By Maharshi Desai, Global GTM Specialist, AWS
Introduction
The healthcare industry, much like other industries, is witnessing a surge in data growth due to Electronic Health Records (EHR), medical imaging, population sequencing, and claims processing. According to the World Economic Forum, hospitals generate 50 petabytes of info every year, yet 97% goes underutilized for various reasons, such as data sitting in silos, being stored as unstructured and semi-structured data, and more. The pandemic emphasized how important technology is for healthcare organizations, prompting them to accelerate digital transformations for growth and innovation. This article discusses how healthcare organizations can apply the Amazon Web Services (AWS) modern data strategy to become data-driven and advance innovation.
Healthcare Business Challenges
Healthcare professionals, providers, and organizations rely on timely, accurate, and secure data to make informed decisions on patient care, manage reimbursements, and control costs. However, there is no straightforward solution due to the following three data challenges:
- Data governance and compliance: Healthcare is highly regulated industry. Security, privacy, and compliance are of the utmost importance for healthcare organizations. This can be a challenge as data privacy policies and compliance standards vary drastically across the world and can make it difficult for healthcare organizations to adopt and innovate quickly at global scale. A number of countries, including the U.S., also have laws to keep data within their borders, which makes data residency a major factor in defining the cloud strategy. IT resiliency, including robust disaster recovery planning, is critical to this industry as organizations face regular threats due to human-made (e.g., ransomware, server configurations errors) and natural (e.g., hurricane, flood, earthquake) causes.
- Data silos: Despite the widespread usage of EHR, more than one-third of clinicians and care providers still struggle with exchanging information. A key contributor to this challenge is the lack of standardization and integration of data management systems to make access to data easier and faster. In addition, data governance can impact the ability to share information globally due to local policies.
- Quality data access: Without data standardization, IT professionals can unknowingly create multiple data stores with different data values for the same patient records, causing data integrity issues across the systems. IT teams will spend a significant amount of time to manually standardize the data for physicians to use in patient care.
AWS Data Strategy
To help address these healthcare challenges, AWS has developed a modern data strategy to assist customers in becoming data-driven organizations. AWS has defined a comprehensive blueprint to aid businesses in safeguarding, accessing, evaluating, and responding to data. AWS has helped customers, including The Minnesota Department of Health (MDH), Philips, Gilead, Genentech, and 3M, break down data silos and provide direction to ensure the highest level of safety and compliance are met.
So, where to begin? The first step to becoming a data-driven organization is to understand success comes from: 1) treating data as an asset, 2) collaborating and sharing data across the enterprise, 3) providing a federated data governance and compliance model between data producer and data consumer, and 4) operationalizing the data strategy to align with people, process, and tools. This is a shift as data has historically been owned and managed by IT, and business users could not locate the necessary data to make decisions. Through data democratization, anyone with the right permission can gain access to data that is easy to understand and straightforward to use.
Next, to implement data democratization, organizations need to understand the data mesh architecture. The AWS Data Mesh architecture empowers business units (organized into domains) to have high ownership and autonomy for the technologies they use, while providing technology that enforces data security policies both within and between domains through data sharing. The data mesh architecture simplifies and automates the business processes and uses fully managed, purpose-built databases to share the information.
Now, when building an end-to-end data strategy, there are three core pillars: 1) comprehensive set of services 2) integrated data sources, and 3) governance strategy. Figure 1 show how AWS data strategy translates to AWS Health for Data.
![](png/aws%20health%20data.png)
Figure 1: AWS Health for Data Strategy
- Comprehensive set of services: The basis of this pillar is to build a foundation of data capabilities to support any use case or business outcome now and in the future. A data strategy built for tomorrow will allow organizations to adapt to changing needs and new opportunities. To build modern data architectures on the cloud, AWS offers the broadest portfolio of 15-plus purpose-built database engines to support the most demanding workloads at a lower cost compared to on-premise databases. These services offer simplification of data management with enhanced features like automatic failover, automated backup and recovery, database upgrades, isolation and security, industry compliance, automated patching, advanced monitoring, routine maintenance, and seamless scaling, in-region and cross-region to name a few. AWS services, including databases, are certified for HITRUST CSF, HIPAA, GDPR, FedRAMP, ISO 27001, ISO 3425, and more to help healthcare organizations adhere to industry compliance and regulations.
