HealthITPrep
CPHIMS guide Concepts Features Pricing FAQ
Sign in Start free
العربية

Health information exchange models: centralized, federated and hybrid

By Adel Moustafa, CPHIMS® · Updated 22 October 2026

How hospitals and clinics share patient data across organisations: the three forms of exchange, the three architectures, and the pieces that make it work, from patient matching to consent.

In short: in a centralized health information exchange (HIE), members send copies of their data to one shared repository. In a federated (decentralized) HIE, the data stays in each member's own systems and is fetched only when a request comes in, usually found through a record locator service. A hybrid keeps some data or indexes centrally and leaves the rest with the members. All of them depend on matching the right patient across organisations.
What an HIE is Three forms Three architectures Side by side The pieces Consent Common mistakes Summary Sources

What is a health information exchange?

The word means two things. As an activity, health information exchange is the electronic sharing of patient information between organisations: hospitals, clinics, laboratories, pharmacies, public health. As an organisation, an HIE is the body (often regional or national) that runs the shared services and the rules that let its members exchange. The aim is that the information follows the patient, so that a clinician in an emergency department can see the medications and allergies recorded by another hospital.

What are the three forms of exchange?

  • Directed exchange: sending information to a known recipient, such as a referral letter or a discharge summary sent securely from the hospital to the family doctor. It is "push".
  • Query-based exchange: searching for and requesting a patient's information from other organisations, typically for unplanned care, such as an emergency visit. It is "pull".
  • Consumer-mediated exchange: the patient collects and controls their own information, for example through a portal or an app, and shares it with whom they choose.

What are the centralized, federated and hybrid architectures?

  • Centralized: members send copies of their data to a central repository run by the HIE. Queries are fast and the data is easy to use for population health and reporting, but members give up some control, the copies must be kept current, and a single store of everyone's data is a larger target.
  • Federated (decentralized): each member keeps its data in its own systems. When a request comes, the HIE finds which members hold records for that patient and the data is fetched from them at that moment. Members keep control and nothing is copied in advance, but the answer depends on every source being online and fast. This is the usual answer when members insist on keeping the data in their own systems.
  • Hybrid: a mix, for example a central master patient index and record locator with the clinical data left at the sources, or a central store for a small summary (allergies, medications, problems) and federated queries for the rest.

The three architectures side by side

CentralizedFederatedHybrid
Where the data livesCopies in one central repositoryAt each member, fetched on requestSome centrally, the rest at the members
Control by membersLowerHighestIn between
Speed of a queryFast, one placeDepends on every source respondingFast for the central part
Use for population analyticsEasiestHardestPossible for the central part

What pieces does an HIE need?

  • Master patient index (MPI): links the identifiers each organisation uses for the same person. Matching uses names, date of birth, sex, address and other details, with deterministic rules (exact agreement on chosen fields) or probabilistic ones (a score of how likely two records are the same person). A false match mixes two people's records; a missed match hides information. Both are patient-safety risks.
  • Record locator service: knows which members hold records for a patient, without holding the records.
  • Provider directory: who and where the recipients are, and how to reach them securely.
  • Standards: HL7 messages and documents (C-CDA), FHIR APIs, and shared code sets, so that what arrives can be used.
  • Security and governance: authentication, audit logs, a data use and participation agreement that every member signs, and rules for who may see what.

In the United States, the Trusted Exchange Framework and Common Agreement (TEFCA) sets common rules so that networks can exchange with each other through Qualified Health Information Networks (QHINs); the first QHINs were designated in December 2023.

How is patient consent handled?

  • Opt-in: the patient's information is shared only after the patient agrees.
  • Opt-out: it is shared unless the patient declines.
  • Variants allow choices by type of data or by organisation, and some categories, such as substance use treatment records in the United States, have stricter rules of their own.

The choice affects how much data is available: opt-out usually gives far more complete records, opt-in gives patients more explicit control.

What do candidates get wrong?

  • Choosing centralized when members want to keep their data. That is the federated model: data stays at the source and is fetched on request.
  • Calling a record locator a repository. It knows where the records are; it does not hold them.
  • Swapping directed and query-based exchange. Directed pushes to a known recipient; query-based pulls on request.
  • Underrating patient matching. Without reliable matching, exchange either shows the wrong patient's data or misses the right one's.
  • Thinking a shared standard is enough. Exchange also needs governance, agreements and trust between organisations.

Summary for revision

  • Forms: directed (push to a known recipient), query-based (pull on request), consumer-mediated (the patient controls).
  • Centralized: copies in one repository; fast, good for analytics, less member control.
  • Federated: data stays at the members, fetched on request through a record locator; most control.
  • Hybrid: central index or summary, the rest at the sources.
  • Needs: a master patient index, standards, security, governance and a consent model (opt-in or opt-out).

Sources

  • Office of the National Coordinator for Health IT (ONC, now ASTP/ONC), HealthIT.gov: health information exchange and its three forms.
  • ONC: the Trusted Exchange Framework and Common Agreement (TEFCA).

Related free practice questions

  • Health information exchange architectures

The site files this concept under the domain Healthcare and Technology Environments.

Try exam-style questions

A 20-question trial exam with an explanation for every option and a score per domain, drawn from 40 free questions. No payment method needed.

Start the trial exam

Questions about the exam: CPHIMS guide · What is CPHIMS? · Exam cost · Eligibility · Passing score · Renewal · Exam day · How long to study · Is it worth it? · What is CAHIMS?

Concepts explained: HL7 v2 vs FHIR · Capitation vs fee-for-service · Safe Harbor vs Expert Determination · Incident vs problem vs change · Backup types, RTO and RPO · Privacy Rule vs Security Rule · CDS five rights · Lewin vs Kotter · SDLC and go-live strategies · Testing types · Earned value (SPI, CPI) · Structure, process, outcome · Six Sigma vs Lean · Levels of interoperability · Bandwidth vs latency · EMTALA · Risk responses · RACI matrix · Build vs buy · Sunk vs opportunity cost · Data governance

The four domains: Healthcare and Technology Environments · Clinical Informatics · Healthcare Information and Systems Management · Management and Leadership

Written by Adel Moustafa, About us

Written and checked against the sources above on 22 October 2026. An independent site, not affiliated with or endorsed by HIMSS. CPHIMS® is a registered trademark of HIMSS.

CPHIMS guide Compare certifications Free practice questions Concepts explained About us Terms of Use Privacy Policy Refund Policy Contact us

© 2026 HealthITPrep. This site is not affiliated with or endorsed by HIMSS. CPHIMS is a registered trademark of HIMSS.