[CQM-2220] Diagnostic Study ECG Results Difficult to Capture Created: 09/30/16  Updated: 05/26/17  Resolved: 05/26/17

Status: Closed
Project: eCQM Issue Tracker
Component/s: Measure

Type: Hosp Inpt eCQMs - Hospital Inpatient eCQMs Priority: Minor
Reporter: Joelencia Leflore Assignee: Joelencia Leflore
Resolution: Answered Votes: 3
Labels: CRP

Attachments: Microsoft Word CQM-2220_Jira comment-response_1.9.17.docx    
Issue Links:
Duplicate
is duplicated by CQM-2019 Diagnostic Study ECG Results Difficul... Closed
Solution: Implementers have expressed difficulty in capturing ECG results through EHR data. An ECG result is generally in a non-discrete field and therefore not easily retrieved electronically, whereas a diagnosis based on the ECG result would more likely be coded as such. So, using a data type of Diagnosis instead of Diagnostic Study might more closely match how the information is recorded. Therefore, we propose adding a principal diagnosis attribute of Acute or Evolving MI to the denominator logic and removing the Diagnostic Study, Performed: Electrocardiogram (ECG) (result: Acute or Evolving MI) data type.

Additionally, because the ‘Acute or Evolving MI’ value set does not include non-STEMI codes, the logic statement that checks for non-STEMIs through the STEMI Exclusions value set is redundant. Limiting the patients to only those with a diagnosis code from the ‘Acute or Evolving MI’ value set negates the additional logic statement to exclude those with a code from the ‘STEMI Exclusions’ value set. Therefore, we also recommend removing the Diagnostic Study, Performed: Electrocardiogram (ECG) (result: STEMI Exclusions) data type to reduce redundancy in the logic. This would also include removal of the STEMI Exclusions grouping value set (2.16.840.1.113762.1.4.1045.36).
Solution Posted On:
2017 Reporting Period EH eCQMs:
CMS53v5/NQF0163
Comment Posted On:
CRP Overview: Based on our interpretation of the information provided in CQM-2019, we believe the issue is how the result (intrepretation) of an electrocardiogram (ECG) is captured and stored in an electronic health record (EHR). It is our belief that this interpretation is documented as free text and not in a discrete field, and further, there may be more than one interpretation of that ECG.

 

We also believe there will be a diagnosis recorded in the EHR based on the interpretation of the ECG that will be codified and retrievable by electronic query. What is not clear is when that diagnosis is documented in relationship to the ECG interpretation. Clinically and based on the current logic construct, there is in general a linear flow of events:

1) An admission

2) An ECG and interpretation

3) Treatment if indicated based on the ECG interpretation

It is fluid as to when the diagnosis would be recorded. It may occur at any point, from concurrent with the ECG, up to and including discharge. However, since review of these records is retrospective, we believe the actual impact of when the diagnosis is recorded (documented) is minimized as long as it is recorded in EHR for the specific encounter. Additionally, QDM will allow us to associate this diagnosis to the encounter. Therefore, we propose modifying the current Denominator logic from:

•AND: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG) (result: Acute or Evolving MI)"
•AND NOT: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG) (result: STEMI Exclusions)"

To (option 1):

•AND: Intersection of:
◦Occurrence A of $EncounterInpatient



Assessment

The initial population includes all patients age 18 or older who have an inpatient encounter less than or equal to 120 days with a principal diagnosis of AMI using the Acute Myocardial Infarction (AMI) value set. This is the first pass to determine the available pool of cases. The denominator then further narrows the pool of cases to the diagnosis of interest by only including cases that (1) had an ECG with an interpretation of Acute or Evolving MI using the Acute or Evolving MI value set, and (2) are ST-segment elevation MIs (STEMI) (using the STEMI exclusions value set to exclude those patients with a non-ST segment elevation MI).

 

The initial check for an AMI diagnosis is required to align with specifications set forth for the AMI measure set. Therefore, this secondary check for Acute or Evolving MIs is needed to further refine the denominator.

