It is obvious that a significant amount of work and thought went into developing these electronic versions of the clinical quality measurement. With that in mind, we appreciate the opportunity to give feedback to the committee that developed the specifications. We are posting a single response but will refer to all 3 measures in this response. We will start by providing general thoughts on the measures and attempt to address some of your specific questions.

In general, we consider these measures to be more suited for the ambulatory setting. We believe that a primary care physician maintains a stronger, more personal influence on a patient in regards to smoking and treatment. Typically, a person does not enter the inpatient setting due to a smoking related illness without a referral from their primary care physician. In this case, we would expect that the PCP had already discussed treatment options and/or prescribed the appropriate medications. While we know, that a certain percentage of people use the emergency department as their primary care provider, we feel that requiring 100% screening and intervention is not feasible and wonder about the effectiveness of decreasing the rate of active smokers/tobacco users. In the study by Rigotti et al (2008) that is cited in your supporting references, their conclusion was that only when supportive contact is continued for 1 month after discharge, are the intra-hospitalization interventions effective. This supports evidence that the anti-tobacco interventions are more effectively managed in the ambulatory setting. While we believe that the practice guidelines are effective in decreasing the rate of tobacco users, we don’t believe the guidelines will be as effective in the inpatient setting. If these measures are to be implemented, we would like to provide the following feedback specific to the measures:

All Measures;

We suggest that when the empty value sets are created, that there is a mechanism in place for the public to provide feedback on the values. We have learned in our experience of implementing eCQMs that the value sets can be constricting thus making it difficult to capture and report specific data elements. Documentation practices can vary across institutions and forcing a clinician to change how they normally document is a significant burden on clinicians and implementers. Feedback to assure the value sets are broad enough to fit varying practices would be extremely helpful.

TOB-1 Use Screening

• Initial population is for all patients >=18 y/o. This does not align with the meaningful use core object of recording smoking status which is all patients >=13 y/o. It is easier for clinical staff to think of a single population, either using tobacco products or not. We recommend aligning one with the other for consistency.
• To make the Glasgow Coma Score (GCS) more efficient, we recommend removing all but "Physical Exam, Performed: Glasgow Coma Score Total" satisfies all: (result <= 8). While we understand the purpose of the GCS granularity to measure severity of the patient's condition, for this measure we question the purpose of including the individual values given that in our EHR and normal clinical practice, the GCS is always totaled. While leaving the more granular scores present may appear not to cause any additional burden, there is a level of mapping and maintenance involved for every data element. In addition, the more data elements the query has to look for, the more inefficient the algorithm becomes. We have learned this with the complex Venous Thromboembolism measures when the reports would take multiple hours to complete for a large healthcare system like ours.

TOB-2 & 3

• "Medication, Order not done: Patient Refusal" for "Tobacco Use Cessation Pharmacotherapy Ingredient Specific" construct is not compliant with the recent QRDA I & III Implementation Guide. I refer to Section 5.2.3.1 (Pg. 30) “Not Done with a Reason” copied below. I’ve bolded the applicable coding that shows when referring to a medication not given, the Value Set OID is to be referenced. The way the TOB algorithm is currently written, it references the “ingredient specific” medication. This forces a physician to select a specific medication that was refused, instead of the current proper way of stating “all of the medications were refused” using the value set OID. We request that ingredient specific reference related to contraindications be removed from all negation constructs.

• For a QDM data element that is not done (when negationInd="true") with a reason, such as "Medication, Order not done: Medical Reason", an entryRelationship to a Reason (templateId: 2.16.840.1.113883.10.20.24.3.88") with an actRelationship type of "RSON" is required. This is specified in the section 3.4 Asserting an Act Did Not Occur with a Reason in the base HL7 QRDA-I, R3 Implementation Guide. To summarize, the following steps shall be followed:
  • Set the containing act attribute negataionInd=”true”
  • Use code/[@nullFlavor="NA"]
  • Set code attribute code/sdtc:valueset="[VSAC value set OID]"
  • Use code/originalText for the text description of the concept in the pattern "None of value set: [value set name]"
  h4 Figure 16: Not Done Example
  <!--Medication administered not done, patient refusal: Drug declined by patient - reason unknown. No "Antibiotic Medications for Pharyngitis" were administered -->
  <act classCode="ACT" moodCode="EVN" negationInd="true">
  <templateId root="2.16.840.1.113883.10.20.24.3.42" extension="2014-
  12-01" />
  <id root="517d5bbb-03a8-4400-8a78-754321641159" />
  <code code="416118004" displayName="Administration"
  codeSystem="2.16.840.1.113883.6.96" codeSystemName="SNOMED-CT"
  <statusCode code="completed" />
  <entryRelationship typeCode="COMP">
  <substanceAdministration classCode="SBADM" moodCode="EVN">
  <manufacturedProduct classCode="MANU">
  <templateId root="2.16.840.1.113883.10.20.22.4.23"
  extension="2014-06-09" />
  <id root="37bfe02a-3e97-4bd6-9197-bbd0ed0de79e" />
  <manufacturedMaterial>
  <code nullFlavor="NA"
  sdtc:valueSet="2.16.840.1.113883.3.464.1003.196.12.1001">
  <originalText> None of value set: Antibiotic Medications
  for Pharyngitis</originalText>

• In addition to the above, we strongly recommend removing all "Medication, Order not done: Patient Refusal" constructs from the algorhythm and only keeping "Medication, Administered not done: Patient Refusal". Please review CQM-225 (https://jira.oncprojectracking.org/browse/CQM-225) and related discussions for rationale.

• For "Risk Category Assessment:” we recommend one category that combines the two into the following:
• Light User: practical counseling during stay/referral to outpatient counseling at discharge
  • Smokeless Tobacco: Any
    Cigar/Pipe: Some days
    Cigarette: < 5 per day/< 0.25 packs per day
  • Heavy User: practical counseling AND cessation medications during stay /referral to outpatient counseling AND received prescription for cessation medications at discharge
    Cigar/Pipe: Every day
    Cigarette: >= 5 per day/>= 0.25 packs per day
    The above 5 categories would be combined into a value set called “Risk Category Assessment: Tobacco Type and Frequency” (Light and Heavy with the descriptions are major categories not to be included in value set).
    Rationale
  • Significantly reduces cognitive burden for providers, as well as number of clicks to document the discrete data fields.
  • Specific frequency unlikely to be accurate.

• In conclusion, we recommend approving these measures for the ambulatory setting after reconfiguration. If the measures are approved for the inpatient setting, please consider the recommended changes to the algorithm to make the data capture and queries more efficient and end-user friendly. We thank you for considering our recommended changes.