Healthcare technology company /

Healthcare technology company

How MEV Designed a Healthcare MDM Foundation and Golden Record Strategy for Pharmacy Claims Data

Client:

Healthcare technology company

Industry:

Healthcare / Pharmacy claims and coverage workflows

Focus:

Healthcare MDM foundation and golden record strategy for payer, PBM, sponsor, and plan entities

Technologies:

Snowflake / SQL / AWS Glue / AWS Entity Resolution

[ about the client/ ]

The client is a healthcare technology company building a platform for pharmacy claims and drug coverage lookup.

Their product needed to process large volumes of pharmacy claims data and support coverage intelligence across PBMs, insurance companies, sponsors, and plans. As more vendor sources were added, the same real-world entities could appear in different forms across datasets, creating the need for cleaner entity resolution and a trusted golden record layer.

What Needed to Be Done

The platform needed a way to reconcile duplicate and inconsistent insurance-related entities across multiple healthcare data sources.

The challenge was not only data volume. Healthcare entity data is messy by nature. The same payer, PBM, sponsor, or plan can appear under different names, formats, identifiers, abbreviations, or source-specific structures. 

MEV needed to design the foundation for a healthcare-specific MDM layer that could turn inconsistent source records into cleaner golden records: authoritative versions of payers, PBMs, sponsors, and plans that downstream lookup and reporting workflows could trust.

The foundation needed to support:

entity standardization
duplicate detection
golden record consolidation
survivorship rules
human review for uncertain matches
cleaner downstream lookup and reporting logic

The goal was to move toward a single authoritative directory of insurance-related entities, reconciled across source systems.

Why It Mattered

As the platform scaled beyond the first vendor sources, duplicate entities became harder to manage. Without a structured MDM approach, the platform could end up with several versions of the same insurance company, PBM, sponsor, or plan.

That would create downstream problems:

less reliable coverage lookup logic
inconsistent reporting
harder source onboarding
manual cleanup work
unclear relationships between payers, PBMs, sponsors, and plans
weaker governance as the platform grew

For a pharmacy coverage intelligence platform, entity consistency matters because lookup results depend on clean relationships between claims history, plan data, payer structures, and coverage information.

A golden record layer gives the platform a cleaner source of truth. Instead of treating every incoming record as equally reliable, the system can consolidate source records into authoritative entities with defined matching and survivorship logic.

What MEV Built

MEV designed the foundational approach for healthcare Master Data Management and golden record consolidation.

The target model was a single authoritative directory of payers, PBMs, sponsors, and plans, consolidated into golden records across multiple sources.

The matching strategy was designed as a staged process:

Deterministic matching using exact and canonicalized keys, identifiers, and standard code sets.
Probabilistic and fuzzy matching using string similarity, addresses, identifiers, and tunable confidence thresholds.
ML-assisted clustering for gray-zone candidates, with mandatory human review and adjudication feedback loops to improve matching quality over time.

In one stream, MEV also worked with source tables that had different schemas and transformed them into a unified golden-record structure. This helped shape the broader MDM approach around standardized entity records instead of source-specific table logic.

AWS Entity Resolution and AWS Glue were identified as planned services for the MDM layer. Governance, lineage tracking, and RBAC were designed as core requirements for the future state.

Technical Flow

01:

Source entities enter the data platform

Payer, PBM, sponsor, and plan records arrive from multiple healthcare data sources with different schemas, naming patterns, and identifiers.

02:

Entity data is standardized

Records are normalized before matching. This includes cleaning names, canonicalizing keys, aligning known identifiers, and preparing comparable fields.

03:

Deterministic rules catch clear matches

The system identifies high-confidence matches using exact keys, canonicalized values, standard code sets, and known identifiers.

04:

Fuzzy matching handles near matches

For records that do not match cleanly, the approach uses string similarity, address comparison, identifier overlap, and tunable thresholds.

05:

ML-assisted clustering supports gray zones

Ambiguous candidates can be grouped for review using ML-assisted clustering, especially where rules alone are not enough.

06:

Human review resolves uncertain cases

Domain experts review low-confidence or gray-zone matches. Their decisions feed back into the model and matching rules.

07:

Golden records become the trusted entity layer

Resolved entities are consolidated into authoritative records for payers, PBMs, sponsors, and plans, giving downstream workflows a cleaner source of truth.

Outcome

The client gained a clear foundation for healthcare-specific MDM and golden record management.

Instead of letting duplicate payer, PBM, sponsor, and plan records spread across the platform, MEV designed a staged entity resolution approach that could support cleaner data relationships as more sources were added.

This gave the platform a path toward:

more consistent insurance entity data
cleaner source onboarding
stronger governance and lineage
better downstream lookup reliability
a scalable golden record model for healthcare coverage workflows

The full implementation status should be described carefully: the case study supports saying that MEV laid the groundwork and designed the foundational MDM approach, not that the complete MDM layer was already fully implemented in production.

This case is part of the full case study

Building a Snowflake-Based Pharmacy Claims Platform for Real-Time Drug Coverage Lookup

[ portfolio/ ]