AWS for Industries

Fast-tracking the Healthcare AI Roadmap: AWS Health Data Accelerator

Many customers struggle to securely ingest, process, govern, and analyze sensitive healthcare data while adhering to vendor and regulatory requirements including HIPAA. In this blog, we introduce a technical offering that enables customers to efficiently solve these challenges.

Solving the implementation challenges of healthcare data strategy

Many healthcare organizations find it difficult to implement their healthcare data strategy and vision due to:

  1. Complexity in securely ingesting sensitive electronic health record (EHR) data while adhering to vendor and regulatory requirements including HIPAA.
  2. Lack of familiarity in architecting cloud data platforms and migrating from on-premises infrastructure.
  3. Shortage of cloud data engineering skills to build scalable, production-grade pipelines.
  4. Difficulty integrating diverse healthcare data sources into a unified platform.
  5. Insufficient know-how on implementing cloud data governance, security and access controls.

To address these challenges and meet strong customer demand, we created the Amazon Web Services (AWS) Health Data Accelerator. This offering makes it straightforward to implement a health data strategy as quick as in 12 weeks—accelerating speed for execution and scalability.

The AWS Health Data Accelerator

The AWS Health Data Accelerator offering enables healthcare organizations to create a secure platform on AWS that supports multi-modal data and healthcare analytics. The Health Data Accelerator offering provides a best-practice reference architecture and prescriptive guidance to optimally configure AWS services for healthcare analytics.

With automated deployment of foundational components (such as secure EHR data ingestion pipelines) the Health Data Accelerator allows customers to adopt state-of-the-art analytics quickly. The reference architecture incorporates essential AWS services for data ingestion, processing, storage, governance, and analytics including machine learning and generative AI.

The AWS Health Data Accelerator offering (Figure 1) makes it possible to securely ingest various data sources including EHR, omics, medical imaging, digital pathology, medical audio recording, and third-party data. It is built on top of the Landing Zone Accelerator for Healthcare, providing security and compliance best practices.

The offering includes a data ingestion process to automatically load EHR data from on-premises data servers into an AWS data lake. It leverages Amazon Simple Storage Service (Amazon S3), handling both historical data migration as well as incremental updates on an ongoing basis. Additionally, Amazon Redshift is deployed as a data warehouse for processed data and can integrate other data sources beyond EHR.

Depending on the data modality, specialized services such as AWS HealthLake, AWS HealthOmics, AWS HealthImaging, and AWS HealthScribe can be optionally deployed. AWS Glue enables data cataloging, integration and harmonization across different data sources. Amazon DataZone natively integrates with AWS Glue to provide an interactive data catalog for end-user data discovery, access control and data lineage tracking, enabling data governance.

Data consumers such as data scientists and administrators can act on the data and develop analytic applications by using Amazon SageMaker AI for machine learning and Amazon QuickSight for business intelligence. Amazon Bedrock can power clinical applications with generative AI capabilities for providers, researchers, and patients.

Some customers are looking to expand their data consumers to include third-party entities outside of their organization. AWS Data Exchange enables external data user access, and AWS Clean Rooms enables collaboration across different organizations.

High-level overview showing the architecture components. Starts with Data Sources on the left (containing EHR, omics data, medical imaging, digital pathology, medical audio recording, and third-party data). Data Sources points to Store, Query, and Analyze components on the right (containing Amazon S3, Amazon Redshift, optional AWS HealthLake, optional AWS HealthOmics, optional AWS HealthImaging, and optional AWS HealthScribe). Store, Query, and Analyze points to Integrate on the right (containing AWS Glue to harmonize, optional AWS Data Exchange for third-party data access, and optional AWS Clean Rooms for data collaboration). Integrate points to Act on the right (containing Amazon SageMaker AI for machine learning, Amazon QuickSight for business intelligence, and Amazon Bedrock for generative AI). SageMaker AI and QuickSight can power analytic applications serving scientists and administrators as end users. Amazon Bedrock can power clinical applications to serve providers, researchers, and patients. AWS Data Exchange and AWS Clean Rooms can serve third-party data consumers. At the bottom, Amazon DataZone spans across the architecture component areas for data catalog and governance.

Figure 1 – High-level architecture overview

During a Health Data Accelerator engagement, the AWS team provides Infrastructure-as-Code solutions along with runbooks and documentation to the customer. The knowledge transfer and training enable customers to scale the solution quickly with additional data sources. Leveraging the Health Data Accelerator, customers are able to launch a wide range of generative AI applications tailored to their business goals and data (Figure 2).

Design cohorts and find clinical trials - Use a chatbot for patient cohort design and analysis. Find relevant clinical trial studies for patient enrollment. Automate data engineering - Ingest unstructured clinical reports and forms to extract medical concepts and generate structured data to expedite patient transfer requests and tumor board reviews. Provide clinical decision support - Use a chatbot that references pharmacogenetics clinical guidelines to help clinicians interpret genetic laboratory test results and prescribe appropriate medications and drug dosage. Democratize data access - Democratize data access, analytics and data visualizations to end-users who are not familiar with programming by using natural language queries. Understand biomedical literature - Use a chatbot to answer clinician and researcher questions based on medical publication and scientific literature data.

