HTAC Pipeline · Step 5 of 7

Condition Cohort Queries

Clinical Data
Deduplication
5
Cohort Queries
6
Suppression
7
Results

For each of the 22 HTAC health conditions, a codeset of OMOP concept codes defines who qualifies as a case. The query engine searches four clinical domains at each site — Condition, Drug, Measurement, and Observation — and returns a patient cohort for stratification.

Why use codesets?

A single diagnosis concept is rarely sufficient to identify all patients with a condition. Diabetes, for example, may appear as a SNOMED diagnosis, a Metformin prescription, or an elevated HbA1c measurement. A codeset with include and exclude codes across multiple domains captures the full clinical picture.

Query logic

For each condition, the engine builds an OR across all included domains: a patient qualifies if they have at least one matching clinical event in any domain during the study period. Excluded codes then remove patients who match an exclusion criterion.

Federated execution

The same cohort definition is executed independently at each site against that site’s local OMOP database—equivalent to shipping one approved analytic package everywhere the definition must match exactly.

Include / Exclude Logic

Step 1 — Inclusion (OR across all domains):
Person qualifies IF ConditionOccurrence.condition_concept_id IN [included_condition_codes]
OR DrugExposure.drug_concept_id IN [included_drug_codes]
OR Measurement.measurement_concept_id IN [included_measurement_codes]
OR Observation.observation_concept_id IN [included_observation_codes]
All filtered to health_system and start_date BETWEEN study_start AND study_end

Step 2 — Exclusion (removes patients matching any exclusion code):
EXCLUDE IF any matching event exists with is_excluded=True concept codes

The cohort engine applies the include and exclude rules above for each site and study window; results are materialized only when an analyst or job requests counts or extracts.

All 22 HTAC Conditions

Total concept codes in the database: 73 (58 Condition, 9 Drug, 6 Measurement, 0 Observation)

Condition Category Total codes Condition Drug Measurement Observation
Atrial Fibrillation Cardiometabolic 3 3 0 0 0
CAD Cardiometabolic 3 3 0 0 0
Heart Failure Cardiometabolic 4 3 1 0 0
Hyperlipidemia Cardiometabolic 4 2 1 1 0
Hypertension Cardiometabolic 4 2 1 1 0
Obesity Cardiometabolic 3 2 0 1 0
Diabetes Endocrine 4 2 1 1 0
Hypothyroidism Endocrine 3 2 1 0 0
HIV Infectious Disease 3 3 0 0 0
Hepatitis C Infectious Disease 3 3 0 0 0
Anxiety Mental Health 3 3 0 0 0
Bipolar Disorder Mental Health 3 3 0 0 0
Depression Mental Health 4 3 1 0 0
PTSD Mental Health 3 3 0 0 0
Schizophrenia Mental Health 3 3 0 0 0
Dementia Neurological 3 3 0 0 0
Stroke Neurological 3 3 0 0 0
CKD Renal 4 2 0 2 0
Asthma Respiratory 4 3 1 0 0
COPD Respiratory 3 2 1 0 0
Opioid Use Disorder Substance Use 3 2 1 0 0
Substance Use Disorder Substance Use 3 3 0 0 0
Total 73 58 9 6 0

Latest demonstration — federated cohort counts

Numerator / denominator (persons with a study-window visit who meet the condition definition).

Site Asthma Depression Hypertension Type 2 Diabetes Opioid Use Disorder
Allina Health 10 / 154 19 / 154 50 / 154 12 / 154 3 / 154
HealthPartners 16 / 137 23 / 137 45 / 137 14 / 137 3 / 137
M Health Fairview 13 / 145 22 / 145 44 / 145 12 / 145 4 / 145
Mayo Clinic 10 / 120 20 / 120 40 / 120 17 / 120 2 / 120
Essentia Health 5 / 103 18 / 103 36 / 103 8 / 103 1 / 103
Hennepin Healthcare 7 / 68 10 / 68 29 / 68 8 / 68 1 / 68
Sanford Health 5 / 85 11 / 85 34 / 85 9 / 85 0 / 85
CentraCare 0 / 60 15 / 60 20 / 60 6 / 60 0 / 60
Children's Minnesota 4 / 34 4 / 34 0 / 34 0 / 34 0 / 34
North Memorial Health 1 / 43 4 / 43 17 / 43 3 / 43 1 / 43
Minneapolis VA Health Care System 3 / 51 16 / 51 23 / 51 3 / 51 0 / 51
No completed demonstration yet. Run the pipeline demonstration to populate these metrics with a synthetic federated run (OMOP → PPRL → deduplication → cohorts → suppression).
← Step 4: Enrichment
← Pipeline overview
Step 6: Suppression →