Remote Patient Monitoring with Internet of Medical Things (IoMT)

Summary

The video discusses a Microsoft Research team's project focused on healthcare data interoperability, using the FHIR (Fast Healthcare Interoperability Resources) open standard. Rashmi Raj from Microsoft showcases how the team is developing tools to securely ingest and normalize data from various IoT medical devices, storing it in a FHIR server via Azure. The project aims to consolidate scattered data sources, creating a comprehensive patient record that can be used for applications like remote patient monitoring and AI-based healthcare solutions. The open-source project encourages community contributions for further enhancement.

Takeaways

• 🔗 Interoperability is a critical challenge in healthcare due to the diverse data coming from various medical devices.
• 💻 Microsoft Research, led by Rashmi Raj, is developing a data platform that uses the FHIR (Fast Health Interoperability Resources) standard to address this challenge.
• 🌐 FHIR (pronounced 'fire') is an open standard rapidly growing for healthcare data exchange and interoperability.
• 🔐 The platform focuses on secure and scalable data ingestion from IoMT (Internet of Medical Things) devices, utilizing Azure services.
• 🏥 Key use cases include remote patient monitoring, clinical trials, telehealth, and home care, all powered by data pulled from various IoT medical devices.
• 📂 The project features the IoMT FHIR connector, an open-source tool available on GitHub, allowing devices to send data to Azure and store it in a FHIR server.
• 🔓 Open-source contributions are encouraged, enabling users to extend and improve the platform by creating new data templates and normalizing device data.
• 🧑‍💻 The platform integrates data from electronic medical records (EMRs), labs, retail, and social determinants of health to build a comprehensive patient record.
• ⚙️ Users can deploy the IoMT FHIR connector via Azure and configure it to manage identity and security, ensuring HIPAA and HITRUST compliance for protected health data.
• 📊 Real-world scenarios include using Postman to query data from the FHIR server, and this data can be integrated into dashboards or used for AI and machine learning applications.

Q & A

What is the main challenge in healthcare data that the video addresses?

The main challenge is data interoperability—how to make sense of data coming from various healthcare devices and applications in a unified way.

What is the focus of Rashmi Raj's team at Microsoft?

Rashmi Raj's team focuses on creating a data platform for interoperability using the FHIR (Fast Healthcare Interoperability Resources) standard, specifically in the context of Internet of Medical Things (IoMT) devices.

What does FHIR stand for, and why is it important?

FHIR stands for Fast Healthcare Interoperability Resources. It is an open standard designed to enable the exchange of healthcare information electronically, which helps with data interoperability in healthcare systems.

What is the Azure IoMT FHIR Connector?

The Azure IoMT FHIR Connector is an open-source project that allows for the secure ingestion and normalization of data from IoMT devices into a FHIR-compliant format, enabling downstream business applications to use the data.

How does the IoMT FHIR Connector ensure security and compliance?

The IoMT FHIR Connector ensures security by using Azure's secure cloud infrastructure, which is HIPAA, HITRUST, and SOC2 compliant, ensuring the safe handling of Protected Health Information (PHI).

What types of data can be ingested and used by the FHIR Connector?

The connector can ingest various types of data, including data from IoMT devices (e.g., heart rate, respiratory rate), Electronic Medical Record (EMR) data, lab results, and social determinants of health data.

What scenarios can benefit from this data platform?

The platform can support scenarios such as remote patient monitoring, telehealth, clinical trials, and home care, by enabling the integration of various health data types into a single system.

How does the open-source nature of the IoMT FHIR Connector benefit the healthcare community?

The open-source nature allows developers and organizations to not only use the connector but also contribute to its development, such as by creating templates for different devices and expanding its capabilities.

What role does Azure API for FHIR play in this solution?

Azure API for FHIR provides a fully compliant and managed service that stores and normalizes the ingested data from IoMT devices, enabling developers to build applications on top of the FHIR API.

What is the significance of creating templates for different devices in this system?

Templates are essential because they map data from specific devices to the FHIR standard, allowing for the normalization of device data so that it can be processed and used in various healthcare applications.

Outlines

00:00

📊 The Importance of Interoperability in Healthcare Data

This paragraph highlights the challenge of managing large volumes of healthcare data from various devices, emphasizing the importance of interoperability. Rashmi from the Microsoft research team introduces a project focused on creating open standards for data management in healthcare. The discussion touches on building solutions around the FHIR (Fast Health Interoperability Resources) standard and mentions a project aimed at solving data challenges through secure, scalable technology for remote patient monitoring and other use cases.