Figure 2 – Example of generative AI applications using AWS Health Data Accelerator

Customer success story: Electronic Caregiver’s journey with AWS Health Data Accelerator

Electronic Caregiver Inc. (ECG) is a leading digital health technology and services company, headquartered in New Mexico, USA. ECG’s mission is to design and deliver innovative, impactful telehealth products and services. They bridge the gap between the provider’s office and the patient’s home to improve clinical outcomes, expand medical access, optimize resource allocation, and deliver value-based care.

ECG partnered with AWS Professional Services and implemented the AWS Health Data Accelerator offering, creating a scalable data and analytics platform on AWS for their EHR, patient surveys and medical IoT data. The Health Data Accelerator offering enabled ECG to accelerate data ingestion of over two million patient records onto their AWS Cloud.

To leverage this healthcare data, Amazon QuickSight dashboards were created for data governance monitoring, financial forecasting based on insurance reimbursements, and patient outreach program analytics. Moreover, ECG took advantage of Amazon Q in QuickSight to quickly generate key insights and executive summaries by using natural language prompts.

Within one week of the Health Data Accelerator implementation, over 20 percent of the entire ECG organization (including senior executives and team members without coding skills) were able to directly interact with the rich patient data. This helped ECG to understand patient compliance with the remote monitoring services more efficiently, because the generative AI capability removed a dependence on internal analytics teams with programming skills.

ECG’s remote monitoring technologies are deployed nationwide at over 100 clinics. Mark Francis, the Chief Product Officer at ECG, recognized the need to demonstrate utilization and impact for their telehealth services across the country. The US federal COVID-19 Public Health Emergency program execution and the industry trend of value-based care intensified this need. Francis took a step further and sought to quantify the business impact brought by the Health Data Accelerator offering. Understanding the level of patient interaction with remote monitoring devices is critical to drive clinical outcome improvements and insurance reimbursements.

To this end, the ECG team analyzed the physiological data captured by the devices (for example, blood pressure, glucose levels, and weight) and looked for clinical anomalies and acute episodes. The analysis revealed where to focus their intervention programs to improve patient compliance and clinical outcomes, leading to an 18 percent increase in insurance reimbursement.

“The data and insights obtained through the AWS Health Data Accelerator offering enables healthcare providers, hospital systems, and payors to proactively manage healthcare services, manage resource allocation, and optimize outcomes. For telehealth, this means expanded access to care, as well as better care at home,” says Francis. “Partnering with AWS Professional Services enabled us to move efficiently into production with our data insights and analytics initiatives. The AWS Health Data Accelerator implementation was an excellent showcase of clear project definition and technical discovery, open collaboration, and rapid and iterative development work.”

ECG plans to expand their use of the AWS Health Data Accelerator offering to support chronic care management, transitional care management, and hospital-at-home use cases.

Conclusion

By addressing common challenges in healthcare data analytics, the AWS Health Data Accelerator empowers healthcare organizations to focus on driving actionable insights. This offering makes it quicker and smoother than ever for healthcare leaders to capitalize on the power of the AWS Cloud and create meaningful improvements in patient outcomes and operational efficiency.

The customer story of ECG also highlights how effective and efficient the AWS Health Data Accelerator can be for companies looking to expand their data insights. Through natural language queries, using machine learning and generative AI services, all levels of the company are able to engage insights without the need for coding. This new ability to experiment and build transformative experiences with valuable insight will help ECG to stay on the leading edge.

We encourage any healthcare organization interested in accelerating their analytics and AI capabilities to reach out to their AWS account team or the AWS Healthcare and Life Sciences sales team.

Further Reading

Jenna Eun

Jenna Eun

Jenna Eun is a Principal Practice Manager for the Health and Advanced Compute team at AWS Professional Services. Her team focuses on designing and delivering healthcare and life sciences data and machine learning solutions for the public sector, including federal, state and local governments, academic medical centers, nonprofit healthcare organizations and research institutions.

Bryan Marsh

Bryan Marsh

Bryan Marsh is a Principal Solutions Architect in the Academic Medical Center team at AWS. He has expertise in enterprise architecture with a focus in the healthcare domain. He is passionate about using technology to improve the healthcare experience and patient outcomes.

Hinel Karia

Hinel Karia

Hinel Karia is an Executive Healthcare Adviser at AWS. She is passionate about driving data modernization and AI/ML adoption in healthcare. She collaborates closely with healthcare executives, guiding them through their digital transformation journeys to enhance patient care, streamline operations, and accelerate medical research.

Mark Francis

Mark Francis

Mark Francis is Chief Product Officer at ECG. Mark’s expertise is in developing, deploying, and scaling AI, ML, and data-infused solutions to improve outcomes and expand access to care in the fields of health and aging.

James Wiggins

James Wiggins

James Wiggins is a Senior Solutions Architect Manager for Global Healthcare at AWS. He is passionate about using technology to help organizations positively impact world health. He also loves spending time with his wife and three children.