You're building a healthcare data warehouse. Your team has modeled retail systems, banking platforms, even e-commerce giants. Healthcare should be easier, right?

Wrong.

Three months into the project, your carefully normalized member table can't handle mid-month plan switches. Claims get adjusted weeks after payment, breaking your financial reconciliation. Provider network attribution shows Dr. Smith treating patients in three cities simultaneously. Your quality measures calculate incorrectly because you didn't model diagnosis code sequencing.

Healthcare data modeling is fundamentally different. Here's how to build logical models for Members, Claims, Providers, and Compliance that actually work in production.

The Four Pillars of Healthcare Data Models

Every healthcare data warehouse centers on four core entity groups:

Members - People enrolled in health plans
Claims - Medical services rendered and billed
Providers - Doctors, hospitals, clinics delivering care
Compliance - Regulatory reporting, quality measures, audits

The fatal mistake? Modeling these like transactional retail data. Healthcare requires temporal tracking, version history, source lineage, and regulatory audit trails that traditional normalized models don't support.

Pillar 1: Member Data Models

Members aren't static records. A member can:

Switch plans mid-month (retroactive coverage changes)
Have multiple coverage types (medical + dental + vision + Rx)
Change addresses while claims are processing (where do you route the EOB?)
Appear in 4 different source systems with 4 different names

The Naive Member Model (Breaks Immediately)

CREATE TABLE member (
  mmbr_id          VARCHAR(50) PRIMARY KEY,
  first_nm         VARCHAR(100),
  last_nm          VARCHAR(100),
  date_of_birth    DATE,
  gender_cd        CHAR(1),
  ssn              VARCHAR(11),
  address_line1    VARCHAR(200),
  city_nm          VARCHAR(100),
  state_cd         CHAR(2),
  zip_cd           VARCHAR(10),
  plan_id          VARCHAR(50),
  coverage_start_dt DATE,
  coverage_end_dt   DATE
);

Why this fails in production:

Updates overwrite history (can't reproduce last month's metrics)
Can't handle mid-month plan changes (member had 2 plans in one month)
No source system tracking (which system is truth for address?)
No family relationships (subscriber vs. dependent)

The Production Member Model

-- Golden record (current truth)
CREATE TABLE member_master (
  mmbr_id              VARCHAR(50) PRIMARY KEY,
  subscriber_id        VARCHAR(50),
  dependent_seq_nbr    INT,
  golden_first_nm      VARCHAR(100),
  golden_last_nm       VARCHAR(100),
  golden_dob           DATE,
  golden_gender_cd     CHAR(1),
  golden_ssn           VARCHAR(11),
  mdm_effective_dttm   TIMESTAMP,
  is_current_flag      CHAR(1),
  created_dttm         TIMESTAMP,
  updated_dttm         TIMESTAMP
);

-- Demographic history (all changes)
CREATE TABLE member_demographic_history (
  demo_history_id      VARCHAR(50) PRIMARY KEY,
  mmbr_id              VARCHAR(50),
  version_nbr          INT,
  first_nm             VARCHAR(100),
  last_nm              VARCHAR(100),
  address_line1        VARCHAR(200),
  city_nm              VARCHAR(100),
  state_cd             CHAR(2),
  zip_cd               VARCHAR(10),
  
  effective_dt         DATE,
  end_dt               DATE,
  created_dttm         TIMESTAMP,
  source_system_cd     VARCHAR(20),
  is_current_flag      CHAR(1),
  
  FOREIGN KEY (mmbr_id) REFERENCES member_master(mmbr_id)
);

-- Enrollment/coverage (temporal)
CREATE TABLE member_enrollment (
  enrollment_id        VARCHAR(50) PRIMARY KEY,
  mmbr_id              VARCHAR(50),
  plan_id              VARCHAR(50),
  coverage_type_cd     VARCHAR(20),
  coverage_start_dt    DATE,
  coverage_end_dt      DATE,
  enrollment_status_cd VARCHAR(20),
  termination_reason_cd VARCHAR(50),
  premium_amt          DECIMAL(10,2),
  
  effective_dt         DATE,
  end_dt               DATE,
  is_current_flag      CHAR(1),
  
  FOREIGN KEY (mmbr_id) REFERENCES member_master(mmbr_id)
);

-- Member months (pre-calculated)
CREATE TABLE member_months (
  mmbr_month_id        VARCHAR(50) PRIMARY KEY,
  mmbr_id              VARCHAR(50),
  plan_id              VARCHAR(50),
  year_nbr             INT,
  month_nbr            INT,
  coverage_days        INT,
  member_months        DECIMAL(5,2),
  full_month_flag      CHAR(1),
  calculation_method   VARCHAR(50),
  
  FOREIGN KEY (mmbr_id) REFERENCES member_master(mmbr_id)
);

Why this works:

✅ Golden record preserves current state
✅ Full demographic history for point-in-time queries
✅ Enrollment periods tracked separately
✅ Member months pre-calculated (no runtime debates)
✅ Source lineage documented

Pillar 2: Claims Data Models

Claims are the most complex healthcare entity:

