Health data are normally found incomplete and inconsistent, and is always unstructured, with records contained in clinical records, laboratory reports, insurance claims, medical images, recorded conversations, and time-series data across disparate formats and systems. For examples heart ECG or brain EEG traces. Each medical care supplier, payer, and life sciences organization is attempting to take care of the issue of structuring the data.
What is an Amazon health lake:
Amazon HealthLake provide healthcare organizations to store, transform and analyse all of their data information in the cloud
It is HIPAA-qualified assistance empowering medical care and life sciences organizations to safely store, transform, query, and examine health data at scale. There are no extra charges for the storage, FHIR questions, coordinated NLP, and FHIR information trade.
Amazon HealthLake helps to avoid putting the hard work of getting sorted out, organizing, indexing, and structuring patient data, to give a total perspective on the health of individual patients and also the whole patient population in a secure, compliant, and auditable manner.
Using the HealthLake APIs, medical care associations can store, health data on the Fast Healthcare Interoperability Resources (FHIR) industry-standard organization from on-premises frameworks to a protected data lake in the cloud.
HealthLake changes unstructured information utilizing particular AI models, similar to regular language preparation, to consequently separate significant clinical data from the information and gives incredible inquiry and search abilities. Associations can utilize progressed examination and ML administrations, like Amazon QuickSight and Amazon SageMaker to break down and get connections, distinguish patterns, and make expectations from the recently standardized and organized information. From early recognition of infection to population health patterns, associations can utilize Amazon HealthLake to direct clinical information examination controlled by AI to further develop minds and lessen costs.
Quickly and easily import your health data, including clinical notes, insurance claims and more
Store it on the AWS Cloud in a secure complaint, and auditable way.
Tag and index unstructured data using specialized machine learning (ML) models
- QUERY and SEARCH:
Query and search the transformed data across all of your health data or an individual patient record
Understand relationships in the data and identify trends, and make predictions with integrated analytics and ML capabilities
Overview of how it works
Amazon HealthLake is supported by a completely managed AWS infrastructure. You will not need to obtain, arrange, or manage a single piece of IT equipment. You should simply create a new data store, which just requires a couple of moments. When the datastore is prepared, you can immediately create, read, update, delete, and query your data. HealthLake discover a simple REST Application Programming Interface (API) available in the popular languages in which customers and partners will be able to easily integrate into their business applications.
HealthLake upholds both structured and unstructured data details ordinarily found in clinical notes, lab reports, insurance claims, etc. The assistance stores this information in the Fast Healthcare Interoperability Resource (FHIR, said as well as ‘fire’) design, a standard intended to empower the trade of health data information. HealthLake is compatible with the most recent version (R4) and right now upholds 71 FHIR asset types, with extra resources to follow.
If the data is already in FIHR format, then it’s perfect, if not, you will be able to reconstruct it with the help of the AWS platform having validated connectors for HealthLX, Redox, Diameter Health, and InterSystems applications. Also helps to convert your HL7v2, CCDA, and flat file data to FHIR, and to upload it to HealthLake easily.
Once the data is uploaded, Healthlake will use an integrated natural language process to extract the entities contained in the documents and store the appropriate metadata. All these include anatomy, medical conditions, medication, protected health information, tests, treatments and procedures. It is matched with the industry-standard ICD-10-CM and RxNorm entities.
After the data has been uploaded, we can start querying it, by setting the parameter values to FHIR resources and extracted entities. If we want to access the information on a single patient and need to export many documents to construct a research dataset. it will only take a single API call
Querying FHIR Data in Amazon HealthLake
Open the AWS console for HealthLake, and click on ‘Create a Data Store’. Then, at that point, I just pick a name for the Database store, and choose to encrypt it with an AWS Managed key. Tick the option in box that preloads sample synthetic data, which is a great way to quickly kick the tires of the service without having to upload my own data.
Sample mentioned below
Following a couple of moments, the information store will get active and can send queries to its HTTPS endpoint. As per the example below, search for clinical notes (and clinical notes only) that contain the ICD-CM-10 entity for ‘hypertension’ with a score of 99% or more. In the engine, the AWS console is sending an HTTP GET request to the endpoint. I featured the related inquiry string.
Sample query for reference.