Healthcare data domain
Skill hollandkevint/data-product-operator/skills/healthcare-data-domain
Healthcare data domain context covering FHIR, HL7, OMOP CDM, real-world evidence, and clinical terminology systems. Use when working on clinical data pipelines, EHR integrations, claims data products, HIPAA-governed data, OMOP transformations, or when the conversation involves PHI, ICD-10, SNOMED, CPT, LOINC, or RxNorm. Skip this skill for non-healthcare data products.From its SKILL.md
npx -y skills add hollandkevint/data-product-operator --skill healthcare-data-domainAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
5.0 KB, ~1.1k tokens by cl100k_base, as published. Nobody here has run it
When This Skill Applies
Activate when the data product involves: electronic health records (EHR), claims/billing data, clinical terminology, patient-level data, OMOP CDM, FHIR/HL7, or any data governed by HIPAA.
Do NOT activate for: general analytics, marketing data, financial data, or non-clinical datasets.
Core Standards
FHIR (Fast Healthcare Interoperability Resources) - Modern API standard for health data exchange. Resource-based (Patient, Observation, MedicationRequest, Condition, Encounter). Use for real-time integrations and patient-facing apps.
HL7 v2 - Legacy messaging standard still dominant in hospital systems. Pipe-delimited segments (MSH, PID, OBX). Expect to encounter this in any EHR integration project.
OMOP CDM (Common Data Model) - Research-optimized schema for observational health data. Core tables: PERSON, VISIT_OCCURRENCE, CONDITION_OCCURRENCE, DRUG_EXPOSURE, MEASUREMENT, OBSERVATION, PROCEDURE_OCCURRENCE. Use for analytics and real-world evidence studies.
Clinical Terminology Systems
| System | What It Codes | Example |
|---|---|---|
| ICD-10-CM | Diagnoses | F32.1 (Major depressive disorder, single episode, moderate) |
| CPT | Procedures | 99213 (Office visit, established patient) |
| SNOMED CT | Clinical concepts | 73211009 (Diabetes mellitus) |
| LOINC | Lab tests/observations | 2345-7 (Glucose, serum/plasma) |
| RxNorm | Medications | 197361 (Sertraline 50mg tablet) |
| NDC | Drug packages | National Drug Code for specific manufacturer/package |
ALWAYS use standard terminology codes rather than free-text descriptions. Map to the appropriate code system for the use case.
OMOP Analytics Patterns
The 30-40 normalized OMOP tables are wrong for analytics dashboards. Transform to star schema with strategic denormalization:
- Fact tables: Drug exposures, visits, conditions (events become facts)
- Dimension tables: Patient, drug, diagnosis
- Pre-calculated cohort definitions for common queries
- NEVER fully denormalize (One Big Table). Healthcare's many-to-many relationships cause exponential row growth.
See domain-reference.md for detailed OMOP table relationships and FHIR resource mappings.
FHIR Gotchas
Common mistakes when working with FHIR resources:
| Gotcha | What Trips You Up | Fix |
|---|---|---|
| Coding vs CodeableConcept | Coding is a single code. CodeableConcept wraps multiple codings with a display text. Most FHIR fields use CodeableConcept. | Always access .coding[0].code, not .code directly. |
| Patient.identifier vs Patient.id | .id is the FHIR server's internal ID. .identifier holds MRNs, SSNs, and other business identifiers. | Query by .identifier.value with the correct .identifier.system. |
| Observation.value[x] | Polymorphic field. Could be valueQuantity, valueString, valueCodeableConcept, or others. | Check the resource profile or test data to know which type your source sends. |
| Bundle pagination | Search results return pages of 20-50 resources. The full result set requires following Bundle.link where relation = "next". | Always paginate. NEVER assume a single Bundle contains all results. |
FHIR-to-OMOP Mapping
When transforming FHIR resources into OMOP CDM:
| FHIR Resource | OMOP Table | Key Mapping Notes |
|---|---|---|
| Patient | person | Map Patient.birthDate → year_of_birth. Gender codes differ between systems. |
| Condition | condition_occurrence | Condition.code → condition_concept_id via SNOMED-to-OMOP vocabulary mapping. |
| Observation | measurement | Lab results map here. Use LOINC code from Observation.code for measurement_concept_id. |
| MedicationRequest | drug_exposure | Map RxNorm codes. MedicationRequest.dosageInstruction → dose_value/dose_unit. |
| Encounter | visit_occurrence | Encounter.class → visit_concept_id. Map inpatient/outpatient/emergency. |
Common LOINC Codes for Vitals
| Vital Sign | LOINC Code | Units |
|---|---|---|
| Blood pressure, systolic | 8480-6 | mmHg |
| Blood pressure, diastolic | 8462-4 | mmHg |
| Heart rate | 8867-4 | /min |
| Body temperature | 8310-5 | Cel |
| Body weight | 29463-7 | kg |
| Body height | 8302-2 | cm |
HIPAA Awareness
CRITICAL: Any data product handling patient data must consider the 18 HIPAA identifiers. See domain-reference.md for the full list. De-identification is required before data leaves a HIPAA-governed environment.
This skill provides general domain context, not compliance advice. Involve your privacy officer and legal team for HIPAA compliance decisions.
What ships with it: 1 file
3.5 KB alongside SKILL.md
- domain-reference.md3.5 KB
Gives 0 of the 12 instructions most healthcare skills give in ~1.1k tokens
Counted across 147 of the 152 authors here whose files we hold, read 2026-08-07
- Export trial data to CSV formatin 11 of 147, across 2 files
- Retrieve trial details using an NCT IDin 11 of 147, across 2 files
- Split clinical datasets strictly by patientin 11 of 147, across 3 files
- Use the ClinicalTrials.gov API v2in 10 of 147, across 1 file
- Search trials by condition, drug, location, status or phasein 10 of 147, across 1 file
- Use maximum page size for bulk data retrievalin 10 of 147, across 1 file
- Extract and summarize key study informationin 10 of 147, across 1 file
- Combine multiple filters for targeted searchesin 10 of 147, across 1 file
- Print and review dataset statistics before modelingin 8 of 147, across 1 file
- Start model development with simple baselinesin 8 of 147, across 1 file
- Match preprocessing processors directly to data typesin 8 of 147, across 1 file
- Monitor validation metrics for task type and class imbalancein 8 of 147, across 1 file
Said here and by no other author read
- activate for electronic health records, claims, terminology, or HIPAA-governed data
- use standard terminology codes rather than free-text descriptions
- map healthcare codes to the appropriate code system
- transform normalized OMOP tables to a star schema
- create fact tables for healthcare events
- pre-calculate cohort definitions for common queries
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.