Get submitted → adjudicated → paid → adjusted → reprocessed
Have 50+ attributes (diagnosis codes, procedure codes, modifiers, revenue codes)
Link to multiple providers (billing, rendering, referring, facility)
Must reconcile to EDI 835 remittance files
Require complete audit trail for CMS audits

The Naive Claims Model

CREATE TABLE claim (
  clm_id               VARCHAR(50) PRIMARY KEY,
  mmbr_id              VARCHAR(50),
  prvdr_id             VARCHAR(50),
  service_dt           DATE,
  total_billed_amt     DECIMAL(10,2),
  total_paid_amt       DECIMAL(10,2),
  claim_status_cd      VARCHAR(20)
);

Why this fails:

No adjustment tracking (claims change weeks after payment)
No line-level detail (claims can have 100+ service lines)
No diagnosis codes (needed for quality measures, risk adjustment)
No provider role tracking (who billed vs. who rendered?)
No reconciliation to financial ledger

The Production Claims Model

-- Claim master (current state)
CREATE TABLE claim_master (
  clm_id               VARCHAR(50) PRIMARY KEY,
  mmbr_id              VARCHAR(50),
  plan_id              VARCHAR(50),
  original_clm_nbr     VARCHAR(50),
  payer_clm_nbr        VARCHAR(50),
  
  service_from_dt      DATE,
  service_to_dt        DATE,
  received_dt          DATE,
  processed_dt         DATE,
  paid_dt              DATE,
  
  total_billed_amt     DECIMAL(10,2),
  total_allowed_amt    DECIMAL(10,2),
  total_paid_amt       DECIMAL(10,2),
  total_patient_resp_amt DECIMAL(10,2),
  
  claim_status_cd      VARCHAR(20),
  claim_type_cd        VARCHAR(20),
  current_version_nbr  INT,
  adjustment_count     INT,
  
  is_current_flag      CHAR(1),
  created_dttm         TIMESTAMP,
  updated_dttm         TIMESTAMP,
  
  FOREIGN KEY (mmbr_id) REFERENCES member_master(mmbr_id)
);

-- Claim version history
CREATE TABLE claim_version_history (
  clm_version_id       VARCHAR(50) PRIMARY KEY,
  clm_id               VARCHAR(50),
  version_nbr          INT,
  
  total_billed_amt     DECIMAL(10,2),
  total_paid_amt       DECIMAL(10,2),
  claim_status_cd      VARCHAR(20),
  
  adjustment_type_cd   VARCHAR(20),
  adjustment_reason_txt TEXT,
  
  effective_dt         DATE,
  end_dt               DATE,
  version_created_dttm TIMESTAMP,
  is_current_flag      CHAR(1),
  
  FOREIGN KEY (clm_id) REFERENCES claim_master(clm_id)
);

-- Claim line detail
CREATE TABLE claim_line (
  clm_line_id          VARCHAR(50) PRIMARY KEY,
  clm_id               VARCHAR(50),
  line_nbr             INT,
  
  service_dt           DATE,
  place_of_service_cd  VARCHAR(10),
  revenue_cd           VARCHAR(10),
  procedure_cd         VARCHAR(20),
  modifier_cd_1        VARCHAR(10),
  units_qty            DECIMAL(10,2),
  
  billed_amt           DECIMAL(10,2),
  allowed_amt          DECIMAL(10,2),
  paid_amt             DECIMAL(10,2),
  deductible_amt       DECIMAL(10,2),
  coinsurance_amt      DECIMAL(10,2),
  copay_amt            DECIMAL(10,2),
  
  rendering_prvdr_id   VARCHAR(50),
  line_status_cd       VARCHAR(20),
  denial_reason_cd     VARCHAR(50),
  
  FOREIGN KEY (clm_id) REFERENCES claim_master(clm_id)
);

-- Claim diagnosis codes
CREATE TABLE claim_diagnosis (
  clm_diag_id          VARCHAR(50) PRIMARY KEY,
  clm_id               VARCHAR(50),
  diagnosis_seq_nbr    INT,
  diagnosis_cd         VARCHAR(20),
  diagnosis_type_cd    VARCHAR(20),
  present_on_admit_cd  CHAR(1),
  
  FOREIGN KEY (clm_id) REFERENCES claim_master(clm_id)
);

-- Claim provider relationships
CREATE TABLE claim_provider (
  clm_prvdr_id         VARCHAR(50) PRIMARY KEY,
  clm_id               VARCHAR(50),
  prvdr_id             VARCHAR(50),
  provider_role_cd     VARCHAR(20),
  
  FOREIGN KEY (clm_id) REFERENCES claim_master(clm_id)
);

Why this works:

✅ Version history preserves all adjustments
✅ Line-level detail captured
✅ Diagnosis codes stored with sequence
✅ Multiple provider roles documented
✅ Complete audit trail for compliance

Pillar 3: Provider Data Models

Providers have:

Multiple identifiers (NPI, state license, DEA, PTAN)
Taxonomy classifications (specialty codes)
Practice locations (primary + satellites)
Network participation (contract dates, reimbursement rates)
Credentialing status (active, suspended, under review)

