At the 2020 Amazon Cloud Technology re:Invent conference, we previewed Amazon HealthLake, which is a fully managed and HIPAA-compliant service. Healthcare and life sciences customers can use this service to aggregate health information from different silos and different formats into a structured centralized Amazon cloud technology data lake, and obtain insights from these data through analysis and machine learning (ML). Recently, Amazon Cloud Technology is very pleased to announce that Amazon HealthLake is officially released for all Amazon Cloud Technology customers.
- Amazon HealthLake：
ability to quickly store, transform and analyze health data of any scale is essential for making sound health decisions. In daily practice, doctors need to follow a chronological view of the patient’s medical history to determine the best treatment plan. In the event of an emergency, providing the medical team with appropriate information at the appropriate time can significantly improve the patient's treatment effect. Similarly, health care and life science researchers also need high-quality standardized data to analyze and build models to determine population health trends or drug test recipients.
Traditionally, most health data is locked in unstructured text such as clinical notes and stored in IT silos. Heterogeneous applications, infrastructure, and data formats make it difficult for practitioners to access patient data and gain insights from it. We built Amazon HealthLake to solve this problem.
If you can't wait to start using the service, you can immediately jump to Amazon HealthLake's Amazon Cloud Technology Console . If you want to know more, please keep reading!
- Amazon Cloud Technology Console:
📢 To learn more about the latest technology releases and practical innovations of Amazon Cloud Technology, stay tuned to the 2021 Amazon Cloud Technology China Summit! Click on the picture to register now~
grand launch of Amazon HealthLake
Amazon HealthLake supported by a fully managed Amazon cloud technology infrastructure. You don't have to purchase, provision, or manage a piece of IT equipment. only needs to create a new data store, and this only takes a few minutes. After the data storage is ready, you can immediately create, read, update, delete, and query data. Amazon HealthLake exposes a simple REST application programming interface (API), which is provided in most commonly used language . Customers and partners can easily integrate it into their business applications.
- Amazon HealthLake：
- The most commonly used language
Ensuring security is the top priority of Amazon Cloud Technology. By default, Amazon HealthLake uses Amazon Key Management Service (KMS) to encrypt data at rest. You can use the key managed by Amazon Cloud Technology, or you can use your own key. including Amazon Cloud Technology employees, from retrieving your plaintext key from the service. For data in transit, Amazon HealthLake uses industry standard TLS 1.2 end-to-end encryption.
- Amazon Key Management Service (KMS)
At launch, Amazon HealthLake supports structured and unstructured text data, which can usually be found in clinical notes, laboratory reports, insurance claims, etc. The service Fast Healthcare Interoperability Resource (FHIR, pronounced "fire") format, which is a standard designed to support the exchange of health data. Amazon HealthLake is compatible with the latest revision (R4). Currently, supports 71 FHIR resource types , and more resources will be supported in the future.
- Rapid healthcare interoperability resources
- 71 FHIR resource types
If your data is already in FHIR format, that's great! If you have not adopted this format, you can do the conversion , or you can use the partner solutions provided in 161cda652a325b Amazon Marketplace At launch, Amazon HealthLake includes connectors that have been validated Redox, HealthLX, Diameter Health, and InterSystems They can easily convert HL7v2, CCDA, and flat file data to FHIR format, and then upload it to Amazon HealthLake.
- Amazon Marketplace
- Diameter Health
When uploading data, Amazon HealthLake uses integrated natural language processing to extract entities present in the document and store the corresponding metadata. These entities include anatomy, medical conditions, medications, protected health information, tests, treatments, and procedures. They also match the industry standard ICD-10-CM and RxNorm entities.
After uploading the data, you can start querying the data by assigning parameter values to FHIR resources and extracted entities. Whether you need to access a patient's information or want to export many documents to build a research data set, you only need one API call.
Let's do a quick demonstration.
Query FHIR data in Amazon HealthLake
Open 161cda652a33e0 Amazon Cloud Technology Console , and click "Create a Data Store". Then, I just choose a name for my data store and decide to encrypt it with a key managed by Amazon Cloud Technology. I will also check the box to preload the sample synthetic data. This is a great way to quickly start using the service without uploading my own data.
- Amazon Cloud Technology Console
After a few minutes, the data store is active and I can send queries to its HTTPS endpoint. In the example below, I look for clinical notes (and only clinical notes) that contain the ICD-CM-10 "hypertension" entity, with a confidence score of 99% or higher. In the background, the Amazon Cloud Technology console will send an HTTP GET request to the terminal node. I highlighted the corresponding query string.
The query only takes a few seconds to run. When I checked the JSON response in my browser, I found two documents. For each document, I see a lot of information: when it was created, the organization it belongs to, who the author is, and so on. I will also see Amazon HealthLake automatically extract a long list of entities, including name, description, and confidence score, and add it to the document.
The document is attached to the response in base64 format.
