Amazon and Google are finally addressing one of healthcare's most pressing issues: unstructured health data.

Clinical contacts with patients, population health statistics and trends, community health outcomes, pharmacological trials, and so on are only a few of the various silos that comprise the concept of “healthcare” as a whole. Even at the individual level—medical records, medication history, previous lab testing and findings, prior procedures—there are numerous silos that comprise a person's healthcare profile.

Given the importance of each of these components individually, it is natural to believe that these disparate data sets should interact with one another and be considered holistically when approaching patient treatment or population health decisions for a community.

This, however, is not the case. In practice, reconciling these many silos to construct a composite database in order to make better-informed judgments is incredibly difficult. This is because these various pieces of information are frequently in their own unique forms, running on their own proprietary software or system, and possibly arranged in incompatible ways.

This is the precise issue that technology companies are currently attempting to fix, with Amazon and Google leading the way.

Amazon HealthLake is a tool that "securely stores, transforms, queries, and analyses health data in minutes." The product delivers end-to-end services, including: “Use natural language processing (NLP) to extract meaning from unstructured data for easy search and querying.” […] Make health-data predictions with Amazon SageMaker machine learning (ML) models and Amazon QuickSight analytics [...] Create a comprehensive and chronological view of patient health information, including prescriptions, treatments, and diagnoses [...] Encourage the use of interoperable standards such as the Fast Healthcare Interoperability Resources (FHIR) format.”

Amazon HealthLake is designed to: convert structured and unstructured data into FHIR format (the site provides examples of lab reports, medical records, insurance claims, doctor's notes, and so on); assist in developing insights and usable meaning from that data in a searchable and indexed format, and promote even deeper predictions and data-driven decision making in conjunction with advanced tools.