 

Further, while the Acute or Evolving MI value set and STEMI Exclusions value sets are completely contained within the broader AMI value set, only the STEMI Exclusion value set includes non-STEMI codes. Therefore, narrowing the denominator to look for a principal diagnosis within the Acute or Evolving MI value set will eliminate any non-ST-segment elevation MIs.

◦Encounter, Performed: Inpatient Encounter (diagnosis: Acute or Evolving MI)
•AND NOT: Intersection of:
◦Occurrence A of $EncounterInpatient
◦Encounter, Performed: Inpatient Encounter (diagnosis: STEMI Exclusion)

Or (option 2) modify and move the AND NOT portion of the above statement to the Denominator Exclusions:

•OR: Intersection of:
◦Occurrence A of $EncounterInpatient
◦Encounter, Performed: Inpatient Encounter (diagnosis: STEMI Exclusion)

These proposed constructs will allow the Acute or Evolving MI diagnosis to occur before or during the encounter and associate the diagnosis to that encounter.

 

If any of the above suggestions are approved, the corresponding metadata definitions will also need to be modified accordingly:

For option 1: Denominator: Initial population with an ECG and resulting diagnosis of Acute or Evolving MI overlapping the inpatient encounter and a primary PCI procedure closest to the inpatient admission that does not start after fibrinolytic therapy

For option 2: Denominator Exclusions: Patients who transferred from another hospital's inpatient, outpatient, or emergency department and ambulatory surgery center facilities or have an ECG and resulting diagnosis of non-ST-segment myocardial infarction (STEMI) overlapping the inpatient encounter are excluded

We may want to also consider adding guidance related to the ECG driven diagnosis:

A secondary (non-principal) diagnosis of Acute or Evolving myocardial infarction or non-ST-segment myocardial infarction is based on the result (interpretation) of the above described ECG.​


We propose changing the denominator logic to add a principal diagnosis of acute or evolving MI using the Acute or Evolving MI value set and either (1) removing the ‘AND NOT: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG) (result: STEMI Exclusions)’ statement to reduce redundancy, or (2) retaining the ‘AND NOT: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG) (result: STEMI Exclusions)’ to more closely resemble the current logic construct.

   

 Proposed logic update – Option 1 (makes logic more concise by reducing redundant logic statements):

   

 IP section:

•AND: Intersection of:
◦Occurrence A of $EncounterInpatient
◦"Encounter, Performed: Encounter Inpatient (principal diagnosis: Acute Myocardial Infarction (AMI))"
 (This casts a large net for the initial population based on a diagnosis code for AMI (containing both acute or evolving MI [ST segment elevation present] patients and non-ST segment elevation MI patients)

 Denominator section:

 · AND: Intersection of:

 o Occurrence A of $EncounterInpatient

 o Encounter, Performed: Inpatient Encounter (principal diagnosis: Acute or Evolving MI)

 (This pulls a subset of the IP population, only those with a ST segment elevation (acute or evolving MI) patients). To clarify, this does not indicate the diagnosis code must be entered twice, only that the diagnosis code is run thru the logic query twice, 1st to determine if it is included in the AMI value set and a 2nd time to determine if it is also in the Acute or Evolving MI value set (AMI subset).

 Remove the current AND NOT logic for the ‘non-ST segment MI’.

   

Proposed logic update – Option 2 (more closely resembles the current logic construct):

   

 IP section:

•AND: Intersection of:
◦Occurrence A of $EncounterInpatient
◦"Encounter, Performed: Encounter Inpatient (principal diagnosis: Acute Myocardial Infarction (AMI))"
 Denominator section:

 · AND: Intersection of:

 o Occurrence A of $EncounterInpatient

 o Encounter, Performed: Inpatient Encounter (principal diagnosis: Acute or Evolving MI)

 · AND NOT: Intersection of:

 o Occurrence A of $EncounterInpatient

   

 o Encounter, Performed: Inpatient Encounter (principal diagnosis: Non-STEMI)

 Please provide your vote by leaving a comment indicating "Yes" or "No."
CRP Recommendations: Implementers have expressed difficulty in capturing ECG results through EHR data. An ECG result is generally in a non-discrete field and therefore not easily retrieved electronically, whereas a diagnosis based on the ECG result would more likely be coded as such. So, using a data type of Diagnosis instead of Diagnostic Study might more closely match how the information is recorded. Therefore, we propose adding a principal diagnosis attribute of Acute or Evolving MI to the denominator logic and removing the Diagnostic Study, Performed: Electrocardiogram (ECG) (result: Acute or Evolving MI) data type.

