We’re going to be sharing a story every week for the 12 weeks of summer, showing you how healthcare organisations are using technology to transform patient outcomes and increase productivity. For the third blog in the series, Phil Allen, Technology Strategist for Healthcare and Life Sciences at Microsoft, shares four practical steps to taking an ethical approach to using AI in healthcare. 

The ethical use of data has dominated headlines in recent times, but we rarely hear discussions over the ethical use of AI.

Microsoft CEO Satya Nadella is one of the few to tackle the tricky issue of ethical AI. In an article for Slate, Satya riffed on Isaac Asimov’s famous ‘Three Laws of Robotics’, producing his own ‘Ten Laws of AI’. It’s a brilliantly considered set of rules that should guide how AI is developed and deployed over the coming years.

Introducing AI and ethics to your business

In a healthcare setting, artificial intelligence promises improved diagnosis, better patient outcomes and, crucially, more time for professionals to spend with their patients. But the industry’s heavy regulation means mass adoption relies on AI data being measured by the same patient efficacy and safety standards as drugs, devices, and diagnostics.

For AI to have the greatest impact, key questions must be asked:

• How do we find, access, and understand this data, integrating it across healthcare, clinical, and research applications to improve the patient experience and outcomes?
• How should we quantify the input, so that the output satisfies both the regulatory and healthcare use case?
• How can we increase AI’s rate of adoption when only 20% of medical data is machine-readable?

The answer: FAIR Data Principles.

FAIR Data Principles, if you haven’t heard of them, are a set of nationally recognised guidelines that qualifies data and metadata for AI applications.

If your organisation wishes to take an ethical approach to AI, it’s worth studying how the healthcare sector is adapting to the technology right now.

Step 1. Make your data searchable

The first step is making your data searchable – or, to put the F into FAIR, ‘findable’. All data requires a persistent identifier that ‘follows’ it wherever it goes, and metadata that explicitly defines what it is. The trick is to make sure it’s a completely unique identifier. Human operators and computers are then able to find and actually use that data – rather than letting it fester, forgotten on a server somewhere.

Step 2. Ensure all data is accessible

Users must easily understand how to obtain the data they need, even if access requires authorisation. But it also means they should be able to access the data without specialist software – it should literally be as simple as clicking a link. Delivering accessibility also means the metadata can still be accessed, even after the core data has been removed (typically because it takes up too much space on hardware or a server).

Step 3. Make data interoperable

Even if people and machines can find and access data, it’s pointless if a computer (or users) can’t read it. Just as you’d expect a business meeting to be conducted in a mutually understood language, data should be created in a commonly agreed programming language. This lets it better integrate with other data and work across multiple systems

Step 4. Your data must be reusable

The idea underpinning reusability is linked to searchability. Remember when you ensured the data you held was rich with contextual metadata that explicitly stated what it was? That now lets a user or computer know whether the existing data is worth using in new ways, e.g. creating datasets for AI training. Another core part of this is to add clear usage licences, so the data can actually be used.

How organisations deliver FAIR-ness

The UK’s Health Data Research, the national institute for health data science, is already adopting FAIR Data Principles. Data on various conditions will then be available in ‘FAIR form’ across all major academic and healthcare partners.

They are not alone.

Microsoft partner Aridhia (www.aridhia.com) is keen to establish data security and consistency norms founded on FAIR principles. The Scottish research company provides a Digital Research Environment powered by Azure, helping teams access, curate, analyse, and publish data in a safe, secure, and audited setting. Teams experiment with data science approaches, building a picture of how a project moves from experiment to implementation.

Project Fizzyo is a ground-breaking example; a multi-disciplinary collaboration establishing the routine collection of data relevant to the treatment of Cystic Fibrosis (CF) in children. By aligning the flow and linkage of data with a CF clinical pathway, Fizzyo can build a dataset using FAIR Principles. Professionals can access these datasets, nurtured and grown over time, to build an evidence base of patient and clinical utility.

Increasing demand for ethical AI across healthcare

Aridhia – which has customers in the healthcare, academic and pharmaceutical fields – has seen substantial demand for machine learning and AI innovations in highly regulated and specialised healthcare environments. This is particularly true where data doesn’t yet conform to the FAIR principles.

Of course, healthcare data is a broad term. At one end of the spectrum, we have academic health science producing greater insight through advanced genomic and imaging technologies. In the middle, there’s ‘traditional’ healthcare data generated through electronic patient records in primary and secondary care. And, for consumers at the other end of the spectrum, we see ever-increasing routine longitudinal data generated through wearable devices.

Initial results from focused AI applications in MRI scan analysis or digital retinopathy in diabetes are encouraging. However, the sector is a complex system; individuals and circumstances are unique. The successful, widespread adoption of ethical, AI-ready data in healthcare demands the introduction of data that is, above all, FAIR.

Find out more

Learn about AI in healthcare

Discover how to maximise the AI opportunity

About the author

Phil Allen is a Technology Strategist for Healthcare and Life Sciences at Microsoft. Passionate about the potential of technology to transform health and healthcare, he has 20 years’ experience at Microsoft. In that time, Phil has helped customers and partners chart a course to the future with technology, to get the best out of their investment.