05:01

🏥 Securing Protected Health Information on Azure

The second paragraph explains how healthcare data, including Protected Health Information (PHI), can be securely stored on Azure using services that comply with HIPAA, HITRUST, and other regulations. The use of FHIR standard enables interoperability, allowing the integration of data from various sources such as electronic medical records (EMR), lab results, and even social determinants of health. The goal is to create a comprehensive longitudinal patient record for advanced use cases like AI-driven healthcare models, ensuring privacy and security.

10:03

⚙️ Setting Up the IoMT-FHIR Connector in Azure

This section walks through the setup process for integrating Internet of Medical Things (IoMT) devices using the FHIR connector in Azure. It details the steps to deploy an FHIR server via Azure API and integrate IoMT devices by configuring them in Azure's IoT Central platform. The paragraph also explains how the FHIR server handles secure data ingestion and device interoperability, using open-source tools available on GitHub to help developers get started.

🔧 Data Templates for Normalizing IoMT Device Data

In this paragraph, Rashmi discusses the process of normalizing data from various IoMT devices using templates that map device data to FHIR-compliant formats. She mentions the availability of templates for common devices and the flexibility to create new templates for other devices. The open-source project encourages contributions to these templates, aiming to simplify the integration of any medical device into the FHIR ecosystem. This data normalization is key for creating interoperable healthcare applications.

💻 Demonstrating FHIR Data Flow with IoMT Devices

This section demonstrates the complete setup of an IoMT device pipeline, from simulating device data in Azure IoT Central to viewing normalized data using Postman. It shows how telemetry data like heart rate and respiratory rate from devices are sent through the IoMT-FHIR connector, securely ingested, and then stored in the FHIR server. The paragraph emphasizes how easy it is to scale this solution to handle millions of devices while ensuring security and privacy for healthcare data.

🛠️ Open-Source Contributions and Feedback

The final paragraph invites the audience to explore the open-source GitHub repository for the IoMT-FHIR connector, contribute templates and code, and provide feedback. Rashmi expresses excitement about the community's potential to build on the team's work and help improve healthcare through interoperability and scalable solutions. The segment concludes with a call to action for developers to engage with the project.

Keywords

💡Interoperability

Interoperability refers to the ability of different systems, devices, or applications to exchange and make use of information. In the video, the challenge of ensuring interoperability in healthcare is highlighted, as data comes from numerous Internet of Medical Things (IoMT) devices. The team’s solution focuses on making sense of all this diverse data by adhering to open standards like FHIR, allowing various healthcare systems to work together seamlessly.

💡FHIR (Fast Healthcare Interoperability Resources)

FHIR is an open standard used for exchanging healthcare information electronically. It aims to simplify interoperability between different healthcare systems. The video discusses how Microsoft’s team builds around this standard to handle and normalize data from IoMT devices, enabling it to be consumed by downstream business applications securely and at scale. The video emphasizes its importance in creating a unified patient record and supporting applications like remote patient monitoring.

💡IoMT (Internet of Medical Things)

IoMT refers to the network of medical devices that are connected to healthcare IT systems via online computer networks. The video explains how these devices produce vast amounts of data, which must be collected, secured, and analyzed. Microsoft’s team has developed an IoMT FHIR connector to ingest and normalize data from these devices, making it easier for healthcare professionals to use the data in real-world applications such as clinical trials or telehealth.

💡Azure API for FHIR

Azure API for FHIR is a managed service by Microsoft that allows healthcare systems to store and manage healthcare data in compliance with FHIR standards. In the video, Rashmi's team uses this API to build a secure and scalable platform for ingesting data from IoMT devices. By integrating this API, they can ensure that the healthcare data is compliant with privacy standards such as HIPAA, while also enabling machine learning and AI-driven healthcare solutions.

💡Data Normalization

Data normalization in the context of this video refers to the process of converting various formats of data from different IoMT devices into a standard format that is FHIR-compliant. This is essential for ensuring that diverse data types can be used together in healthcare applications. The video highlights the importance of using templates to map device data to FHIR, enabling interoperability across various healthcare systems.