In the healthcare industry, the EHR system is surrounded by a number of clinical systems and they exchange the data across the systems to work effectively. Organizations traditionally have used relational databases for every solution, which includes structured and unstructured data. Storing unstructured data, such as medical images and claim images, to relational databases as a blob of objects is very expensive and also difficult to analyze, contributing to 97% of data being underutilized. By moving medical images to object storage and into a non-relational key-value databases, like Amazon DynamoDB, will improve accessibility and performance at scale, whereas drug discovery and GxP compliance may be better off using non-relational graph databases, like Amazon Neptune. For organizations that have outgrown their FAIR (Findable, Accessible, Interoperable, and Reusable) standards, like Genentech, consider consolidating and transferring your existing data to a managed, centralized searchable repository. Genentech achieved the scalability and efficiency needed by using AWS services, such as Amazon Simple Storage Solution (Amazon S3) to store data, Amazon Relational Database Service (Amazon RDS) to handle the associated metadata, and Amazon OpenSearch Service to index and quickly search.
- Integrated data sources: This pillar represents building integrated data solutions across the organization, so stakeholders can easily access it, no matter where it resides. Creating a holistic view of the organization's data sources simplifies access and analysis. This view may be multi-dimensional and cross-domain, and the goal is to analyze the data to create a meaningful outcome. AWS provides services, such as Amazon Aurora, Amazon Redshift, AWS Lake Formation, AWS Glue, Amazon S3, and Amazon Athena, to create a data mesh architecture. With a modern data architecture on AWS, organizations can efficiently: 1) build scalable data lakes, 2) ensure compliance via unified data access, security, and governance, and 3) scale their systems at a low cost without compromising performance.
For example, global bio-pharmaceutical company, Gilead, adopted a data mesh approach to improve agility, accelerate insight generation, and increase its return on investment. According to Marc Berson, chief information officer at Gilead, “The primary reason that we chose AWS was its passion for innovative transformation. With AWS, we have developed an enterprise data solution to create better access to, and analysis of, data across the organization using a data mesh approach.” Outside of the data mesh, Gilead used AWS Data Exchange, which makes it simple to find, subscribe to, and use third-party data in the cloud, to provide massive data transfer speeds for its data marketplace.
In addition to AWS Data Exchange, AWS provides a number of services designed for integrating healthcare specific data. Amazon HealthLake is a HIPAA-eligible service that provides FHIR APIs that help healthcare and life sciences companies securely store, transform, transact, and analyze health data in minutes for a chronological view at the patient and population level. Amazon Omics helps organizations store, query, analyze, and generate insights from genomic, transcriptomic, and other omics data. In late 2022, AWS also announced Amazon Aurora Zero-ETL to enable near-real time analytics and machine learning (ML) using Amazon Redshift. With Amazon Aurora zero-ETL, transactional data is automatically and continuously replicated seconds after it is written into Amazon Aurora and seamlessly made available in Amazon Redshift, where data analysis can begin immediately. To help with cross-organizational data sharing, AWS Clean Rooms allows organizations and their partners to more easily and securely collaborate and analyze their collective datasets—without sharing or copying one another’s underlying data.
AWS is also improving data quality and simplifying ML implementation with scalable infrastructure, tools, services, and industry ML models. Some common use cases addressed by purpose-built artificial intelligence (AI) services include: 1) contact center intelligence, 2) medical context understanding (such as CD-10-CM, RxNorm, and SNOMED CT) with advanced text analytics and natural language processing using Amazon Comprehend Medical, 3) medical speech to text conversion using Amazon Transcribe Medical, 4) automatically extract printed text, handwriting, and data from any document using Amazon Textract, 5) online fraud detection, 6) business metrics analysis, and more.
- Governance strategy: This pillar applies to the overall governance of the data architecture so that teams can operate effectively. Governance protects patient privacy with data policies. AWS enables organizations to establish the right governance—one that balances control and access—to empower users to access data when and where they need it. For example, Amazon DataZone is a data management service that enables you to catalog, discover, govern, share, and analyze your data across accounts and supported regions. Amazon DataZone simplifies the experience across AWS services including Amazon Redshift, Amazon Athena, AWS Glue, and AWS Lake Formation. It will collaborate on data projects through a unified data analytics portal that gives a personalized view of all the data while enforcing governance and compliance policies.
When Philips built Philips HealthSuite, a secure and HIPAA-compliant digital cloud infrastructure on AWS, the intent was to make it simpler for oncology teams to deliver personalized therapy to patients by integrating genomic data with other modalities like imaging, digital pathology, and clinical data. It was fundamental to have security controls that ensured HIPAA compliance to support the hybrid, multi-tenant system connecting experts across centers and with near real-time updates.
Next Steps
Data-driven organizations consider data a valuable asset, enable governed access to integrated data sources, and use it to make more informed decisions. By adopting a modern data strategy, healthcare organizations can unlock the underutilized data to increase clinical and operational effectiveness, advance insights and discoveries, and improve patient care. AWS can help organizations through their end-to-end modern data strategy with a comprehensive set of purpose-built databases, advanced analytics, and machine learning services. Let’s start building together and focus on what really matters—the patients.
For more information, go to AWS Health for Data or contact sales.