OHDSI Home | Forums | Wiki | Github

AHRQ Utilization Flag


(Gowtham Rao) #1

AHRQ maintains some concepts in what may be considered AHRQ vocabulary.

e.g. AHRQ Utilization flag

CCS-Services and Procedures provides a method for classifying Current Procedural Terminology (CPT®) codes and Healthcare Common Procedure Coding System (HCPCS) codes into clinically meaningful procedure categories.
More than 9,000 CPT/HCPCS codes and 6,000 HCPCS codes are collapsed into 244 clinically meaningful categories that may be more useful for presenting descriptive statistics than are individual CPT or HCPCS codes. For example, CCS-Services and Procedures can be used to identify populations for procedure-specific studies or to develop statistical reports providing information (such as charges and length of stay) about relatively specific procedures.

They may be considered concept-ancestors of CPT4 and HCPCS, that currently have no meaningful hierarchy in OMOP Vocabulary. I am not sure if these AHRQ vocabulary should be ‘standard concepts’ or ‘classification concepts’ in OMOP vocabulary.

Source: https://www.hcup-us.ahrq.gov/toolssoftware/ccs_svcsproc/ccssvcproc.jsp

Thoughts on including this?

(Christian Reich) #2


Probably easy to include this as a new Classification Vocabulary for those CPT and HCPCS. But what I think would be even more useful is to use them for domain assignments. Can you check it out?


@Dymshyts and gang are building a comprehensive Procedure hierarchy. May not help you with your AHRQ friends, though.

(Dmytry Dymshyts) #3

here is the table itself

On a first look it seems to be useful. So we may add these concepts as classification concepts, ancestors of the corresponding CPT4 and HCPCS.

I’ll test this and return with the results soon.

(Dmytry Dymshyts) #4


Is there a rule that we decide to add something to the OHDSI vocabulary if there’s a use case only?

(Gowtham Rao) #5

Use case

CPT4 and HCPCS are vocabulary that is used to code procedures/services. The rates of these services and procedures are tracked for various reasons including health economic analysis. E.g. how often is CT abdomen performed per 1000 persons per year. Using a concept hierarchy makes standardized analytics easier. Alternative would be to specify concept sets - this approach is difficult to standardize

(Gowtham Rao) #6

The same CCS classification applies to

  1. CPT4 and HCPCS: https://www.hcup-us.ahrq.gov/toolssoftware/ccs_svcsproc/ccssvcproc.jsp

  2. Icd 10 cm procedure codes: https://www.hcup-us.ahrq.gov/toolssoftware/ccs10/ccs10.jsp (only CCS, not multi level)

I recommend, incorporating both vocabulary into the same classification vocabulary.

(Dmytry Dymshyts) #7

I also find the relationships between ICD10CM conditions to CCS CATEGORY.
Should we also add them?

(Gowtham Rao) #8


(Christian Reich) #9

Of course. Why would we otherwise do it?

(Christian Reich) #10

The problem is this: It is a CPT/HCPCS-specific categorization. The procedure hierarchy will turn CPTs and HCPCSs into source concepts, rendering these non-interoperable between countries. We don’t want that. Any good ideas?

(Gowtham Rao) #11

Could you please elaborate? Right now CPT4, HCPCS and ICD proceudre codes are mix of ‘standard’, ‘non-standard’ concept_id’s


(Christian Reich) #12

Right now. But we are building the Procedure hierarchy, and then they will be Source Concepts (which they should have been from the beginning).

(Gowtham Rao) #13

Maintained by OHDSI?

(Christian Reich) #14

Right now, yes. Will try to convince the better suited authorities (SNOMED or so) to take that over.

(Thomas White, MD, MS, MA, CHIE) #15

Re-opening this thread about CCS (since it is listed as currently unavailable in Athena, and only seems to be using the ICD-9-CM codes)

There are two Clinical Classification systems maintained by AHRQ:
(1) Diagnostic CCS - this is a multi-level hierarchy, with available CSV files mapping all ICD-9-CM and ICD-10-CM codes to the diagnostic CCS categories
(2) Procedure CCS - There are separate mappings for CPT codes and ICD-10-PCS codes.

Both CCS categories are clinically and analytically useful. It is convenient to analyze and report findings based upon CCS categories (and/or roll-up levels) rather than individual procedures or diagnoses. The CCS categories also helpful to define initial concept sets (adding exclusions as needed). In additional, it would be useful for the FeatureEngineering package to create features for each diagnostic and procedure CCS categories (e.g. as utilization and cost summaries by those categories for the short/mid/long look-back periods).

Are there plans to resume making the CCS vocabularies and mappings available?

(Gowtham Rao) #16

SNOMED is considered to offer a better parent/child hierarchical relationship for condition domain. SNOMED is a standard vocabulary for condition domain and is supported by OHDSI software.

Procedure maybe an opportunity as we dont have a great standard vocabulary.

(Christian Reich) #17


As @Gowtham_Rao said. We have discussed CCS several times, each time not making a decision to move forward. All the merits you mention are well understood. The downsides are:

  • It’s based on ICD, and we are based on SNOMED, and bolting them together will be stress
  • ICD-9-CM, ICD-10-CM, ICD-10-PCS and CPT4 are all US-based standards, and not that useful for the international community. The CMs are drastically inflated compared to the original ICDs from the WHO, and for the procedures we are working on some kind of a common hierarchy. But it’s not easy.

Happy to reopen the debate.