Master Unit Index¶
Software tool — instantiates Exhaustive Population Mapping
Maintains one deduplicated, versioned, access-controlled record per real unit as the registry the whole enumeration reads and writes against.
A Master Unit Index is the persistent store that holds the enumeration's answer: one authoritative record per real unit, keyed on a resolved identity, so that every source contributing to the map points at the same row for the same unit. Its defining property is that it is the single place where a unit exists exactly once — it enforces one-record-per-unit as an invariant of the store itself, versions every change so the registry has a history rather than only a present, and wraps the whole thing in access controls because a complete unit-level list is exactly the kind of legible artifact that can harm the people it names. It is not the machinery that decides whether two records are the same unit, and it is not the count; it is the durable, guarded, deduplicated ledger those processes read from and write into.
Example¶
A hospital network operating five sites runs a Master Patient Index so that a patient who was born at one hospital, treated at another, and admitted through a third emergency department is one person in the system, not three.[n1] Each source feed — registration, lab, radiology, billing — arrives with its own local identifier. The index binds them to a single enterprise record under a resolved identity key, and stores every merge and every demographic correction as a versioned event, so a mistaken merge of two different "Maria Garcia" records can be un-merged by replaying history rather than by guesswork.
The guardrail is not decoration. Because the index is the one place that can enumerate every patient, it enforces field-level access — a scheduling clerk sees contact details but not diagnoses, and every read of a flagged record is logged. When the network later needs to notify every patient exposed to a recalled device lot, the value of the index is exactly that the query returns each affected person once, with current contact information, behind an audited access path — not a pile of overlapping lists nobody can reconcile under deadline.
How it works¶
- One resolved record per unit. The index keys on an identity, not on a source's local ID, so many inbound records collapse to one enterprise unit.
- Store the resolution, don't perform it. It persists the deduplication decision handed to it; the adjudication of hard cases happens upstream and lands here as a committed result.
- Version everything. Merges, splits, and attribute corrections are recorded as events, giving the registry an auditable history and a way to reverse mistakes.
- Guard exposure at the store. Access is role-scoped and logged, because the completeness that makes the index useful is also what makes it dangerous if exposed wholesale.
Tuning parameters¶
- Match-commit threshold — how confident a match must be before two records are collapsed in the store. Set it high and duplicates persist; set it low and distinct units get wrongly merged.
- Versioning depth — full event history versus last-known-good snapshots. Deep history enables clean reversals but costs storage and complexity.
- Access granularity — table-level versus field-level permissions. Finer control protects sensitive attributes but raises administrative burden.
- Survivorship rules — when two records merge, which source's value wins per field. Good survivorship keeps the golden record current; bad rules overwrite good data with stale.
When it helps, and when it misleads¶
The index is indispensable whenever downstream action is per unit and the same unit is described by many systems — notification, billing, care continuity, reconciliation — because it is the one artifact that can answer "every unit, exactly once, with current attributes." Its versioning also makes the registry corrigible: mistakes have a history and therefore a path back.
Its characteristic failure is trusting the golden record as truth when it is only a reconciliation — merge one wrong pair and two people become one, splitting an identity that later action depends on.[n2] The store also tends to accumulate silent staleness: a record can be internally consistent and thoroughly out of date. And a complete, queryable unit registry is a standing exposure risk that a convenient "just give everyone read access" quietly amplifies. The guarding discipline is to keep merges reversible, treat the golden record as a maintained hypothesis rather than a fact, and default access closed.
How it implements the components¶
A Master Unit Index fills the durable-store slice — the artifact the archetype's other machinery reads and writes:
unit_level_registry_or_map— its core: one persistent, attributed record per unit, with identifiers, provenance, and update status.unit_identity_and_deduplication_key— it keys on a resolved identity and enforces one-record-per-unit as a store invariant (storing the key, not deciding the hard cases).privacy_and_exposure_guardrail— role-scoped, logged, field-level access controls govern who can see which units and attributes.
It stores the resolved result but does not adjudicate which look-alike records are the same unit or log the merge/split decisions — that exception_and_revision_path work is Duplicate Resolution Queue's — and it does not estimate how many units were never entered at all, which is the missingness_and_overlap_diagnostic produced by Capture-Recapture Check.
Related¶
- Instantiates: Exhaustive Population Mapping — the index is the corrigible unit-level registry the archetype exists to produce and maintain.
- Consumes: Duplicate Resolution Queue supplies the adjudicated identity decisions the index commits and stores.
- Sibling mechanisms: Census Protocol · Enumeration Area Map · Capture-Recapture Check · Door-to-Door or Field Sweep · Administrative Record Linkage · Duplicate Resolution Queue · Coverage Gap Heatmap · Late-Unit Inclusion Window · Enumeration Quality Backcheck
Editorial Notes¶
Form Classification¶
Form family: Record, Log & Register
Rationale: Master Unit Index operates as a durable record, ledger, register, or trace whose value depends on preserving actual state or history because it maintains one deduplicated, versioned, access-controlled record per real unit as the registry the whole enumeration reads and writes against.
Independent corroboration: The frozen evidence defines Master Unit Index as 'Maintains one deduplicated, versioned, access-controlled record per real unit as the registry the whole enumeration reads and writes against', so its operative form is Record, Log & Register.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Statistics & Experimental Design
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Specialized
Rationale: A master unit index is fundamentally an enumeration or sampling-frame control: one current record must represent each real unit so coverage and duplication can be measured. Authority control, database engineering, and official-register administration materially shape how that frame is maintained.
Related originating lineages:
- Computer Science & Software Engineering — Retained as a formative lineage independently identified as primary: Master unit and master patient indexes are enterprise record-linkage systems that enforce one versioned identity per real unit under controlled access.
- Library & Information Science — Authority control contributed deduplication, versioning, and stable identification.
- Public Administration & Policy — Civil registries and administrative enumeration materially shaped population-scale systems of record.
Review resolution: The U.S. Census Bureau identifies the Master Address File as the source frame from which sample units are selected and describes its continual updating. That makes statistical enumeration primary; the other retained domains explain identity control, implementation, and official-register lineage rather than downstream uses. The alternates are retained only as formative or independently established origins, not because the mechanism can be applied there. origin_mode=cross_disciplinary_synthesis states the provenance relationship; domain_reach=specialized separately records breadth because established use remains concentrated in a bounded professional context. confidence=high reflects the strength and specificity of the evidence; encyclopedia_synthesis=false because the entry generalizes an established mechanism without inventing a new composite.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
- https://www.census.gov/programs-surveys/sipp/methodology/sampling.html — Primary Census methodology identifies the Master Address File as the sampling frame and explains unit selection and updating.
Notes¶
[n1] A Master Patient Index (MPI, or enterprise MPI) is the healthcare system of record that maintains one canonical patient identity across separate registration, clinical, and billing systems — the standard institutional example of a deduplicated master registry. ↩
[n2] In master-data management, the golden record is the single reconciled version of an entity assembled from many sources. It is a construct, not an observation: its quality is only as good as the survivorship and matching rules that built it, which is why a wrong merge propagates as if it were fact. ↩