💡Remote Patient Monitoring

Remote Patient Monitoring (RPM) is the use of technology to monitor patients’ health data outside traditional healthcare settings, such as at home. In the video, it is mentioned as one of the key use cases for the IoMT FHIR connector, where data from IoMT devices can be integrated into healthcare systems securely, enabling doctors to track vital signs and manage patient health remotely in real time.

💡Azure IoT Central

Azure IoT Central is a Microsoft platform that simplifies the creation and management of IoT solutions by offering pre-built templates and integration tools. The video showcases its use to simulate devices like a smart vitals patch in the healthcare scenario. It is a key part of the solution where data from medical devices is ingested, standardized, and then passed on to the Azure API for FHIR for further processing.

💡HIPAA Compliance

HIPAA (Health Insurance Portability and Accountability Act) compliance refers to adhering to regulations that protect patient health information. The video emphasizes that the Azure platform and the tools developed by Rashmi's team ensure that data is stored and managed in a HIPAA-compliant manner. This is critical for maintaining the privacy and security of sensitive patient information in healthcare applications.

💡Longitudinal Patient Record

A longitudinal patient record refers to a comprehensive and continuous record of a patient’s health information over time. The video discusses how the integration of data from IoMT devices, electronic medical records (EMRs), and other healthcare data sources can help build a longitudinal patient record. This can be used to improve healthcare outcomes by creating a holistic view of the patient's health for more informed decision-making.

💡Machine Learning in Healthcare

Machine learning in healthcare refers to the use of algorithms and statistical models to analyze healthcare data for predictive insights. The video explains how Microsoft’s platform, once it normalizes the IoMT data and stores it in a FHIR-compliant format, can apply AI and machine learning to create new healthcare solutions, such as predictive models for patient outcomes, remote health monitoring, and personalized care.

Highlights

The critical challenge in healthcare is ensuring interoperability of data generated by numerous IoT devices.

Rashmi Raj from Microsoft Research discusses the development of an open-source project aimed at solving data interoperability in healthcare.

Microsoft's team is focused on building a platform using the FHIR (Fast Healthcare Interoperability Resources) standard to ingest and manage IoT data in Azure.

The project allows for secure and scalable ingestion of data from various IoT medical devices using the FHIR open standard.

Azure IoT Hub and Central services connect devices securely, and data is normalized and stored in a FHIR server for further application use.

The FHIR open-source project allows developers not only to use the tool but also contribute to it, helping expand its use cases.

A demonstration of the system shows how data from IoT medical devices can be mapped, normalized, and stored using the IoMT FHIR connector.

By using Azure's compliant cloud, healthcare providers can securely store Protected Health Information (PHI) and meet HIPAA and other regulatory standards.

The system integrates various healthcare data, including Electronic Medical Records (EMR), lab data, and social determinants of health data, to create a comprehensive longitudinal patient record.

The solution enables new healthcare use cases, such as remote patient monitoring, clinical trials, telehealth, and home care scenarios.

The demo simulates a medical device (Smart Vitals Patch) sending biometric data like heart rate and respiratory rate, which is normalized and stored in a FHIR server.

Postman is used to query the normalized heart rate data stored in the FHIR server, showing real-time ingestion and retrieval of IoT medical data.

The system is designed for scalability, allowing providers to connect and manage data from a few devices to millions, ensuring both security and scalability.

Azure IoT Central provides pre-configured templates for healthcare applications, speeding up the setup of remote patient monitoring solutions.

The open-source nature of the project invites the community to extend, contribute, and build new templates for various devices like Fitbit and Garmin.

The project is helping to advance healthcare technology by enabling the integration of IoT data into machine learning models for AI-driven healthcare solutions.

Transcripts

one critical thing in healthcare is data

data coming very often from tons of

different types of devices a big

challenge for building applications on

top of that data is interoperability how

do you make sense of all that data

Rashmi from the Microsoft research team

is here to show us a project her team is

developing to rationalise and to build

appropriately around an open standard

coal fire with a nice ash or connector

that's today on the IOT show

[Music]

