor how you feel
Virtual assistants and chatbots are as useful for patients in healthcare settings as they are for consumers and customers in other areas. There are lots of tools for building chatbots, but for something as important as health, it’s vital to get the language right, to make sure that you don’t miss a patient presenting with a serious condition or misunderstand what they say about symptoms. That’s why LUIS, the Language Understanding Service in Azure Cognitive Services, is a particularly useful option: you can make sure you include the phrases and questions patients typically use and match them to the correct medical terminology.
The Huddle templates for Teams use LUIS with the Microsoft Bot Framework, giving you a good example of how to build a medical bot. If you don’t want to build and run your own bot, you can also use LUIS with the new Microsoft Health Bot service. This has a portal where you can create bots for everything from scheduling appointments to triaging medical conditions using symptom checkers or helping patients find information about specific diseases based on medical content from trusted sources. It’s a way to put a friendly front end on something that’s often scary, while keeping this very personal information confidential.
This is the first in a series of three posts about Microsoft’s innovations in Healthcare. Read the second blog in the series, “What Microsoft is doing in healthcare research“
Grey Matter has a Cloud Services team who are experts in managing and enabling Microsoft cloud technology and Azure-related projects. If you need their technical advice or wish to discuss Azure, O365, Visual Studio, or your use case, contact them on +44 (0)1364 654200.