Why is patient matching so important today?
Population health. Value-based care. Digital front doors. Interoperability mandates. All depend on knowing who is who — yet Pew found up to 1 in 5 patients go unmatched within a single organization, and match rates can fall to 50% between organizations.1 Three generations of patient matching technology have chased that gap: MPI, EMPI, and UMPI. Each compares two imperfect records. Referential matching, the newest generation, brings in outside truth: an independent, continuously updated identity source.
Patient matching technology, compared
What is a Master Patient Index (MPI)?
A master patient index (MPI) is the patient matching engine built into a single system—most often an electronic health record (EHR) — that links records using deterministic matching, an exact or near-exact comparison of demographic fields like name, birth date, and address.
Benefits of Master Patient Index (MPI) solutions
- No added cost or implementation. An MPI ships inside systems you already own—no separate product to buy or vendor to manage.
- Dependable within its own walls. Inside a single application with consistent registration workflows, an MPI handles routine cases well.
Master Patient Index (MPI) challenges
- Blind beyond one system. An MPI cannot tell you whether the patient in your billing system is the person in your lab system.
- Defeated by imperfect data. Out-of-date addresses, maiden names, or missing fields break the comparison, and no configuration recovers data that isn’t there.
- Silent failures. When records don’t line up, they sit unlinked with nothing flagging the gap, so duplicates accumulate unnoticed.
- Resolution falls to the front desk. Ambiguous matches land with registration staff, who decide under time pressure without the patient’s full history.
What is an Enterprise Master Patient Index (EMPI)?
An enterprise master patient index (EMPI) is an patient matching system that sits outside any single application and links patient records across an organization’s EHRs, data warehouses, and analytics platforms using probabilistic matching, a statistical weighing of partial matches rather than an exact one.
Benefits of Enterprise Master Patient Index (EMPI) solutions
- Matches across the enterprise. An EMPI reaches across applications, so it can answer the question an MPI cannot.
- Statistical rather than exact matching. Probabilistic algorithms weigh partial matches instead of requiring every field to align, catching more true matches than an MPI.
- Stewardship moves to trained hands. Unresolved records route to a data steward in health information management rather than to the registration desk.
Enterprise Master Patient Index (EMPI) challenges
- Still comparing two flawed records. An EMPI weighs only the data both records contain, so placeholder birth dates like 1/1/2000 still defeat it.
- A review queue that only grows. Every new data source adds to the manual “potential duplicate” queue, so stewardship burden scales rather than shrinks.
- The re-tuning tax. Most EMPIs need re-tuning by a data scientist or vendor consultant whenever a population shifts or a new source comes online.
- Heavy to stand up and maintain. Deployments carry large upfront costs, months of implementation, ongoing database administration, and periodic upgrade projects.
What is a Universal Master Patient Index (UMPI)?
A universal master patient index (UMPI) is a patient matching system that resolves patient identity across organizations rather than within a single enterprise, so separate institutions can recognize the same person in each other’s records.
Benefits of Universal Master Patient Index (UMPI) solutions
- Identity that travels between organizations. A UMPI answers whether the patient at one health system is the person another plan or exchange already knows.
- Keeps pace with modern data sources. Not tied to one enterprise’s stack, a UMPI ingests patient portals, telehealth, claims, CRM, and SDOH data.
- Cloud delivery, faster to stand up. Delivered as cloud services rather than installed software, UMPIs remove the hardware and upgrade projects on-premises deployments require.
Universal Master Patient Index (UMPI) challenges
- The term isn’t standardized. “UMPI” means different things to different vendors, so two products carrying the label may share nothing beyond the acronym.
- Wider scope is not better accuracy. A UMPI running deterministic or probabilistic algorithms compares two imperfect records over a longer distance — the same ceiling.
- Cross-organizational matching requires governance. Resolving identity between institutions depends on data sharing agreements, consent management, and privacy controls an EMPI never negotiates.
- Value depends on participation. An exchange can only match members’ patients if those members contribute data, so the benefit compounds with adoption.
What is referential matching?
Referential matching is a patient matching approach that resolves identity by comparing each record against an independent, continuously updated reference database of known identities rather than against another record. Because that database already knows who is who, the reference data works as an answer key — confirming that two records belong to the same person even when the records themselves share almost no data in common.
Benefits of referential matching
- Fixes the input, not the reach. Wider scope does nothing about out-of-date or incomplete demographic data. Referential matching repairs the comparison itself.
- Delivers the automation UMPI only promises. A universal index running probabilistic algorithms still queues hard cases. Reference data clears them without a human.
- Improves matching at every scope. Accuracy rises inside a single EHR, inside an enterprise EMPI, and between institutions, so no cross-organizational program is required.
- Removes tuning instead of relocating it. Probabilistic matching needs re-tuning per population or new source. Reference data does that work continuously.
Referential matching challenges
- Only as good as the reference data behind it. Breadth, accuracy, and currency of the database decide everything, so vendors’ results differ widely
- Coverage varies by population. Reference databases are thinnest where identity footprints are thinnest—newborns, pediatric patients, and people who recently arrived in the country.
- Requires sending demographic data to an outside service. Involving a third party brings privacy review, business associate agreements, and security diligence into the evaluation.
- Resolving identity doesn’t clean source systems. Referential matching links the right records, but data quality issues in each application remain for owning teams.
Choosing the right patient matching technology
This isn’t a ranking where the newest acronym wins. Deterministic and probabilistic matching — the algorithms inside MPIs and most EMPIs — share one structural ceiling: both can only compare the data two records already contain. When demographic data is out of date, incomplete, or inconsistent, no amount of tuning recovers what was never captured.
Three questions decide the choice:
- How far does matching need to reach? One application, every application in your enterprise, or across organizational lines.
- How clean is the data going in? The dirtier the source data, the less algorithm tuning can do for you.
- How much manual review can you absorb? Each generation handles unresolved records differently—your front desk, your stewards’ backlog queue, or automated resolution.
Answer those three honestly and the same conclusion tends to surface: the constraint is the data quality going in, not the distance matching has to travel.
How Verato Referential Matching® resolves what MPIs, EMPIs, and UMPIs miss
Every generation of this technology describes the same underlying job: proving that two records belong to the same person. Verato® does it without needing the two records to look alike. Verato Referential Matching® checks each record against Verato Carbon®, a reference database of more than 300 million U.S. identities, and runs inside Verato MDM Cloud™ alongside automated stewardship, enrichment, and governance.2
See it in practice: one health system moved beyond a legacy EMPI, and a national insurance provider automated duplicate resolution to save $18M.
Book a demo to see how much of your stewardship queue Verato MDM Cloud™ clears automatically.
¹Bipartisan Policy Center (2012) and ONC/Audacious Inquiry (2014), as reported by The Pew Charitable Trusts, “Enhanced Patient Matching Is Critical to Achieving Full Promise of Digital Health Records” (2018).
2Verato Carbon™ curates 300M+ identities and covers more than 85% of the U.S. population, based on Verato’s internal platform data (as of June 2026).