 

 

Additionally, because the ‘Acute or Evolving MI’ value set does not include non-STEMI codes, the logic statement that checks for non-STEMIs through the STEMI Exclusions value set is redundant. Limiting the patients to only those with a diagnosis code from the ‘Acute or Evolving MI’ value set negates the additional logic statement to exclude those with a code from the ‘STEMI Exclusions’ value set. Therefore, we also recommend removing the Diagnostic Study, Performed: Electrocardiogram (ECG) (result: STEMI Exclusions) data type to reduce redundancy in the logic. This would also include removal of the STEMI Exclusions grouping value set (2.16.840.1.113762.1.4.1045.36).

 

This would impact five lines of Denominator logic and one QDM Data Element as follows:

•Denominator =
◦AND: Initial Population
◦AND: Intersection of:
◾Occurrence A of $EncounterInpatient
◾Encounter, Performed: Inpatient Encounter (principal diagnosis: Acute or Evolving MI)
◦AND:
◾OR:
◾AND: "Occurrence A of Encounter, Performed: Emergency Department Visit" <= 1 hour(s) ends before or concurrent with start of Occurrence A of $EncounterInpatient
◾AND:
◾OR: Most Recent: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG)" <= 1 hour(s) starts before start of "Occurrence A of Encounter, Performed: Emergency Department Visit"
◾OR: First: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG)" starts during "Occurrence A of Encounter, Performed: Emergency Department Visit"
◾OR: First: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG)" starts during Occurrence A of $EncounterInpatient
◾OR:
◾AND NOT: "Encounter, Performed: Emergency Department Visit" <= 1 hour(s) ends before or concurrent with start of Occurrence A of $EncounterInpatient
◾AND:
◾OR: Most Recent: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG)" <= 1 hour(s) starts before start of Occurrence A of $EncounterInpatient
◾OR: First: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG)" starts during Occurrence A of $EncounterInpatient
◦AND: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG) (result: Acute or Evolving MI)"
◦AND NOT: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG) (result: STEMI Exclusions)"
◦AND:
◾OR:
◾AND: First: "Occurrence A of Procedure, Performed: PCI" <= 1440 minute(s) starts after start of Occurrence A of $EncounterInpatient
◾AND NOT: "Encounter, Performed: Emergency Department Visit" <= 1 hour(s) ends before or concurrent with start of Occurrence A of $EncounterInpatient
◾AND NOT: "Occurrence A of Procedure, Performed: PCI" starts after start of ("Medication, Administered: Fibrinolytic Therapy" during Occurrence A of $EncounterInpatient )
◾OR:
◾AND: "Occurrence A of Encounter, Performed: Emergency Department Visit" <= 1 hour(s) ends before or concurrent with start of Occurrence A of $EncounterInpatient
◾AND: First: "Occurrence A of Procedure, Performed: PCI" <= 1440 minute(s) starts after start of "Occurrence A of Encounter, Performed: Emergency Department Visit"
◾AND NOT: "Occurrence A of Procedure, Performed: PCI" starts after start of ("Medication, Administered: Fibrinolytic Therapy" starts after start of "Occurrence A of Encounter, Performed: Emergency Department Visit" )
Data Criteria (QDM Data Elements)
•Attribute: "Result: STEMI Exclusions" using "STEMI Exclusions Grouping Value Set (2.16.840.1.113762.1.4.1045.36)"

This issue was sent to CRP for public comment and voting. We received three positive votes for this proposed logic change, and two supportive comments that looking at the principle diagnosis rather than the ECG result “was a move in the right direction”. No negative votes or comments were received.​
Last Commented Date:

 Comments   
Comment by Joelencia Leflore [ 05/26/17 ]

This issue has been addressed in the 2017 Annual Update.