[Music]

hi everyone you're watching the

internet-of-things show and I'm Olivia

your host as you might know there is a

lot of IOT devices out there and the

medical space is no exception there's

tons of devices that produce data in in

the healthcare systems in the medical

environments and there's a big challenge

today which is all that data needs to be

consumed by various systems and we need

interoperability and rashmi rise from

the Microsoft Research Group is here and

it's going to tell us about some work

that your team is doing

Rashmi thanks for joining the show -

that sure so for our audience can you

rapidly introduce yourself and tell us

what your team is doing at Microsoft

okay so my name is Rashmi Raj I am part

of health cloud and data team our team

is very focused on creating data

platform for interrupt interoperability

using fire open standard fire is rapidly

growing a standard stands for fast

health interoperability resources

my team is specifically focused on

creating tools for ingesting and pulling

data from these IO empty internet of

medical things which are devices

applications sensors pull that data

bring it to Azul with security and

scalability and in able to create huge

cases like remote patient monitoring

clinical trials tele health home care

scenarios so this is a full team of you

guys working on that yep I think it's

brilliant so we have we have different

services like azure IOT hub and central

that allow connecting these devices

securely as you were saying but the key

of your project really is about

ingesting the data coming from various

sources of these IOT are UMT are your

medical things and and then you know

anything that data and doing something

out of it in a very in Tripura balay so

let's dive a bit into you know what

would the team is producing and others a

connector for a sure that allows

ingesting that data let's jump into the

details and I know you have a nice demo

of how things are working okay so as I

said earlier one of the biggest

challenges

in health care has been lack of data

interoperability so we built a fire

server as open-source and released it on

github we also have Azure API for fire

fully compliant and managed service to

enrich the ecosystem of fire and

enabling the data ingestion from IO

empty device edge we have created IO

empty fire connector for Azure it's

available at open source on github this

allows to again connect these devices

with security and scalability ingest the

data normalize the data in fire and

store it in fire server and then

downstream business applications can

consume and huge open fire API for

building applications on top of it also

and and so basically the fact that this

open source means that people not only

can use it right but also they can

contribute exactly and there's a there's

that standard actually is a standard

body that actually works in that myself

contributes to in the formulate open

source project yes cool what I want to

see how that works

sounds good so in so in real world as I

was saying you will have device it so if

you see here high level architecture you

have devices and then you use the IMT

fire connector which is open source

project which will allow you to ingest

data and persist it in fire server and

then you will use tools to query the

data and create your business

applications so how's it done today I

mean like when you have these equipments

like they're set where they're sending

data where's that data lending and how

do medical staff and and and personnel

interact with their data today

so right now data is very scattered and

that's the reason why we are building on

it right now if you see in IO empty

space there is no clear standard either

but what we are building will allow

answering a question to bring this data

and interoperable way on Azure and Azure

is against a cure compliant cloud so you

can bring pH I protected health

information data and store it and our

services are fully HIPAA compliant as

well as high trust and sacto compliant

as well

I see so so basically you get all the

benefits of that as your compliance with

these requirements and the other thing

you get as well I guess is the ability

as you're building an infrastructure or

a solution to aggregate different

sources of data different types of

devices and offer to your customers

which would be in a hospital or you know

a healthcare facility or something an

actual you know one solution versus a

scattered set of the apps and things to

maintain and so forth right

exactly and on top of it since we are

using fire standard you can bring EMR

electronic medical record data along

with devices data you can bring lab data

we are also looking at retail data we

are also looking at social determinants

of health data as well and bring all of

them with interoperability with the goal

to create a longitudinal patient record

so that you can use that data and create

a I machine learning model for improving

healthcare with mobility security and

privacy at the core of all of that so

for the demo we will use IOT Center for

simulating the device we will go to

github and deploy io empty connector we

will use as your API for fire and then

we will use postman to query the data

and see the data in the fire store so

let's go to Azure portal so here I'm at

the azure portal I go as your API for

fire and I add and create as your API

for fire server I have already created

with the basic configuration so you can

see I have a fire server created the

important thing is the endpoint that we

will use it when we deploy IO empty

connector so that's the first step we

have fire server now okay in the second

step I go to get up this is the IMT fire

github repository here I look at the end

of the presentation for this one for

people to find it sure and here you have

we have rich document about the

architecture how to use it how to debug

it

for easy deployment you just go to

getting started and you click deploy to

Azure when you deploy to Azure it will

bring you to basic configuration screen

where you can choose resource group you

can put service name the important thing

is to use your fire server URL so the

fire server that we created in the