The post AI in Healthcare: 4 steps to taking the ethical approach appeared first on Microsoft Industry Blogs - United Kingdom.

We’re going to be sharing a story every week for the 12 weeks of summer, showing you how healthcare organisations are using technology to transform patient outcomes and increase productivity. For the second blog in the series, Phil Allen, Technology Strategist for Healthcare and Life Sciences at Microsoft, will be sharing an example of how cloud computing and automation can help deliver high-quality, patient-centric care.

I’m always excited to discover how people are harnessing Microsoft Azure in a healthcare setting. DrDoctor, which delivers patient-facing automated technology, is one such example.

The company is working with NHS Wales and hospitals including Nottingham University Hospital to digitally transform the health service. This is achieved by automating key procedures and letting patients take greater control over their own care. Success comes by embracing both public cloud and open standards to make interoperable, future-proof solutions that are scalable and secure.

Within 20 years, the company hopes to see a new type of hospital; hospitals that work in radically different ways, removing time-consuming paperwork and using digital tools for better patient-clinician interaction and managing clinical risk.

One of DrDoctor’s core beliefs is in changing systems from the inside out. This isn’t a quick fix, then, but a cultural shift. One that creates a secure healthcare infrastructure that works for both patients and practitioners by adapting current processes and adding incremental value. Putting cutting-edge technology into the hands of providers delivers the first tremors of a fundamental shift in the quality and cost of delivering long-term, high-quality, patient-centric care.

Moving to cloud computing

Since NHS Digital released their guidance for NHS organisations adopting public cloud, Microsoft Azure has become the go-to cloud computing platform for DrDoctor. In a recent interview, Tom Whicher, DrDoctor’s co-founder, told Health Tech Newspaper that the business pursued this strategy because it “allows us to leverage [Microsoft’s] machine learning technology, scale much more quickly and to reach more patients. We want to make sure we’re always using, developing on and giving patients the latest and best in high tech.”

Warming to his subject, Whicher added that, “the biggest technology challenge for the NHS is beginning to implement a new set of standards. Going forward, all of our interfacing will be [Fast Healthcare Interoperability Resources] FHIR compliant. In order for the NHS to achieve its ambitions, it needs to adopt FHIR, two-way integration, and open standards at scale. That means getting lots of legacy systems to integrate with newer tech and solving lots of those dreary back-office problems.”

Helping overcome that challenge is Microsoft’s Azure API for FHIR, a fully-managed, standards-based and compliant healthcare data platform. It’s not by accident that the cloud computing platform is capable of powering Internet of Medical Things (IoMT) scenarios, population health research projects, AI-powered diagnostic solutions and more. With security and privacy fully embedded into the service, conforming to global health privacy and security standards, users have full control over how patient data is used and stored.

Clinical analytics can be pulled together from multiple records systems, then normalised using common models and specifications. Leveraging such data in AI workloads means users gain the insights needed to power new systems of engagement. This includes clinician and patient dashboards, diagnostic assistants, population health insights, and connected healthcare scenarios, such as Remote Patient Monitoring.

All of which lets healthcare organisations bring clinical health data into the cloud based on the interoperable data standard FHIR. It’s what allows professionals to more easily respond to changing business dynamics.

Creating change in healthcare

DrDoctor was determined to push ahead with FHIR as the standard for interoperability in UK healthcare. Microsoft have developed an open-source FHIR server that uses Azure services to deliver a fully operational API. This is backed by a data repository, offering an enormous head-start for those moving across to FHIR.

Alberto Amati, Senior Developer at DrDoctor and self-proclaimed ‘cloud expert’, explained in his blog how this has been enabled over the data-based Health and Social Care Network (HSCN). By deploying an Azure Application Gateway in a new subnet in the HSCN VNet, and only configuring it with a private IP address, the business could use it to control access to the FHIR App Service. The team used VNet Service Endpoint combined with a new VNet Integration. This ensured the Microsoft Azure Cosmos DB could only be accessed by DrDoctor’s own FHIR service.

Despite recent business innovations, it’s Whicher’s belief that, “we need to ensure that we bring our hospital partners with us on this journey. It isn’t easy creating change in healthcare… We need to make sure we stay on top of the latest technologies which is why we’ve invested in moving all of our product to Microsoft Azure.”

By leveraging the Azure Cloud and Microsoft’s FHIR API and Server, automated technologies, like the DrDoctor platform, create an environment where hospital teams can, in Whicher’s words, “refocus their time and focus on all the value-adding work for patients.”

Find out more

Explore Azure for healthcare

Discover how virtual consultations are improving patient care

About the author

Phil Allen is a Technology Strategist for Healthcare and Life Sciences at Microsoft. Passionate about the potential of technology to transform health and healthcare, he has 20 years’ experience at Microsoft helping customers and partners chart a course to the future with technology and get the best out of their investment.

The post DrDoctor prescribes automation for healthcare appeared first on Microsoft Industry Blogs - United Kingdom.