Comment by Derrick Owusu (Inactive) [ 01/30/17 ]

Thank you for all of the feedback, the CRP voting period is now over. We will review these comments and discuss amongst EH measure developers.

Comment by Cathy Campbell (Inactive) [ 01/27/17 ]

We add our vote for option 1 as well.

Comment by Lynn Perrine [ 01/25/17 ]

Thank you for voting, your feedback is greatly appreciated. The value set information can be obtained from the Value Set Authority Center (VSAC) or the United States Health Information Knowledgebase (USHIK) available through the eCQI Resource Center. You will need a free UMLS license to view these sites.

The Acute Myocardial Infarction (AMI) value set OID is 2.16.840.1.113883.3.666.5.3011
The Acute or Evolving MI value set OID is 2.16.840.1.113883.3.666.5.3022

Links:
VSAC @ https://vsac.nlm.nih.gov/ > search codes by OID
eCQI Resource Center @ https://ecqi.healthit.gov/eh/ecqms-2017-reporting-period > go to the row for the AMI8a measure and click on the Version Detail link under the USHIK Version Links column (last column to the right) > login to USHIK > click the Data Criteria tab to view the value sets and associated codes.

Comment by Alex Liu [ 01/23/17 ]

Thank you for the clarification. We will also add a vote for Option 1. Our only concern would be to ensure that value set for "Attribute: Principal Diagnosis Acute Myocardial Infarction (AMI)" contains the comprehensive list of STEMI diagnoses that will be omitted from the value set for "Attribute: Principal Diagnosis Acute or Evolving MI". Are there any working lists for these two value sets that are available for public viewing?

Comment by Nancy Walker (Inactive) [ 01/23/17 ]

Vote for Option 1; We propose changing the denominator logic to add a principal diagnosis of acute or evolving MI using the Acute or Evolving MI value set and either (1) removing the ‘AND NOT: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG) (result: STEMI Exclusions)’ statement to reduce redundancy

Comment by Lynn Perrine [ 01/17/17 ]

While the logic statements regarding the encounter with the principal diagnosis attribute are different between the two options, the intent is the same. As you pointed out, the ‘Acute or Evolving MI’ value set does not include non-STEMI codes, which allowed us to compose Option 1. By limiting the patients to only those with a diagnosis code from the ‘Acute or Evolving MI’ value set would negate the additional logic statement to exclude those with a code from the ‘STEMI Exclusions’ value set.

However, we included Option 2 as a possible consideration because it is closest to the current ECG logic constructs that does include logic that would exclude those patients with a non-STEMI code.

As part of our update, we determined it would be more transparent to change the name of the ‘STEMI Exclusions’ used for the ECG result value set to ‘Non-STEMI’ for the principal diagnosis value set. If we proceed with Option 2, this is how the value set name will be updated.

We hope this additional information is helpful as you consider these options. Thank you.

Comment by Cathy Campbell (Inactive) [ 01/17/17 ]

We appreciate the additional information, but agree that neither option can be voted upon as currently stated. We support looking to the principal diagnosis and the removal of the ECG result, however the options as stated are not quite clear enough to make a decisive choice. Would definitely recommend additional consideration and follow up for this issue, as it should not remain as is.

Comment by Howard Bregman (Inactive) [ 01/13/17 ]

Thank you for this clarification.

We think you are moving in the right direction by looking at principle diagnosis instead of the ECG result. However, we still don't understand the two options - they don't seem to be equivalent regarding what is being measured. Your statement above suggests that the acute and evolving MI value set does not contain non-STEMI values, but if that is the case, then the two statements of option 2 don't seem to be necessary. Also we are not clear why option 1 no longer looks at a non-STEMI exclusion.

At this point we can't vote for either option but are still open to futher consideration.

