Transforming mHealth applications in Asia Pacific Japan using AWS Outposts

Healthcare organizations are seeking to leverage advanced technologies like Internet of Things (IoT), big data analytics, and cloud computing to improve patient outcomes and increase efficiency. However, stringent data privacy regulations, data sovereignty considerations, and the need for low latency access to critical applications present challenges for adoption. Monitoring devices utilized in mobile health (mHealth) applications often face latency challenges and data residency considerations when relying solely on centralized cloud solutions. AWS Outposts offers a hybrid cloud approach, bringing AWS infrastructure on-premises to enable low latency access to services while maintaining data sovereignty.

AWS Outposts allows organizations to run AWS infrastructure and services on-premises, bringing the agility, breadth of services, and pace of innovation of the AWS cloud to virtually any data centre, co-location space, or on-premises facility. As Dr. Werner Vogels noted in his 2022 technology predictions, Outposts is a key enabler of what he calls the “everywhere cloud,” allowing AWS’s infrastructure, services, APIs and tools to be deployed in customer-controlled environments while being fully managed by AWS.

The proliferation of mobile health (mHealth) applications and Internet of Things (IoT) devices are transforming healthcare delivery. mHealth apps on smartphones and wearable sensors enable continuous remote monitoring of vital signs, chronic medical conditions, fitness data and more. This leads to better health outcomes through timely interventions. However, analysing huge volumes of real-time, heterogeneous and noisy physiological data poses computational and privacy challenges.

To address these concerns, hospitals are turning to on-premise edge computing solutions hosted within their premises. These edge solutions aggregate data from various sources within the hospital, ensuring data privacy and compliance with various health regulations.  Additionally, AI-enabled edge solutions facilitate real-time data processing and offer clinicians timely notifications about abnormal patient trends and behaviours, enhancing hospital productivity and resource utilization while ensuring patient data security. AWS Hybrid cloud services can address these needs while enabling healthcare systems to harness cloud innovation.

Hybrid cloud computing combines on-premises infrastructure with public cloud to enable data processing, analytics, storage, and workloads to span edge devices, on-site data centres, and the cloud. This allows optimization across metrics like security, latency, throughput, and cost. For healthcare, applications areas include:

• Real-time collection and analysis of vital signs, telemetry, genomic data
• Edge analytics on imaging data to enable time-sensitive diagnosis
• Local processing of patient privacy information with cloud analytics on de-identified data

Applicability of AWS Outposts in Mobile healthcare scenarios

Latency Sensitivity: Real-time clinical decision making relies on continuous physiological data streams from medical devices and wearables monitoring patient vital signs or medical asset tracking. To meet sub-second latency for arrhythmia detection or falls prevention, initial processing and alerts must occur locally at the edge before streaming to cloud analytics. With AWS Outposts, healthcare providers can process critical data locally, reducing latency and ensuring timely response for interactive mHealth applications. By leveraging robust connectivity from Outposts back to the parent AWS Region, we seamlessly integrate with cloud-based analytics platforms and AI/ML services.

Data Sovereignty and Compliance: Regulations like Health Insurance Portability and Accountability Act (HIPAA) mandates data security and privacy protections when dealing with electronic protected health information (ePHI). Local edge infrastructure allows ePHI and other sensitive data to be stored and processed on-premises. Only de-identified, aggregated outputs are streamed to the cloud after local processing. AWS Outposts address this challenge by enabling hospitals to maintain control over data locality while leveraging cloud services for computation and analytics. By hosting mHealth data on Outposts, healthcare organizations can comply with regional privacy laws and regulations and offer local data processing and low latency compute for critical mHealth applications.

Distributed Data Sources: Healthcare IoT deployments and mobile health use cases mean thousands of medical devices, wearables, and remote assets generating continuous telemetry across multiple geographical sites. AWS Outposts makes it easier to deploy AWS locally for distributed data sources and provides hybrid integration with the public AWS cloud.

Enhanced Security: Security is paramount in healthcare, where patient confidentiality and data integrity are of utmost importance. Outposts inherits the same compliance certifications, audits and security best practices as the public AWS cloud in terms of the infrastructure, software and operational policies. This includes capabilities like encrypted storage, roles-based access control etc.

AWS Outposts architecture for low latency analytics at hospitals

The following diagram shows the architecture of AWS Outpost for low latency analytics at hospitals.

This reference architecture allows healthcare data from edge endpoints like wearables and hospital equipment to be streamed in real-time for analytics while maintaining local control over sensitive patient data. Low latency access to critical workloads is enabled via local data processing. Hospitals can optimize application deployment across edge and cloud while meeting regulatory compliance. The flow is as follows.

1. Wearable and smart devices connect to Mobile applications which leverages AWS IoT Mobile SDK.

2. Devices in pharmacy/clinics leverage Amazon FreeRTOS as their operating system and leverage AWS IoT libraries for secure connectivity.

3. An Amazon EC2 instance on Outposts acts as an edge gateway. It runs AWS IoT Greengrass and AWS IoT SiteWise Edge to facilitate connectivity with the hospital network. Local dashboards and custom applications are deployed on AWS IoT Greengrass for monitoring and processing.

4. Clinical applications deployed on AWS Outposts can consume data locally from AWS IoT Greengrass.

5. EMR on Outposts can be configured to process and analyse data at the edge. This is essential for scenarios like continuous monitoring of patient vital signs or immediate detection of abnormal trends, enabling timely clinical interventions. EMR on Outposts processes the incoming data to extract relevant observations and convert the data into FHIR format. Batches of FHIR observation data are stored in Amazon S3 on Outposts.

6. All the telemetry data from the Outposts resources are trasnmitted to Amazon CloudWatch and Amazon CloudTrail in the AWS Region to which the Outposts is anchored for auditing and monitoring purposes. Besides the individual resource level capacity CloudWatch metrics, CapacityExceptionsAlerts are also populated and detailed in CloudWatch metrics for AWS Outposts.

Key Design considerations

Connectivity Reliability: Robust connectivity is essential for the successful deployment and operation of AWS Outposts. AWS Outposts extend the AWS infrastructure, including services such as Amazon EC2 and Amazon S3, to on-premises locations. To ensure optimal performance and reliability, a high-quality network connection is crucial.

Data Governance: Data governance is a critical aspect of managing and protecting data throughout its lifecycle. When utilizing AWS Outposts, which extends AWS services to on-premises locations, it’s essential to establish robust data governance practices. Some of the best practices include understanding regulatory requirements, encryption. access control, metadata tagging, auditing and monitoring.

Scalability: Auto-scaling capabilities for edge compute, storage, and analytics are needed to dynamically serve potentially thousands of streaming medical devices per hospital or clinic. Ensure that the provisioned resources align with your application’s requirements and can handle peak loads. This may involve periodically reviewing and adjusting resource allocations based on changing needs.

Conclusion

AWS Outposts provides a useful solution for deploying mobile health (mHealth) applications in areas across Asia Pacific that lack proximity to an AWS region. Also, by establishing local computing infrastructure, hospitals can overcome latency challenges, ensure compliance with data residency regulations, and enhance data security while leveraging the scalability and innovation of AWS cloud services.

To learn more about AWS Outposts, including information on common use cases and deployment practices, review the AWS Outposts documentation, the AWS Outposts User Guide and the AWS Outposts Data Residency eBook.