API

Google Healthcare Natural Language API

Unlock insights from unstructured data in clinical notes, scientific literature, lab results, radiology reports and more 

Valuable information is often only available in unstructured text that is hard to process at scale. Customers can use NLP API to extract structured insights, such as identifying patients with the presence of specific cancer phenotypes, from clinical records. 

Google is a third-party product: By requesting information, you consent to your information being shared with Google.

Knowledge extraction maps text to medical concepts

Healthcare NLP API identifies and classifies diagnoses (ICD-10), medications (RxNorm), procedures (LOINC) and more, enabling you to unlock valuable insights hidden within your text data..

Relation extraction identifies and connects related medical attributes

Healthcare NLP API identifies and connects related medical attributes, such as patient-disease, drug-condition and procedure-anatomy relationships. This allows for a deeper understanding of patient data.

Knowledge linking standardizes medical concepts so they can be analyzed across systems

Ensure interoperability and accurate data analysis across your healthcare ecosystem. Our knowledge-linking feature standardizes medical concepts by mapping synonyms and different coding systems (e.g., SNOMED CT, ICD-10-CM) to standardized medical ontologies.

Developer resources

Google Healthcare NLP API

Related resources
Whitepaper

Best-in-class NLP performance

Google Healthcare NLP achieve best-in-class NLP performance with leading precision, recall and F1 scores (statistical measures to rate performance).

Article

How NLP Can Solve the Enterprise Challenge of Unstructured Data

Applied natural language processing (NLP), using serverless software components on Google Cloud, provides an efficient way to identify and guide clinical abstractors toward a prioritized list of patient medical documents.

Frequently Asked Questions

Q: Can I use Google’s Healthcare Natural Language API while I wait for it to become available on Optum AI Marketplace?

A. Yes. You can test out a demo of Google’s Healthcare Natural Language API by following these instructions.

Q: What kind of information can this API extract?

A: Medical concepts: Diagnoses (like diabetes), medications (like aspirin), procedures (like appendectomy) 

Relationships: Links between concepts (like patient suffers from disease or drug treats condition) 

Contextual clues: Things like the certainty of a diagnosis (likely vs. unlikely), when something occurred (past, present, future), and who it relates to (patient, family member)

Q: What formats does the API support for input and output?

A: The API accepts medical text as input (up to 20,000 characters). You get results in JSON format, which is a common way to represent data. You can also get the output in FHIR R4 format, a standard for exchanging healthcare information.

Q: What languages and medical vocabularies are supported? 

A: Currently, the API only supports English text. However, it understands a wide range of medical vocabularies, including ICD-10, SNOMED CT, RxNOrm and LOINC, which ensures it can accurately interpret and code the information it extracts. 

Marketplace updates

Subscribe to The Spark

We're adding new products and services to rapidly expand our marketplace and the potential is huge. Be an early adopter and stay ahead of the trend with our bi-weekly e-updates.