AWS面向普通用户推出健康湖：Amazon HealthLake

AWS公司今天宣布Amazon HealthLake（亚马逊健康湖）上架，医疗机构利用Amazon HealthLake云服务可以将来自不同系统的医疗信息集中到一个地方并可以利用这些信息改善病人的护理。

Medical data is often spread across multiple applications. At a hospital, clinical notes, lab reports and insurance claims may be kept in different databases, which can make it difficult for healthcare professionals to access the information in a timely manner. AWS says that using HealthLake to centralize such disparate records in a single data repository simplifies information access.

医疗数据往往都分散在多个不同的应用程序里。医院的临床记录、实验室报告及保险索赔可能保存在不同的数据库里，因此医疗专业人士可能难以及时获取信息。AWS表示，利用HealthLake可以将不同的记录集中在一个数据存储库里并可以简化信息访问。

“In their daily practice, doctors need a complete chronological view of patient history to identify the best course of action,“ AWS artificial intelligence and machine learning evangelist Julien Simon wrote in a blog post today. “During an emergency, giving medical teams the right information at the right time can dramatically improve patient outcomes. Likewise, healthcare and life sciences researchers need high-quality, normalized data that they can analyze and build models with, to identify population health trends or drug trial recipients.”

AWS人工智能和机器学习推广人士Julien Simon今天在一篇博文中表示，“医生在日常工作中需要查看病人的病史的完整时间顺序，以便找到最佳医治方案。医疗团队在急诊时能在正确的时间获取正确的信息则可以极大地改善救治病人的效果。医疗保健和生命科学研究人员也需要高质量的、规范化的数据，他们同样可以利用这些数据进行分析和建立模型，达到确定人口健康趋势或药物试验对象的目的。”

HealthLake provides a set of connectors that can automatically import records from commonly used healthcare applications. Once the records are in the service, natural language processing algorithms organize the data to make it easier to work with.

HealthLake提供的一套连接器可以自动从常用的医疗保健应用程序导入记录。这些记录进入HealthLake服务后，系统里的自然语言处理算法可以组织数据，令数据更容易操作。

The algorithms scan medical documents such as clinical notes and extract the patient information they contain. HealthLake then indexes the information similarly to how a search engine indexes web pages. This process turns the raw unstructured records into a form that can be easily searched later on for specific items even if the size of the database grows to several petabytes over time.

这些算法先扫描诸如临床笔记的医疗文件并提取记录所包含的病人信息。然后，HealthLake再对这些信息进行索引，类似于搜索引擎对网页的索引。这个过程将原始的非结构化记录转成一种带格式的数据，得到的数据库规模可能最后达到几个PB，但却仍可以在以后轻松搜索到特定的数据。

The result, according to AWS, is that healthcare professionals gain the ability to obtain information about patients’ medical histories faster than by manually sifting through documents.

据AWS称，与手动筛选文件相比，HealthLake令医疗保健专业人员可以更快地获得病人的病史信息。

HealthLake also enables healthcare organizations to transfer the information to other systems for further processing. The platform stores medical data in an industry-standard data format known as FHIR. It provides the ability to move records to many clinical applications without the need for complicated formatting changes. The data transfer itself, in turn, is performed using an application programming interface that AWS provides as part of the HealthLake feature set.

HealthLake也可以令医疗机构将信息迁移到其他系统进行进一步处理。HealthLake平台上医疗数据的存储用的是名为FHIR的行业标准数据格式。FHIR提供了将记录迁移到许多临床应用的功能，无需进行复杂的格式化修改。而数据传输本身则是通过AWS提供的应用编程接口进行，是HealthLake功能集的一部分。

One of the use cases AWS promises to enable with the service is harnessing artificial intelligence to improve patient care. Researchers are working on ways of using AI algorithms to automate tasks such as diagnosing medical conditions and planning treatment plans. HealthLake can ease the logistics of streaming medical data from a healthcare provider’s systems to the AI algorithms that analyze it to generate recommendations.

AWS该项服务可望实现的一个用例是利用人工智能改善病人护理。研究人员目前在研究如何利用人工智能算法自动完成诊断病情和规划治疗计划等任务。医疗数据需要从医疗服务提供者的系统里以数据流的形式进入人工智能算法，算法再分析这些数据生成建议，HealthLake可以简化数据的传送。

AWS manages the cloud infrastructure hosting each HealthLake environment to simplify maintenance for customers. When the data in HealthLake is at rest, it’s encrypted using the AWS Key Management Service, which allows healthcare providers to scramble records either with an encryption they generate themselves or with one supplied by the cloud giant. TLS 1.2 encryption protects the data while it’s moving across the network.

（编辑：广州站长网）

【声明】本站内容均来自网络，其相关言论仅代表作者个人观点，不代表本站立场。若无意侵犯到您的权利，请及时与联系站长删除相关内容!