Which managed FHIR store should I pick?

Usually the one matching your existing cloud (AWS HealthLake, Azure Health Data Services, or Google Cloud Healthcare API) because the value is in native integration with that cloud's analytics, ML, and data services. If you're cloud-agnostic, weigh their specific FHIR features and pricing.

Can I move from a managed store to self-hosted later?

You can export FHIR data and load it into HAPI, but migrating means re-creating any provider-specific analytics pipelines and integrations. Plan for that cost up front if avoiding lock-in is a priority; keeping your access layer abstracted helps.

Are AWS HealthLake, Azure, and Google's FHIR stores HIPAA eligible?

Yes. All three are HIPAA-eligible services offered under the cloud provider's BAA. That covers the infrastructure and the managed service; you're still responsible for configuring access control, encryption, audit logging, and application-level safeguards correctly. HIPAA eligibility is the starting point, not automatic compliance.

How much does a managed FHIR store cost?

Pricing is usage-based (typically a mix of data stored, requests/operations, and features like analytics or NLP) so it scales with volume rather than a flat license. Self-hosted HAPI trades those service fees for infrastructure plus the team to run it. Model both against your expected data volume before assuming one is cheaper.

Can I run HAPI FHIR in the cloud?

Yes. Self-hosted doesn't mean on-premises: you can run HAPI in containers on any cloud (AWS, Azure, GCP) or on-prem. The distinction from a managed store is that you own deployment, scaling, upgrades, and security, which is exactly the control-versus-operational-burden trade at the heart of this decision.

How long does it take to build a FHIR data store?

It depends on how much data you load and from where, which FHIR version and US Core profiles you must support, the searches and bulk exports your apps need, and whether you self-host or use a managed store. Mapping legacy data into FHIR resources is usually the largest piece, so we size it during scoping.

Do you sign a BAA for FHIR data store work?

Yes. When the store will hold protected health information for a covered entity or business associate, we sign a business associate agreement before any PHI is shared. If you use a managed store, you also sign the cloud provider's own BAA for that service.

Can you integrate a FHIR data store with our existing systems?

Yes. We load data from your EHR, HL7 v2 feeds, claims or device pipelines into the store, map it to FHIR resources and expose it to your apps through FHIR APIs and SMART on FHIR authorization. Existing systems keep running while the store fills.

Can we start with a fixed-scope MVP?

Yes. We scope a first release to one workflow, such as one data source loaded into the store and one app reading it through FHIR search. More sources, bulk export and analytics follow once that path is proven with your own data.

How do you use AI in delivery without risking quality?

AI tools help our engineers draft code and tests, but a person reviews every change, and it passes the same automated tests and security checks as any other code, including FHIR validation against your profiles. Which tools may see your code or data is agreed with you at the start.

A home office server with a data centre campus outside the window

Decision Guide

FHIR data stores compared (HAPI / HealthLake / Azure / Google)

Choose a managed FHIR data store (AWS HealthLake, Azure Health Data Services or Google Cloud Healthcare API) to run a scalable FHIR server without managing infrastructure. Choose self-hosted HAPI FHIR for full control, cloud portability and deep customization, at the cost of running it yourself.

15-Minute Scoping Call All Comparisons

What is a FHIR data store?

Quick answer

A FHIR data store is a server that stores health data as HL7 FHIR resources and answers FHIR API reads, searches and bulk exports. You can self-host one, such as HAPI FHIR, or use a managed store from AWS HealthLake, Azure Health Data Services or Google Cloud Healthcare API. Agnotic Technologies builds and integrates either option.

Criterion	HAPI (self-host)	Managed cloud
Operational burden	You run it	Provider runs it
Cost model	Infra + your team	Managed service fees
Control / customization	Full	Within service limits
Cloud portability	High (any host)	Locked to that cloud
Scaling & HA	Your responsibility	Handled by provider
Analytics / ML integration	Build it	Native to the cloud
Compliance (infra)	You own more of it	Shared with provider

What are self-hosted and managed FHIR stores?

What is HAPI (self-host)

HAPI FHIR is the leading open-source FHIR server (Java), which you deploy and operate on your own infrastructure or containers. You control the version, storage, extensions, and scaling, with no managed-service fees, in exchange for owning deployment, upgrades, security, and operations.

What is Managed cloud

Managed FHIR stores (AWS HealthLake, Azure Health Data Services and Google Cloud Healthcare API) are hosted FHIR servers run by the cloud provider. They handle scaling, availability and much of the compliance surface, and connect to each cloud's analytics and ML services.

Self-hosted vs managed FHIR data platforms

Operational burden is the core trade

A managed FHIR store means the provider handles scaling, high availability, patching and infrastructure compliance, and you consume an endpoint. Self-hosted HAPI FHIR means you own all of that. Small teams without deep DevOps gain the most from managed; teams that want control self-host.

Portability vs lock-in

HAPI FHIR runs anywhere, on-premises, in any cloud or hybrid, so you're not tied to a vendor. Managed stores bind you to AWS, Azure or Google, and migrating off is non-trivial. If exit options matter, self-host; if you're committed to one cloud, lock-in matters less.

Analytics and ecosystem integration

Managed FHIR stores shine when you want more from the data: HealthLake integrates NLP and analytics, and Azure and Google connect to their data and ML stacks. Rebuilding those pipelines on HAPI FHIR is engineering you'd own, so managed helps when analytics is central.

Cost shape

Self-hosting trades service fees for infrastructure plus the people to run it, which can be cheaper at scale if you have the team. Managed pricing is predictable and includes operations, but per-store and per-operation fees grow with usage. Model both against your real data volume.

When should you self-host or use a managed store?

When to use Self-Hosted (HAPI FHIR)

You need full control over configuration, extensions, and FHIR version.
You want cloud portability and to avoid managed-service lock-in.
Per-store or per-operation cloud fees don't fit your cost model.
You have the DevOps capacity to run, scale, and secure a stateful server.
You're building on-prem or in a multi/hybrid-cloud environment.

When to use Managed (HealthLake / Azure / Google)

You're already on AWS, Azure, or GCP and want native FHIR without running servers.
You want the provider to own scaling, availability, and infra-level compliance.
You value built-in analytics/ML integrations (e.g. HealthLake's NLP, BigQuery export).
You'd rather trade some control and cloud lock-in for lower operational burden.
You need to stand up a compliant FHIR store quickly with a small team.

What should a FHIR server proof of concept test?

A FHIR server proof of concept should run your own app's queries and a realistic volume of your data against each candidate: HAPI FHIR, AWS HealthLake, Azure Health Data Services or Google Cloud Healthcare API. These gaps usually surface only after go-live.

Test US Core profile validation, the searches you really send (chained parameters, _include and _revinclude), Bulk Data $export and import at your volume, subscriptions, SMART on FHIR authorization, history and audit export, data residency, and a full export into another server to price your exit.

Which sources does this comparison rely on?

AWS HealthLake / Google Cloud Healthcare API FHIR Integration & API Development Healthcare Data Analytics Healthcare Cloud Migration & Modernization Choosing a cloud provider for healthcare workloads HIPAA requirements for software teams

Frequently Asked Questions

Yes. HAPI is a mature, widely-used open-source FHIR server that runs in production at scale. The caveat is that you own deployment, scaling, security, and upgrades, so it's production-ready for teams with the DevOps capacity to operate it.

Not sure which fits your build?

Tell us about your product, integration targets, and timeline. We'll map the trade-offs to your context and recommend the path that gets you to a compliant launch fastest.

15-Minute Scoping Call Explore Solutions