previous step I've already saved it here

you remove slash metadata here you just

need the basic URL okay so I om t

connector is deployed after the

deployment what we need to do is we need

to take the identity of Io empty

connector and save it in fire so that

IMD connector can write data to fire sir

security thank you why we are sure that

we know the source of data is recognized

exactly so I come to fire server go to

authentication and add the managed

identity of the connector here so now we

have server we have connector which is

ready and configured to talk to each

other now devices so for the demo we

will use IOT central for simulating

devices so I can went to IOT central

created an application here you can

create custom application or you choose

pre create a template yes so we have the

new templates for healthcare exactly so

and that's what I'm going to use so we

have a template continuous patient

monitoring I'll choose it and create it

when I create this application it

creates it comes with a dashboard as

well as to pre created devices for my

demo we are going to use smart vitals

patch so I'll go to device template and

for those we're not familiar with azure

IOT in general azure RT central is that

quote-unquote not turnkey solution but

it's what we call a a solution platform

so you very rapidly get to work with

simulators and that's what you're

showing right now okay that will send

the same kind of data an actual device

would send yes and there's a way of

describing it device capabilities in IOT

central that device will be compliant

with and will be able to send data to

your azure IOT central then then you're

gonna consume in exactly

will consume it so if we use here is

smart vitals patch you can see it had

telemetry a biometric data in PH i did i

was talking about about my health

information for example heart rate

respiratory rate it also has data for

device monitoring I got device battery

device temperature device firmware

version for medical and this scenario we

will focus on the biometric or telemetry

data so let's see that data as an

example in JSON format so if you see

this is an example of the data coming

from a smart vitals patch you see heart

rate respiratory rate and other data for

my yeah once you have data now we need

to create two templates so that this

data can be extracted and mapped to open

fire standard okay for in the same I

have gone to the same IMT fire

repository the github repo here we have

sample template which will give you a

good example of how this template looks

like as well as there is a detailed

document on how to create a new template

for your own devices so these are two

template one for device one for fire

mapping so if you see device 1 I am

picking up heart rate and other

information or extracting it from that's

the way you normalize the data that's

coming from different devices you know

which device is the data format and then

you normalize to something that is fire

compliant basically exactly and so when

we are creating it we were debating

should we create for a few devices or

should we create it in such a way that

you can bring any device so the way it

is now you can bring any device as long

as you have the template yeah you can

pick it up will normalize it once again

it's an open source project people want

to contribute these templates these

converters for data normalization they

can definitely contribute to the project

exactly and as we move we also plan to

contribute and create for like very

common variables like fitbit Carmen

create those templates here so that

people can get started

so we have those two templates you take

these two templates save it into the

storage account for the IMT connector so

now if you see and to end we have fire

server we have the connector and we have

seen

device I go to the simulator device

again the IOT central application and I

export this data to IMT connector IMT

connector is listening on event hub okay

so I go to event up I select I am D

connector I have already created one

here and I choose the device data as I

said I'm interested in telemetric the

heart rate respiratory rate bhi data i

save it nice so easy is that exactly and

you have we have not device connected

and fire server and the data is flowing

through yes

now with this let's go to now postman

and query as your api of fire and CD

data especially hard right data so

postman here so I come to postman here

and I am querying again open API

standard and querying the heart rate

with the line code okay and I send it

and it comes back and as I scroll I can

see the heart rate here and I can see

the data value for the heart rate here

as well awesome in a matter of minutes

you actually built an actual you know

fire compliant solution obviously behind

that postman example you would have an

actual application was a dashboarding

with the venting workflows and so on as

you had in your diagram at the beginning

exactly or you can take data from fire

and you can do a I machine learning to

again go into scenarios like telehealth

remote patient monitoring and the key

here that you were saying is the

security and scalability which azure IOT

platform gives you start with a few

devices and you can scale it to hundreds

and millions of devices and store it

with privacy on as you're awesome

that that's fantastic work you guys are

doing leveraging the azure IOT

technology adding on top of that you're

part of months of research so you need

input and feedback so what's your ask to

our audience today

so the ask is first let go and check out

the github repository it's open source I

tried out extend it

ad templates firebugs ad code base and

give us feedback we are very excited to

see the work that we are doing how it's

helping the community to improve the

healthcare awesome so there's a short

link to get to that hit up repo if you

don't catch it in the video is aka dot

ms / IOT show / iom t-connector as IO

medical things in terms of medical

things connector thanks rash me for

joining us on the show today hope to see

you soon and the work your guys are

doing you know on another episode thanks

everyone for watching don't forget to

subscribe and see you soon

[Music]