Regional health networks face the open-source vs commercial MPI question with sharper trade-offs than larger systems do. The budget pressure usually favours open source, the operational ownership often does not, and the governance complexity sits somewhere between a single trust and a national programme. The honest answer depends less on philosophical preferences than on the network's actual staffing, governance capacity, and risk tolerance. The comparison below walks through the trade-offs as they actually show up in 2026 deployments.

For broader context on master patient index selection, the broader FHIR reference shelf is the right entry point to this comparison.

The Real Question Behind the Decision

The open-source vs commercial MPI question for a regional network is really three questions:

• Does the network have, or can it hire, engineering and operations staff who will own the MPI as a long-term operational responsibility?
• Does the network's governance model favour collective ownership of the matching policy or vendor-managed policy?
• What is the network's risk tolerance for matching incidents, and is that tolerance shaped by clinical safety or by regulatory exposure?

A network with strong engineering, collective governance, and a tolerance for owning operations does well on open source. A network with thin engineering capacity, vendor-led governance, and a sharp risk model usually does better commercially.

Where Open-Source MPI Wins

Open-source MPIs (HAPI FHIR EMPI, OpenEMPI, the identity layers in Medplum, and the various community-maintained options) give a regional network three things that matter:

• Zero recurring licence cost, which compounds across multi-year deployments and is hard to argue with for budget-pressured networks.
• Full control over matching policy, with the ability to tune the policy collectively across the network's providers.
• A verifiable code base for clinical safety reviews, which simplifies the conversation with each provider's safety officer.

The honest cost is the engineering time required to operate the MPI, keep it current, and respond when matching incidents happen. A network that genuinely cannot staff this work will struggle with open source.

Where Commercial MPI Wins

Commercial MPIs (Smile CDR MDM, NextGate MatchMetrix, Verato, Rhapsody Patient Index, IBM's healthcare MPI) tend to be the right pick when:

• The network lacks the engineering staffing to own the MPI as a long-term operation.
• The governance model prefers a single point of accountability for matching accuracy and incident response.
• The risk tolerance demands a vendor on the hook, particularly for regulated reporting workflows.
• The procurement model favours a single support contract over multi-component ownership.

The trade-off is a recurring fee, less freedom to customise matching policy, and the standard vendor-lock-in concerns.

A Pragmatic Decision Approach

Most regional networks land in one of three patterns:

• Open-source MPI (usually HAPI EMPI) for networks with strong engineering and a collective governance model.
• Commercial MPI (often Smile CDR MDM or NextGate) for networks that prefer vendor accountability.
• A hybrid where the network runs an open-source MPI for the bulk of the matching work and contracts a commercial referential matching service (often Verato) for the edge cases where local demographic data is too thin.

The third pattern is more common than either pure approach, because it gets most of the cost advantage of open source while bringing commercial coverage to the specific cases where open source struggles.

The Trap to Avoid

The trap is picking an open-source MPI to save money and then not staffing the operational work, which produces a stale matching policy and a growing backlog of unresolved stewardship cases. The mirror trap is picking a commercial MPI for reassurance and then not using the support services, which produces a network paying for a relationship it never engages with.

The honest test is to look at the network's realistic five-year staffing plan and pick the operating model the plan can sustain.

For broader strategic context on master patient index selection, the practical guide to master patient index for FHIR in 2026 is the right back-reference. For the closely related matching style decision that interacts with the open-source vs commercial choice, the deterministic vs probabilistic patient matching comparison is the natural companion read.

Sources

• MDM module overview (open-source MPI baseline) - HTML docs, HAPI FHIR
• commercial MDM module documentation - HTML docs, Smile CDR
• Patient Deduplication Architectures - HTML docs, Medplum