It is no secret that the U.S. healthcare system is incredibly complex, cost-ineffective, and esoteric. The three core entities of the system which are the patient, provider, and payer are locked in a losing cycle of balancing providing the highest quality care possible while managing costs. The patient is caught in a cycle of rising premiums and decreasing access to quality care. The providers are frantically attempting to manage costs in an ever-changing legal and economic landscape. The payers are acting as the fiduciaries to regulate cost while playing to the patient’s needs through the lens of cost management. Nobody is winning except perhaps the shareholders of the publicly traded entities.
The situation
The healthcare system requires a reset. Advances in technology can help to facilitate that reset. There are numerous areas where technology can help drive healthcare change, from implantable medical devices to remote surgical techniques. As a simple starting point, data standardization and data interoperability can help to drive cost efficiencies and improve patient care.
For decades, payers and providers have struggled with data exchange protocols. Data exchange has often taken the form of document exchange, adding layers of unneeded complexity to translating data files and manual processes. Organizations often have multiple data ‘sources of truth’, and the focus is on data processing instead of insights and innovations. The introduction of FHIR, an HL7 standard for exchanging healthcare information electronically, was a way to address these issues. By uniting and standardizing clinical and claims data under a common format, FHIR provides a foundation for data interoperability and data analytics and is driving advancements in broad-based initiatives across the healthcare landscape, including implementing value-based care, increasing patient safety, improving operational efficiency, and enhancing care coordination.
FHIR provides several benefits and improvements as a modern healthcare standard, including facilitating interoperable exchange with legacy standards, lower overhead, shorter learning curve, an ability to transmit only the necessary pieces of information, potential for patient-mediated data, and an energized community of supporters and implementers.
In recent years, CMS has pushed the adoption of the FHIR format through a series of mandates. However, organizational acceptance has been slow and spotty. Often, the barrier to adoption by institutions is the required time and financial investment to migrate EHR data to FHIR and build the analytics framework to realize the value. The slow pace of data acquisition and integration to drive adoption weighs on organizations’ already strained resources.
Speed to insights
Amazon HealthLake has helped to address the concerns about adoption by providing a scalable architecture that allows organizations to focus on the differentiated problems of activating data to support various use cases around interoperability, healthcare analytics, and patient insights. The Amazon HealthLake service helps providers and payers gain insight into their members, patients, and operations within days instead of months. The service enables granular data access and governance to internal teams as well as 3rd party integrations to accelerate innovation while maintaining security controls.
For organizations that need to store and manipulate healthcare data from different sources, Amazon HealthLake is a standardized healthcare data layer that improves data standardization and portability of clinical data. By uniting and standardizing clinical and claims data under a common format, Amazon HealthLake facilitates rapid data acquisition and integration from diverse sources via a scalable architecture and allows organizations to focus on the differentiated problems of activating that data to support various use cases around interoperability, healthcare analytics, and patient insights.
Turning insights into actions
The use cases of Amazon HealthLake span the care continuum—from preventative care through machine learning to easy analysis of historical patient/member cohorts for driving future workflows.
Pariveda explored one example in partnership with AWS: developing an AI/ML model to predict chronic congestive heart failure. The model’s objective was to predict the likelihood of a patient receiving a chronic congestive heart failure diagnosis in the next three clinical encounters.
The team leveraged Amazon HealthLake to ingest and store EHR data in FHIR format. The team defined the specific patient cohorts through a set of Amazon HealthLake APIs leveraging FHIR elements such as demographics, encounter details, and diagnosis codes. The resulting cohort data was fed to Amazon SageMaker to train ML models for scenario-driven predictive analytics and used for building business intelligence dashboards in Amazon QuickSight.
Below are some screenshots of the development process. If you’d like more details, we’d be happy to discuss further.
Snapshot of the CloudFormation template
Snapshot of the data domains
Example of the direct query into Amazon QuickSight
Snapshot of an early version of the Data Wrangler data flow
The results from the model provide insights that enable and drive real actions. For example:
- In a provider scenario, the rendering provider would receive an alert for a high-risk patient with a corresponding action to adjust their care plan and follow-ups accordingly.
- In a payer scenario, a program could be implemented to engage discharged cardiac patients and proactively manage care coordination activities for the patient to avoid adverse health events.
- Advanced analytics can provide insights into defining the “Next Best Action” for the patient based on continuous model training.
Conclusion
Amazon HealthLake provides a foundation to drive change across the healthcare landscape by turning data into actions quickly. Through its HIPAA eligibility, powerful query and search capabilities, integrated medical natural language processing, and ability to define patient cohorts for quick analysis, Amazon HealthLake provides payers, providers, and life sciences a foundation for data interoperability, data sharing, and big data analysis.
Pariveda is looking forward to our continued partnership with AWS and engaging clients to develop solutions based on Amazon HealthLake’s ability to drive actionable innovation and improvement initiatives.
