Section 2.2: Code Sets and Terminologies

2.2.1 The “Why”: From Ambiguity to Precision

In your pharmacy practice, you are a master of managing ambiguity. A patient says they take “the little blue water pill.” A doctor’s office faxes a handwritten prescription for “MS 15mg PO Q4H PRN.” You use your clinical knowledge, experience, and contextual clues to translate these ambiguous statements into a precise, verifiable medication order. You are the human engine of interoperability. But this process, while a testament to your skill, is not scalable and is fraught with inherent risk. What if “the little blue water pill” was triamterene/HCTZ and not furosemide? What if “MS” was misinterpreted as magnesium sulfate instead of morphine sulfate?

The entire field of health informatics is, in many ways, an attempt to solve this single, monumental problem: how can we represent clinical information in a way that is completely unambiguous to a computer? If we can solve this, we can build systems that automatically check for errors, identify trends, and share information seamlessly between providers, hospitals, and even countries, without the risk of misinterpretation. This is the “why” behind standardized code sets and terminologies. They are the agreed-upon dictionaries and grammar books that allow different computer systems to have a meaningful conversation about a patient’s health.

Think of these terminologies not as a technical burden, but as the most powerful safety tool in your informatics arsenal. They are the foundation upon which every advanced feature is built. An automated allergy check is impossible if the EHR doesn’t have a specific code for “penicillin.” A report on A1c control in diabetic patients is meaningless if one system records the lab test with one name and a second system uses another. In the previous section, we learned about the “filing cabinets” (databases) where data is stored. In this section, we will learn about the precise, standardized labels that must be on every file within those cabinets to make the system work. Mastering these “languages”—RxNorm for drugs, SNOMED CT for clinical ideas, LOINC for labs, and ICD-10 for diagnoses—is the transition from being a skilled user of health information to being an architect of it.

2.2.2 RxNorm: The Universal Language of Medications

We begin with the terminology that is most central to your expertise: RxNorm. Developed and maintained by the National Library of Medicine (NLM), RxNorm is the designated standard for medications in the United States. Its purpose is to be a normalized naming system for clinical drugs and to support semantic interoperability. In simpler terms, RxNorm’s job is to make sure that “Lisinopril 10mg Oral Tablet,” “Prinivil 10mg Oral Tablet,” and “Zestril 10mg Oral Tablet” can all be understood by a computer as representing the same fundamental clinical concept.

The RxNorm Concept Unique Identifier (RxCUI)

The heart of RxNorm is the RxCUI, a unique integer that serves as the identifier for a single concept. Every concept in RxNorm, whether it’s an ingredient, a brand name, or a specific clinical drug, has its own unique RxCUI. This is the “UPC” we discussed. When two systems are exchanging medication information, they don’t need to worry about name variations. If they both send the RxCUI, they are guaranteed to be talking about the same thing.

The Building Blocks: RxNorm Atom Types (TTYs)

RxNorm achieves its power by breaking down the idea of a “drug” into its component parts, or “atoms.” Each atom has a specific Term Type (TTY). Understanding these TTYs is the key to mastering RxNorm. While there are many TTYs, we will focus on the most important ones for a pharmacist.

Masterclass Table: Core RxNorm Term Types (TTYs)

TTY	Term Type Name	Description	Pharmacist’s Viewpoint & Example
IN	Ingredient	The active ingredient of a drug, without any strength or dose form information. This is the highest level of abstraction.	This is what you use for drug-class based alerting. Example: Lisinopril [RxCUI: 29046]. An alert for “ACE Inhibitors” would be built on a list of IN concepts.
SCD	Semantic Clinical Drug	The most important concept for clinical use. It represents the Ingredient + Strength + Dose Form. It is brand-agnostic.	This is the level at which most clinical decisions are made and most orders are written. It is the concept of a “generic” drug. Example: Lisinopril 10 MG Oral Tablet [RxCUI: 311044].
BN	Brand Name	A commercial brand name for a drug. It is linked to the ingredient(s).	This is the name the patient or provider might use. Example: Prinivil [RxCUI: 10582]. RxNorm knows that Prinivil’s ingredient is Lisinopril.
SBD	Semantic Branded Drug	This represents a specific branded product: Brand Name + Ingredient + Strength + Dose Form.	This is the brand-name equivalent of an SCD. It’s how you link a specific brand to its generic clinical concept. Example: Prinivil 10 MG Oral Tablet [RxCUI: 312563].
GPCK	Generic Pack	A non-branded drug package, typically containing multiple medications.	This is for combination products or kits. Example: Tri-Lo-Marzia (a multi-phasic oral contraceptive pack).
BPCK	Branded Pack	A branded drug package, such as a manufacturer’s retail box.	This is the specific package you would dispense. Example: Z-Pak [RxCUI: 214197], which contains specific tablets of Azithromycin.

2.2.3 SNOMED CT: The “Google” of Clinical Medicine

If RxNorm is the comprehensive dictionary for medications, then SNOMED CT (Systematized Nomenclature of Medicine – Clinical Terms) is the encyclopedia for everything else in healthcare. It is a massive, multinational, and multilingual terminology designed to represent virtually any clinical idea, from a patient’s complaint to a complex surgical procedure. While you, as a pharmacist, will live primarily in the world of RxNorm, you must understand SNOMED because it provides the essential clinical context for every medication order.

