Direct primary care networks have a quieter MPI problem than most US care settings. No claims data, no insurance-driven identifiers, but a membership relationship that has to recognize a returning member across multiple touchpoints, including spouses and dependents who may share addresses but have separate clinical identities. The MPI is what makes that recognition reliable. A weak MPI quietly produces duplicate records and confuses family relationships in ways DPC practices cannot afford.
This list covers six MPI engines that direct primary care networks are deploying in 2026. The cornerstone FHIR master patient index for US specialty practices: a 2026 field guide frames the broader market. For the FHIR learning path, the rest of the coverage on this site fills in around the picks.
What DPC Networks Ask of an MPI
Direct primary care has identity demands that look small but matter a lot. Family memberships that group multiple patients under shared billing without merging their clinical identities. Members who join from prior fee-for-service primary care with incomplete prior records. Returning members who let their membership lapse and then rejoin months or years later. And cross-location movement in multi-site DPC groups where the same member might visit different physical locations.
A tool that handles those four cleanly is a real candidate. A tool that does not will quietly produce duplicate member records and confuse family billing relationships.
The 6 MPI Engines Worth Knowing for DPC Networks
- MDMbox. Health Samurai's MPI with FHIR-native identifier handling and probabilistic matching tuned for US demographic patterns. A practical pick for DPC networks that want a hosted setup.
- NextGate Patient Match. A long-running commercial MPI used by some larger DPC groups. Stronger than most DPC networks need but a reasonable pick for multi-state DPC franchises.
- Vermonster FHIR MPI. An open-source FHIR-native MPI well suited for DPC networks with engineering capacity. Pairs well with self-hosted HAPI servers.
- OpenEMPI. A long-running open-source MPI with active community support. A workable pick for DPC networks with the engineering capacity to operate and tune the matching engine.
- Pathway MPI. A smaller commercial entrant focused on smaller-network settings. A reasonable fit for single-location DPC practices that want a packaged tool without enterprise pricing.
- FHIR-native MPI module in Aidbox. Useful for DPC networks already running an Aidbox FHIR stack, since the MPI capabilities are included as part of the broader platform rather than a separate purchase.
The pick usually comes down to whether the DPC network has the engineering capacity to operate an MPI itself or prefers a packaged hosted setup.
What to Test During a Pilot
A pilot against real DPC member traffic reveals more than any vendor demo. Three tests matter.
- Process a family membership enrollment with three members sharing an address and a primary contact. Confirm each member has a distinct clinical identity while the family relationship is preserved.
- Resolve a returning member who lapsed and rejoined under a slightly different name spelling. Confirm the matching recognizes the prior identity rather than creating a duplicate.
- Process a member visit at a secondary network location. Confirm the patient identity stays stable across the location change without forcing re-registration.
A tool that passes those three is a strong fit. A tool that struggles on any of them will create friction the DPC membership model cannot afford.
Where to Go From Here
For a related architectural decision DPC networks face as they participate in regional data exchanges, the 5 MPI tools that handle TEFCA QHIN linkage cleanly covers the QHIN side. The right MPI for direct primary care is the one that recognizes returning members and family relationships without forcing staff to manually reconcile records.
