• 01/07/2022

What are data standards and why do we need them?

Data standards - as the name suggests are agreed upon guidelines for recording /storing data elements or datasets. Adhering to data standards makes the sharing and exchange of data easy and streamlined. Let's look at an example to understand the value of data standards. Here is a hypothetical case where three hospital systems are following three different "standards" for storing the 'Gender' of a patient.

Now, imagine you are an IT company that has been tasked to develop a claim adjudication engine which is compatible with all three hospital systems above. Though it's not impossible to build such a system, however, not having data standards makes the task challenging. Also, lack of data standards is potentially a source for errors when combining data across multiple systems, e.g., if the developer assumes that 0/1 representation in case of Hospital B is same as that of Hospital A then this will lead to a huge systemic issue.

Key data standards in health informatics

There are six key data standards in healthcare:

1. Systematized Nomenclature for Medicine - Clinical Terms (SNOMED CT)
2. International Classification of Diseases (ICD-10)
3. Current Procedural Terminology (CPT)
4. Logical Observation Identifiers Names and Codes (LOINC)
5. National Drug Code (NDC)
6. RxNorm
Briefly, SNOMED CT is the most comprehensive, multilingual clinical healthcare terminology in the world. ICD and CPT are widely used for medical billing. LOINC provides significant details about clinical tests. NDC is a US-specific standard for medications maintained by the US Food and Drug Administration (FDA). RxNorm is US-specific terminology in medicine that contains all medications available on the US market.

Deep-dive into SNOMED CT [Systematized Nomenclature for Medicine - Clinical Terms]

The development of SNOMED Clinical Terms traces its roots to a project begun in the 1960s at National Institutes of Health (NIH) to use natural language processing (NLP) to machine code pathologists’ free text dictated notes. SNOMED CT is the most comprehensive, multilingual clinical healthcare terminology in the world. It is essentially an ontology representing relationships among its concepts. SNOMED CT has 350,000+ concepts with over 1.3 million relationships (a relationship is an association between a source concept and a destination concept). Given its scope it is heavily referenced in FHIR Resources.
Let's look at SNOMED CT through an example:

Link to SNOMED CT browser:

If we search for the term ‘hypertension’ in the SNOMED CT browser we get 379 matches. The first match is ‘hypertension’ with an FSN of ‘Hypertensive disorder, systemic arterial (disorder)’.

Now, if we click on ‘Hypertension’ we see the concept has an SCTID is 38,341,003. Below it we see that SNOMED CT recognizes 14 synonyms for this disorder. These synonyms are useful in processing text notes that might contain one or more of these alternative terms.
Further we see the related parent and children concepts for ‘hypertension’ in the right panel - the hierarchical nature of SNOMED CT is quite visible here.
Hypertension is a child of ‘Disorder of cardiovascular system (disorder)’. Hypertension, in turn, has 25 more specific sub-disorders called children. Such a parent-child hierarchy can be useful for grouping patients for analysis.

The browser also provides a diagrammatic hierarchical view (as shown below). It shows how SNOMED CT explicitly reveals important computable clinical relationships. For example: using this hierarchy we can build a computational rule that hypertension is associated with an increased blood pressure.

It is evident that SNOMED CT can be a very valuable tool for usefully grouping and analyzing patients and their medical conditions. Given the importance of SNOMED CT, HL7 has developed a SNOMED CT Implementation Guide for FHIR to show developers, analysts and system architects features of FHIR which can be used to their full potential. The guide also highlights several potential issues that should be carefully considered along with appropriate steps to make to avoid them. The goal the implementation guide is “to maximize both the rich expressivity of SNOMED and the semantic operability when exchanging FHIR Resources between systems”.

Interoperability: FHIR and SNOMED CT

SNOMED CT can be thought of as a dictionary, specifying exactly what medical concepts mean. It does not, however, tell you how to assemble those concepts within a medical record in a way that can be transmitted to another system and understood. An information model is required to achieve this; a standard for how records and messages should be laid out and populated. This role is being played globally by HL7 FHIR. By combining these two world leading standards, systems using SNOMED CT and FHIR together are able to communicate clear and unambiguous meaning in a standard way that can be automatically understood worldwide.

We shall cover other data standards in subsequent posts!


This article has been inspired by Mark L. Braunstein's book: "Health Informatics on FHIR: How HL7's API is Transforming Healthcare (Second Edition)"