Why is a patient on lisinopril? Because they have the SNOMED concept “Hypertension.” Why was an EpiPen ordered? Because the patient has a history of “Anaphylaxis” (a SNOMED concept) due to an “Allergy to peanuts” (another SNOMED concept). SNOMED provides the structured language to document these clinical realities in a way a computer can understand and act upon.

The Three Pillars of SNOMED CT

SNOMED’s architecture is built on three core components:

• Concepts: A concept is a unique clinical idea, represented by a numeric SNOMED CT ID. There are over 350,000 concepts, organized into hierarchies like “Clinical finding,” “Procedure,” “Body structure,” and “Substance.”
• Descriptions: Each concept has multiple descriptions, or text-based ways of referring to it. For example, the concept for “Myocardial infarction” has descriptions like “Heart attack” (a synonym) and “MI” (an abbreviation). This allows clinicians to use familiar language while the system stores the data under a single, unambiguous concept ID.
• Relationships: This is where SNOMED’s true power lies. Concepts are linked to each other through a network of defined relationships. For example, the concept “Viral pneumonia” has an is_a relationship to “Pneumonia” and a causative_agent relationship to “Virus.” This rich network allows for very sophisticated data queries (e.g., “Find all patients with any type of infectious lung disease”).

Why SNOMED Matters for Medication Safety: Structuring Allergy Data

The “free text allergy” problem we discussed in the previous section is solved by SNOMED. A properly designed EHR doesn’t just have a text box for allergies. It uses SNOMED to capture three discrete pieces of information:

1. The Substance: The specific allergen is coded using a SNOMED concept from the “Substance” hierarchy (e.g., the concept for “Penicillin”).
2. The Reaction: The specific manifestation is coded using a SNOMED concept from the “Clinical finding” hierarchy (e.g., “Hives,” “Anaphylaxis,” “Nausea”).
3. The Criticality: A flag indicating the severity (e.g., Mild, Moderate, Severe).

When this structured data is captured, you can build incredibly powerful and precise CDS rules. You can build a rule that fires a “hard stop” alert if a patient with a documented “Anaphylaxis” reaction to “Penicillin” is ordered any drug belonging to the penicillin class. You can build a different, less severe “warning” alert if a patient with a “Nausea” reaction is ordered the same drug. This level of nuanced safety is impossible without the granularity provided by SNOMED CT.

2.2.4 LOINC & ICD-10: Coding the Observations and Diagnoses

Rounding out our core set of terminologies are LOINC, which covers the “what did we find?” part of medicine, and ICD-10, which covers the “what do we call it for billing?” part.

LOINC: The Universal Language of Labs

Logical Observation Identifiers Names and Codes (LOINC) is the standard for lab tests, clinical measurements (like vital signs), and other observations. Its purpose is to solve the same problem RxNorm solves for drugs: a hospital’s internal lab system might call a test “K+”, “Serum Potassium”, or “Potassium, S”. LOINC provides a single, universal code that represents the exact same test regardless of who is performing it. Every LOINC code is defined by 6 main parts that leave no room for ambiguity.

Masterclass Table: Deconstructing a LOINC Code

Let’s deconstruct the LOINC code for a common and critically important lab test for pharmacists: the INR for a patient on warfarin.

Part	Description	Example Value for INR Test	Pharmacist’s Interpretation
Component	What is being measured?	INR	We are measuring the International Normalized Ratio.
Property	What characteristic of the component is being measured?	Tme Rto (Time Ratio)	We are measuring a ratio based on time.
Time Aspect	Is it a point-in-time measurement or over a duration?	Pt (Point in time)	This is a snapshot, not a 24-hour collection.
System	What is the specimen or subject?	BldP (Blood, Platelet poor plasma)	The test is performed on platelet poor plasma from a blood sample.
Scale	How is the result expressed? Quantitative, Qualitative, etc.	Qn (Quantitative)	The result will be a number (e.g., 2.5).
Method	What technique was used? (Optional)	Coag (Coagulation)	This was a coagulation method.
Resulting LOINC Code for INR: 6301-6 INR in Platelet poor plasma by Coagulation method

Why this matters: By using the LOINC code 6301-6, an anticoagulation management clinic can electronically receive lab results from five different commercial labs and have their system automatically interpret every single one correctly, triggering alerts for sub- or supra-therapeutic values without any human intervention or risk of misinterpretation.

ICD-10-CM: The Language of Billing and Classification

Finally, we have the International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM). While SNOMED is used to capture the rich detail of a clinical encounter, ICD-10 is used to classify that encounter into a standardized category for the purposes of billing, public health surveillance, and statistical analysis. It is the code that goes on the insurance claim.

As a pharmacist, you won’t typically be assigning ICD-10 codes, but you will constantly be working with them. They are often required on prior authorization forms to justify the medical necessity of a drug. A patient’s list of ICD-10 codes provides a concise, if not clinically detailed, summary of their major health problems.

2.2.5 The Grand Unification: A Patient Encounter in Four Languages

The true power—and complexity—of health informatics comes from how these terminologies work together to paint a complete, structured picture of a patient’s care. No single terminology is sufficient on its own. They are designed to be a complementary ecosystem. Let’s walk through a single, common clinical scenario and see how each code set plays its critical role.