The Production Provider Model

-- Provider master
CREATE TABLE provider_master (
  prvdr_id             VARCHAR(50) PRIMARY KEY,
  npi_id               VARCHAR(10) UNIQUE,
  tax_id               VARCHAR(20),
  prvdr_type_cd        VARCHAR(20),
  
  last_nm              VARCHAR(100),
  first_nm             VARCHAR(100),
  organization_nm      VARCHAR(200),
  
  primary_taxonomy_cd  VARCHAR(20),
  primary_specialty_cd VARCHAR(50),
  
  is_active_flag       CHAR(1),
  created_dttm         TIMESTAMP,
  updated_dttm         TIMESTAMP
);

-- Provider practice locations
CREATE TABLE provider_location (
  prvdr_location_id    VARCHAR(50) PRIMARY KEY,
  prvdr_id             VARCHAR(50),
  location_type_cd     VARCHAR(20),
  address_line1        VARCHAR(200),
  city_nm              VARCHAR(100),
  state_cd             CHAR(2),
  zip_cd               VARCHAR(10),
  phone_nbr            VARCHAR(20),
  
  is_primary_flag      CHAR(1),
  effective_dt         DATE,
  end_dt               DATE,
  
  FOREIGN KEY (prvdr_id) REFERENCES provider_master(prvdr_id)
);

-- Network participation
CREATE TABLE provider_network (
  prvdr_network_id     VARCHAR(50) PRIMARY KEY,
  prvdr_id             VARCHAR(50),
  plan_id              VARCHAR(50),
  network_id           VARCHAR(50),
  participation_type_cd VARCHAR(20),
  contract_start_dt    DATE,
  contract_end_dt      DATE,
  reimbursement_type_cd VARCHAR(20),
  
  network_status_cd    VARCHAR(20),
  termination_reason_cd VARCHAR(50),
  is_accepting_new_patients_flag CHAR(1),
  
  FOREIGN KEY (prvdr_id) REFERENCES provider_master(prvdr_id)
);

Pillar 4: Compliance & Quality Models

Healthcare requires regulatory reporting:

HEDIS measures
Stars ratings
Risk adjustment (HCC)
Prior authorization tracking

-- Quality measure definitions
CREATE TABLE quality_measure (
  measure_id           VARCHAR(50) PRIMARY KEY,
  measure_cd           VARCHAR(20),
  measure_nm           VARCHAR(200),
  measure_desc         TEXT,
  measure_category_cd  VARCHAR(50),
  
  measurement_year     INT,
  numerator_logic      TEXT,
  denominator_logic    TEXT,
  exclusion_logic      TEXT,
  
  is_active_flag       CHAR(1)
);

-- Member measure compliance
CREATE TABLE member_quality_measure (
  mmbr_measure_id      VARCHAR(50) PRIMARY KEY,
  mmbr_id              VARCHAR(50),
  measure_id           VARCHAR(50),
  measurement_year     INT,
  
  in_denominator_flag  CHAR(1),
  in_numerator_flag    CHAR(1),
  is_excluded_flag     CHAR(1),
  
  gap_exists_flag      CHAR(1),
  gap_close_dt         DATE,
  
  calculated_dttm      TIMESTAMP,
  
  FOREIGN KEY (mmbr_id) REFERENCES member_master(mmbr_id),
  FOREIGN KEY (measure_id) REFERENCES quality_measure(measure_id)
);

The Integration Layer

Your model must explicitly connect:

Member → Claims via mmbr_id
Claims → Providers via claim_provider
Claims → Quality via diagnosis/procedure codes
Providers → Network via provider_network
Members → Enrollment via member_enrollment

Critical Query: Member Attribution

SELECT 
  m.mmbr_id,
  m.golden_first_nm,
  p.npi_id,
  p.organization_nm
FROM member_master m
JOIN member_enrollment me ON m.mmbr_id = me.mmbr_id
  AND me.is_current_flag = 'Y'
JOIN provider_network pn ON me.plan_id = pn.plan_id
  AND pn.participation_type_cd = 'PCP'
  AND pn.network_status_cd = 'ACTIVE'
JOIN provider_master p ON pn.prvdr_id = p.prvdr_id
WHERE m.is_current_flag = 'Y';

Compliance Requirements

Every table needs:

Audit trail: created_by, created_dttm, updated_by, updated_dttm
Temporal validity: effective_dt, end_dt, is_current_flag
Source lineage: source_system_cd, source_file_id, load_dttm

Conclusion

Healthcare data models must handle:

Temporal complexity - Data changes over time
Regulatory compliance - Audit trails required
Multi-source integration - Conflicting data
Reconciliation - Must match external sources

Stop modeling healthcare like retail. Model for healthcare reality.

Checklist

Members:

Demographic history
Enrollment periods
Member months calculation
Source lineage

Claims:

Version tracking
Line-level detail
Diagnosis codes
Provider roles
EDI reconciliation

Providers:

NPI identifier
Practice locations
Network participation
Credentialing status

Compliance:

Quality measures
Audit trails
Temporal validity

Check every box or your model breaks in production.

mdatool.

Logical Data Models in Healthcare: Members, Claims, Providers, and Compliance