I saved the string into a text file and then decoded it using the command line tool, then I saw the following:
Nesser 先生是一位 52 岁的白种男性，有很多既往病史，包括冠状动脉疾病、房颤、高血压、高脂血症，就诊于北急诊科，主诉寒战、恶心、急性左腹疼痛和左腿有些麻木
This document is completely correct. As you can see, querying and retrieving the data stored in Amazon HealthLake is very simple.
- Amazon HealthLake
Analyze the data stored in Amazon Health
You can export data from Amazon HealthLake , store it in a Amazon Simple Storage Service (Amazon S3) bucket, and use it for analysis and ML tasks. For example, you can use Amazon Glue transform data, use Amazon Athena query data, and use Amazon QuickSight visualize the data. You can also use this data to build, train, and deploy ML models Amazon SageMaker
- export data
- Amazon Simple Storage Service (Amazon S3)
- Amazon Glue
- Amazon Athena
- Amazon QuickSight
- Amazon SageMaker
following blog post shows you the end-to-end analysis and ML workflow based on the data stored in
- Population health applications with Amazon HealthLake: Analytics and monitoring using Amazon QuickSight
- Building predictive disease models using Amazon SageMaker with Amazon HealthLake normalized data
- Build patient outcome prediction applications using Amazon HealthLake and Amazon SageMaker
- Build a cognitive search and a health knowledge graph using Amazon AI services
Last but not least, this self-paced seminar will show you how to use Amazon HealthLake to import and export data, how to use Amazon Glue and Amazon Athena to process data, and how to build an Amazon QuickSight control panel.
Now, let’s take a look at the results our customers have achieved with Amazon HealthLake.
- Self-paced seminar
Customers are already using Amazon HealthLake
Chicago-based Rush University Medical Center is an early adopter of Amazon HealthLake. They used this service to build a public health analysis platform on behalf of the Chicago Department of Public Health. The platform aggregates, merges, and analyzes data related to patient admissions, discharges and transfers, electronic laboratory reports, hospital capacity, and clinical care documents of patients with new coronary pneumonia (COVID-19) treated in Chicago hospitals from multiple hospitals. 17 of Chicago's 32 hospitals are currently submitting data, and Rush plans to integrate all 32 hospitals by this summer. For more information, see this blog post .
- Rush University Medical Center
- Blog post
Recently, Rush launched another project aimed at identifying the communities at the highest risk of high blood pressure, understanding the social determinants of health, and improving healthcare services. To this end, they collect various data, such as clinical notes, ambulatory blood pressure measurements in the community, and medical insurance claim data. Then, the data is ingested into Amazon Health Lake and stored in FHIR format for further analysis.
Dr. Bala Hota, vice president and chief analytics officer of Rush University Medical Center, said: “We don’t need to spend time building unrelated projects or rebuilding existing projects. This allows us to enter the analysis phase faster. Amazon HealthLake has indeed accelerated We deliver the insights we need to deliver results to the masses. We don’t want to spend all our time building infrastructure. We want to provide insights."
Cortica's mission is to bring revolutionary healthcare services to children with autism and other developmental disabilities. Today, Cortica uses Amazon HealthLake to store all patient data in a standardized, secure, and compliant manner. Using these data to build an ML model, they can track the progress of the patient's treatment through sentiment analysis, and can share the child's progress in language development and motor skills with parents. Cortica can also verify the effectiveness of the treatment model and optimize the drug treatment plan.
Ernesto DiMarino, Head of Enterprise Applications and Data at Cortica, told us: “In just a few weeks, not months, Amazon HealthLake helped us create a centralized platform that can safely store patient medical history and medications. History, behavior evaluation and laboratory reports. This platform allows our clinical team to gain a deeper understanding of the patient’s care progress. Using predefined notebooks in Amazon SageMaker and data from Amazon Health Lake, we can apply machine learning models to track and Predicting the progress of each patient in achieving their goals is impossible using other methods. Through this technology, we can also share HIPAA compliance with our patients, researchers, and healthcare partners in an interoperable manner Standard data will further advance important research in the treatment of autism."
MEDHOST provides products and services to more than 1,000 healthcare institutions of various types and sizes. These customers want to develop solutions that standardize patient data in FHIR format, and build control panels and advanced analytics to improve patient care services, but this task is currently difficult and time-consuming.
Pandian Velayutham, Senior Director of Engineering at MEDHOST, said: “With Amazon HealthLake, we can create a compliant FHIR data store in just a few days instead of weeks, and integrate natural language processing and analysis to improve the hospital’s operational efficiency and provide Better patient care services to meet customer needs."
Amazon HealthLake is currently available for the US East (N. Virginia), US East (Ohio) and US West (Oregon) regions.
- Amazon HealthLake:
Try to learn our self-paced seminar and provide us with feedback. As always, we look forward to your feedback. You can send feedback through your usual Amazon Support contacts or post your feedback to Amazon Cloud Technology Forum .
- Self-paced seminars:
- Amazon Cloud Technology Forum:
Author of this article