Frequently Asked Questions

General Questions

• What is TermHub?

  • Cloud TermHub is an always on terminology management service that provides standard healthcare terminology content for payer, provider, and research oriented digital health applications. At its core, TermHub is a catalog of terminology, concept mapping, and subset (or value set) content that automates management and distribution for “terminology-enabled” applications in healthcare.

  • On top of that core set of content, TermHub is an informatics engine that provides capabilities for search and retrieval via native APIs as well as industry standard FHIR® R4 and R5 APIs. It also supports bulk modes of operation (look up a list of codes all at once), history services (understand change over time), and downloading of content in a variety of formats.

• What is AutomapAI?

  • AutomapAI is TermHub’s intelligent mapping engine that helps convert non-standard content into standardized clinical codes. It supports mapping:

    • Local, proprietary, or legacy codes

    • Free-text terms or labels (e.g., “fasting glucose”)

    • Even full paragraphs of clinical text, such as open-ended responses or provider notes

    This makes AutomapAI useful for a wide range of scenarios—from standardizing homegrown lab codes to translating narrative assessments into SNOMED CT, LOINC, RXNORM, or ICD-10-CM.

    Whether you're cleaning up legacy data, prepping for interoperability, or building clinical decision support, AutomapAI accelerates the process with consistent, scalable mappings.

• Does TermHub support on-prem deployment?

  • Open TermHub refers to the self-hosted, open-source deployment of TermHub using Docker. It lets organizations run their own instance of the platform, complete with user interface and API access, on-premises or in a private cloud. It includes:

    • UI for browsing and searching standard terminologies

    • FHIR® Terminology Server

    • Syndication to Cloud Termhub for automatic terminology management

    • A clean, containerized deployment

    • No enterprise license required

    It’s perfect if you want full control over your data and environment without relying on TermHub’s hosted service.

API Limits and Performance

• What is the rate limit for a single account?

  • Cloud TermHub doesn’t enforce a hard rate limit, but it expects users not to overwhelm the service (for example, by behaving like a DDoS attack). If you anticipate high request volumes, we recommend deploying the Open TermHub container with your licensed data for local, scalable use.

• How many requests can I send at once?

  • Cloud TermHub handles hundreds of near-simultaneous requests without issue. We’ll share detailed benchmarks soon.
    
    At very high volumes, response times may increase, and we may apply throttling through our web application firewall.

• Do rate limits automatically increase with higher usage tiers?

  • No. TermHub’s pricing isn’t tied to usage tiers.
    
    For high-volume needs, we recommend deploying a local Open TermHub container with your managed content, which you can scale horizontally to handle any load.

• What happens if I exceed capacity? Are requests queued or rejected?

  • There are no explicit limits today, so there isn’t a defined failure mode. If the system becomes overloaded, responses may slow down and could eventually start failing.

Authentication and Access

• How do I refresh OAuth2 tokens dynamically?

  • Cloud TermHub uses username/password authentication under the OAuth2 spec. A successful login returns an access token (valid for 2 hours) and a refresh token (valid for 48 hours). Use the OAuth refresh flow with your refresh token to get a new access and refresh token pair as needed.

Bulk Search Features

• How many entries can I include in a single/bulk lookup call?

  • There is not a strict limit on how many entries can be present in a single lookup call and is mostly limited by available memory.

    Certainly thousands of entries can be provided, but the more that are provided, the longer it will take to get a result. There is generally a 5 minute timeout on calls, so any request taking longer than that will return with a gateway timeout (504).

• Does response time grow linearly with entity count, and what overhead comes from querying multiple terminology sets?

  • Currently, Cloud TermHub doesn’t parallelize lookups for multiple values, so response time scales roughly linearly with entity count.

    However, overhead from requesting multiple terminology sets at once is minimal.

• If we query multiple vocabularies (like SNOMED, ICD-10, and RXNORM), are they processed sequentially, and does that slow things down?

  • No. Multiple vocabularies are searched simultaneously, so performance is roughly the same as querying a single terminology.

Terminology Management

• What is the process for uploading new or private terminologies into TermHub?

  • Cloud TermHub supports BYOD (Bring Your Own Data).

    Right now, this is handled manually: you provide your terminology files to our team, and we load them into TermHub for use within your organization’s projects. This kind of data is available ONLY for your account.

    API and UI-based upload features are on our roadmap.

• How often are official terminologies updated, and is there a changelog or webhook for new releases?

  • Cloud TermHub automatically updates standard vocabularies with clear publisher endpoints, such as SNOMED CT, SNOMED CT US, RXNORM, and LOINC.

    Terminologies pulled from UMLS follow its May and November release cycle.

    Less frequently updated terminologies are refreshed on demand when new versions are identified. Changelogs and webhooks for release notifications are on our roadmap.

Advanced Functionality and AutomapAI

• Beyond basic lookup, what other functions are available (e.g., cross-vocabulary mapping)?

  • Cloud TermHub supports find/get operations across terminologies, mapsets, subsets, and concepts, plus history services to track changes across versions. It offers export features, bulk concept lookups (including export), and a complete set of read-only FHIR® R4 and R5 services with Swagger APIs.

    For early access to advanced capabilities—like cross-vocabulary mapping through “AutomapAI”—you can explore the Preview Features Project.

• How is the confidence score calculated? What scale is used? What defines an “exact” vs. “fuzzy” match?

  • In bulk search, the confidence score is simply a raw output from the search engine and isn’t particularly meaningful.

    For AutomapAI, the score is generated by a multi-step algorithm, producing values from 0 to 1. Generally, scores above 0.85 indicate high-quality matches.

• Can TermHub analyze a sample corpus and recommend the minimal terminology set for a target coverage?

  • Not yet. This is a potential future use case, and we expect AutomapAI to eventually support scenarios like discovering which terminologies best cover your data.

    We’re also open to prioritizing such features under partnerships or cost-sharing agreements.

• Does TermHub support risk coding in AutomapAI, like HCC scoring?

  • Not out of the box.

    Right now, AutomapAI focuses on semantic standardization of clinical records and concept-based indexing of narrative content, such as journal articles or help documentation.

    However, risk coding is a valid use case for the AutomapAI platform, and it could be configured or tuned to support it. We’re also open to prioritizing such features under partnerships or cost-sharing agreements.