List of Different Types of Health Data

Both newcomers and experienced folks alike are challenged by the complex standard vocabularies, coding systems, semi-structured, and unstructured data in healthcare. If you don’t know what I mean yet, just take a few minutes and look at the National Library of Medicine’s impressive attempt at putting all the vocabularies in one place and mapping it all together. Its called the Unified Medical Language System® (UMLS®) Metathesaurus®  and it has over 150 classifications, code sets, thesauri, and lists of controlled terms in the biomedical domain. (I know right?)

You rarely have the opportunity to choose which type of health care data you have to use for your analysis. You simply make the best of what is available.  Usually what is available depends on the setting, purpose, and subject. Setting has to do if the data supports activities in inpatient, outpatient, nursing, etc. environments.  Purpose would be activities like clinical care, billing, and research.  Subject has to do with whether the data is about: diagnosis, medications, etc.   Keep in mind that many health data code systems and vocabularies cut across these domains.

How the data is captured and stored is also very important to how you can use it in analysis.  Health data can be structured, unstructured or semi-structured.  Structured data is captured and stored as discrete coded values. For example, a laboratory result for a patient’s Hemoglobin A1c might be stored in a database as [4548-4] [7.9]  where 4548-4 is the LOINC Code for the specific test, and 7.9 is the result value.  When you store your data in a structured way, it makes it very easy for computer systems to pass data back and forth.  Most diagnosis, labs, procedure orders, medications and many others are stored as structured data.  Structured data is the easiest to do analysis.

Unstructured data on the other hand, is captured and stored as free text. For example “The patient’s a1c was 7.9″.  Unstructured data is what we are used to reading (you are reading it now) and can be very expressive, but computers are not very good at reading English, yet. Progress notes, radiology reports, pathology results, etc. are all usually unstructured data.  Until only a few years ago, clinically captured data like blood pressure or pain scales was embedded in unstructured text.  Now, however, most EHRs will accommodate these as structured data.
Semi-Structured data is a happy mix between the two. A data entry interface may have a combination of structured data capture and free text. If the system has the ability to help structure the free text by applying dictionaries to ensure everybody uses the correct vocabulary, etc. it is even better.

Types of Structured Health Data

Data Coding Systems

Type of Information
Coding System/Standard Vocab
Sample Data
Used for
 
DiagnosisICD-9-CM250.01
V90.83
Billing, some research and population health.Being phased out in 2015 in US. Codes are usually assigned by professional medical coders after reviewing the health record.
DiagnosisICD-10-CMZ77.22
I21.02
Billing, some research and population health.Used Worldwide except in US.
Codes are assigned by medical coders.
DiagnosisMDC (Major Diagnosis Category )17
08
BillingComputed from ICD-9-CM or ICD-10-CM
DiagnosisDiagnosis-relateMS-DRGs (that’s diagnosis related groups) for Dummies groups (MS-DRG, etc.)69
242
Inpatient BillingComputed from ICD-9-CM or ICD-10-CM
DiagnosisSNOMED CT(Systematized Nomenclature of Medicine--Clinical Terms) simple chronic anemia (disorder)
pneumonia due to Staphylococcus aureus (disorder)
Problem Lists, Medical RecordAssigned by provider
ProcedureICD-9-CM44.31
76.0
Billing, some research and population health.Being phased out in 2015 in US. Codes are usually assigned by professional medical coders after reviewing the health record.
ProcedureICD-10-CM7W02X0Z
B2230ZZ
Billing, some research and population health.Used Worldwide except in US.
Codes are assigned by medical coders.
ProcedureAPC (Ambulatory Payment Classification)0370BillingComputed based on CPT and HCPCS
ProcedureCPT (Current Procedural Terminology)Billing outpatient and inpatient
ProcedureHCPCS (Healthcare Common Procedure Coding System)Billing outpatient and inpatientPronounced "hicpix". Assigned by coder and sometimes clinical staff.
MedicationNDC (National Drug Code)0067-6238 Pharmacy, Medical RecordSingle code represents drug, strength, dosage, route, packaging, etc.
MedicationRxNorm and RxCUIsPharmacy, Medical RecordAn ontology of several vocabularies.
LaboratoryLOINC (Logical Observation Identifiers Names and Codes)19254-2
11556-8
Laboratory, Medical Record
This table lists the common data coding systems used in health care

Let me know if you see any missing items, and I will add them.

Unstructured data types to be added soon.

 

If you like this post, please share it using the social links to the left. That’s how we can grow the site. Thanks!

Leave a Reply

Your email address will not be published. Required fields are marked *