Amazon Web Services is rolling out a new tool designed to make it simpler for healthcare organizations to work with their data. Amazon HealthLake, which AWS describes as “HIPAA-eligible,” was launched during Amazon’s annual re:Invent conference.
The new service, which is now available in preview, was designed to help healthcare and life sciences organizations aggregate data across organizational silos and differing formats into a single AWS data lake. HealthLake is capable of understanding and extracting medical information and functions including rules, procedures and diagnoses in real-time.
The service will normalize the information its customers provide automatically using machine learning, the tech giant said. HealthLake options include allowing organizations to copy health data from on-premise EHRs.
AWS is targeting customers such as health systems, pharmas, clinical researchers and health insurers, which will be able to spot trends and unexpected patterns within health data. Among the things they should be able to do with this platform is make more accurate and specific predictions about the progression of disease, the efficacy of drugs being tested in clinical trials in the appropriateness of insurance premium levels.
The new service identifies each piece of clinical information as it enters the system, tags and indexes events and a timeline view with standardized labels allowing it to be easily searched. HealthLake then structures all the data into FHIR format. Once the data is standardized, healthcare organizations are able to use that data to get a more complete view of individual patient health along with that of entire populations, the company said.
For example, HealthLake can use natural language processing and ontology mapping to figure out whether a patient has been properly prescribed a drug. For example, by pulling out information on blood glucose levels and cross-referencing that data with physician notes, lab reports and claims data, providers can do a better job of managing individual cases of diabetes.
Collection of data like these can be queried on an ongoing basis and it is searched using standard data science skills or pulled into Amazon’s SageMaker to train models that can forecast the number of diabetes cases providers can expect year-over-year in the population they manage.
This announcement follows on what seems like an inexorable march into healthcare services by Amazon. Its most recent offering, the April launch of a free-to-access COVID-19 data lake by AWS, followed the late 2019 launch of its Amazon Transcribe Medical offering. Transcribe Medical uses AI to capture both dictated documentation and conversations with patients, then translate what it hears into text. Plus, this doesn’t include everything Amazon is doing on the pharmacy side.
What will be interesting to see, to some extent even more than these traditionally solid healthcare analytics ventures, is how Amazon will leverage its massive investments in voice assistant Alexa to capture healthcare spaces (such as home monitoring for the elderly) which aren’t as addressable by traditional enterprise data tools and technologies.
Given how crazy 2020 has been, I’m not surprised that we’ve heard little about advances in voice assistant technology. However, when Amazon starts making healthcare headlines in 2021, expect to see Alexa involved.