Comment by Lynn Perrine [ 01/13/17 ]

Thank you for your comment and request for clarification. Yes, we are proposing removal of the denominator logic statement "AND: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG) (result: Acute or Evolving MI)" as well as “AND NOT: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG) (result: STEMI Exclusions)" to reduce redundancy in the logic. We have provided two different denominator logic options for which implementer feedback and preference would be appreciated. Option i makes the logic more concise by reducing redundant logic statements while Option 2 more closely resembles the current logic construct. Please see attached file for full details. Thanks!

Comment by Howard Bregman (Inactive) [ 01/09/17 ]

Unfortunately the CRP proposal is confusing. We assume that this part will be removed as well, right? "AND: "Occurrence A of Diagnostic Study, Performed: Electrocardiogram (ECG) (result: Acute or Evolving MI)"

Please clarify.

Comment by Derrick Owusu (Inactive) [ 10/18/16 ]

Thank you for all of the feedback, the CRP public comment period is now over. We will review these comments and discuss amongst EH measure developers.

Comment by Howard Bregman (Inactive) [ 10/17/16 ]

As we weren't asked about consolidating the value sets, I do think that changing the measure with either options 1 or 2 is a significant improvement to the specifications.

I do agree with Floyd that you should abandon the pretense that the diagnosis has to be in anyway associated with the ECG. It's not going to be recorded in that way (you are already agreeing with that) so there is no benefit to requiring that it be associated by guidance or any other means.

Comment by Floyd Eisenberg (Inactive) [ 10/17/16 ]

Comment for ESAC: I agree with the above comments that the two value sets are sufficiently close (Acute Myocardial Infarction) and "Acute or Evolving MI" have significant overlap so there is no benefit and potential additional work by requiring double documentation. I also agree with the comment about temporal requirements. From the last statement in the CRP request it seems the anticipated guidance expects the condition (acute or evolving or non-stemi) to be identified in EKG. Until such time as routine structured EKG interpretations are available the ability to relate the diagnoses to the EKG results can be managed only via abstraction. The guidance cannot be supported.

Comment by Susan Wisnieski (Inactive) [ 10/14/16 ]

I agree with the above comments as well, but since the CRP Overview does not outline temporal timing, I would like to put in a friendly request for critical consideration to the temporal requirements for the diagnoses evaluated against the encounters. We of course want it to be clinically accurate, but temporal requirements such as overlap for diagnoses can be complicated to accurately capture.

Comment by Cathy Campbell (Inactive) [ 10/13/16 ]

I agree with the comments above. If the codes within the Acute Myocardial Infarction (AMI) value set are all contained within the Acute or Evolving MI value set, there is no need for the information to be documented twice.

Comment by Howard Bregman (Inactive) [ 10/11/16 ]

Agree with above (we are co-workers, so you can count that as one vote). Neither option seems practical although they are an improvement from current state. They just require that the same diagnosis be documented twice for no benefit.

Comment by Alex Liu [ 10/11/16 ]

The Reporting Year 2017 value sets for "Acute or Evolving MI" and "Acute Myocardial Infarction (AMI)" are almost exactly the same. In fact, all codes within "Acute or Evolving MI" are entirely contained within "Acute Myocardial Infarction (AMI)".

The IPP requirement for CMS53 is that the patient has a principal diagnosis of "Acute Myocardial Infarction (AMI)". Adding a secondary diagnosis check for "Acute or Evolving MI" seems out of place, since the majority of cases will result in dual documentation of the same diagnosis code. It seems to me that the entire following check could be removed from the denominator statement:
•AND: Intersection of:
◦Occurrence A of $EncounterInpatient
◦Encounter, Performed: Inpatient Encounter (diagnosis: Acute or Evolving MI)

For STEMI exclusion diagnosis, we would request also allowing for the Diagnosis template to be used as follows:
•OR: Intersection of:
◦Occurrence A of $EncounterInpatient
◦ Union of:

  • Encounter, Performed: Inpatient Encounter (diagnosis: STEMI Exclusion)
  • Diagnosis: STEMI Exclusion overlaps (or some other QDM temporal timing) Occurrence A of $EncounterInpatient
Generated at Mon Sep 22 00:16:01 UTC 2025 using Jira 10.3.9#10030009-sha1:eff8913ed2270ee44ab422c3609af4c4f36536d0.