NQF-Endorsed Measures for Endocrine Conditions: Cycle 3, 2015

Transcription

1 NQF-Endorsed Measures for Endocrine Conditions: Cycle 3, 2015 TECHNICAL REPORT November 2015 This report is funded by the Department of Health and Human Services under contract HHSM I Task Order HHSM-500-T0008 1

2 Contents Executive Summary...3 Endocrine Measure Evaluation: Cycle 3 Review: December 2014 January Table 1. Endocrine Cycle 3 Measure Review Summary... 4 Comments Received... 4 Overarching Issue... 5 Summary of Measure Evaluation... 5 Appendix A: Details of Measure Evaluation...7 Measures Endorsed Comprehensive Diabetes Care: Blood Pressure Control (<140/90 mm Hg) Optimal Diabetes Care Measures Withdrawn from Consideration Appendix B: Endocrine Portfolio Use in Federal Programs Appendix C: Project Standing Committee and NQF Staff Appendix D: Measure Specifications Comprehensive Diabetes Care Blood Pressure Control (<140/90 MM HG) Optimal Diabetes Care Appendix E1: Related and Competing Measures (tabular format) Appendix E2: Related and Competing Measures (narrative format)

3 NQF-Endorsed Measures for Endocrine Conditions: Cycle 3, 2015 TECHNICAL REPORT Executive Summary This is the third in a series of 3 reports describing NQF's measure evaluation project for Endocrine conditions. NQF selected this project to pilot more frequent submission and evaluation of measures than what is possible in our current 3-year measure maintenance process. This 22-month project included 3 full endorsement cycles, allowing for the submission and evaluation of both new and previously endorsed measures every 6 months. The background and description of the project, review of NQF's Endocrine portfolio, and the results of the cycle 1 and cycle 2 evaluations are available on NQF's project webpage. In cycle 3 of this project, the Standing Committee evaluated the following 2 measures undergoing maintenance review against NQF s standard measure evaluation criteria: 0061: Comprehensive Diabetes Care: Blood Pressure Control 0729: Optimal Diabetes Care The Committee recommended both measures as suitable for endorsement, and both were subsequently endorsed. Brief summaries of the measures reviewed during cycle 3 of the project are included in the body of this report; detailed summaries of the Committee s discussion and ratings of the criteria are included in Appendix A. 3

4 Endocrine Measure Evaluation: Cycle 3 Review: December 2014 January 2015 In cycle 3 of the Endocrine Measure Evaluation pilot, the Endocrine Standing Committee evaluated 2 diabetes measures undergoing maintenance review against NQF s standard evaluation criteria. One measure was an intermediate clinical outcome measure, and the other was an all-or-none composite measure. The Committee discussed these measures during two conference calls held on January 22 and 28, The Committee s discussion and ratings of the criteria are summarized in Appendix A. Table 1. Endocrine Cycle 3 Measure Review Summary Maintenance New Total Measures under consideration Measures recommended for endorsement Measures not recommended for endorsement Reasons for not recommending N/A N/A N/A Comments Received Comments Received Prior to Committee Evaluation The pre-evaluation comment period was open from December 20, 2014 to January 12, No preevaluation comments were received for the measures under review in this cycle of the project. Comments Received After Committee Evaluation The 30-day post-evaluation comment was open from March 5 to April 3, During this commenting period, NQF received comments from 5 member organizations. The Committee discussed these comments during the Committee's post-comment call that was held on April 16, One comment expressed support of measure #0061: Comprehensive Diabetes Care: Blood Pressure Control (<140/90 mm Hg). The remaining comments pertained to measure # 0729: Optimal Diabetes Care, and each was critical of the measure. Two comments specifically addressed the glucose control component of the composite measure, stating that the blood glucose threshold of <8% for patients <75 years of age does not conform to current clinical practice guidelines for glycemic control. Two comments were critical of the composite measure itself, citing concern that use of the composite measure could mask the individual care processes that most need improvement. Two comments noted the inability to document HbA1c levels between 8-9 % using CPT-II codes. Two comments suggested a need for including sociodemographic factors in the risk-adjustment approach for the measure. Finally, one comment suggested the need for additional detail regarding moderate or high intensity statin dosing in the description of statin use for the measure. 4

5 Overarching Issue The only overarching issue for the measures evaluated in this cycle of the project was that of competing measures. Specifically, the one individual measure that was evaluated in this cycle (blood pressure control) is also included as a component in the all-or-none composite measure. NQF staff asked the Committee to discuss whether there is justification for continued endorsement of the individual measure if the composite retains endorsement. The Committee ultimately agreed that while the composite measure is useful to assess patient-centric performance across a variety of clinical areas, endorsement of the individual measure also can be beneficial, particularly for users who want to focus on blood pressure control components specifically or for those who have data collection constraints and cannot use the composite. The Committee therefore recommended continued endorsement of both the individual measure and the composite measure. Summary of Measure Evaluation The following brief summaries of the measures and the evaluation highlight the major issues that were considered by the Committee. Details of the Committee s discussion and ratings of the criteria are included in Appendix A. Diabetes Care 0061: Comprehensive Diabetes Care: Blood Pressure Control (NCQA): Endorsed Description: The percentage of patients years of age with diabetes (type 1 and type 2) whose most recent blood pressure level taken during the measurement year is <140/90 mm Hg. Measure Type: Outcome; Level of Analysis: Clinician: Group/Practice, Clinician: Individual, Health Plan, Integrated Delivery System; Setting of Care: Ambulatory Care: Clinician Office/Clinic; Data Source: Administrative claims, Electronic Clinical Data, Electronic Clinical Data : Electronic Health Record, Electronic Clinical Data: Laboratory, Electronic Clinical Data: Pharmacy, Paper Medical Records This measure has been NQF-endorsed since 2002 and is used for public reporting, payment, and quality improvement programs at the health plan level, and in public reporting at the clinician level. The measure aims to improve the quality of care for patients with diabetes by assessing whether their blood pressure was adequately controlled (<140/90 mm Hg). High blood pressure is a major risk factor for microvascular and macrovascular complications in patients with diabetes. Evidence provided by the developer included a reference to 5 clinical practice guideline recommendations (published in The Journal of the American Medical Association (JAMA) in 2014 by panel members appointed to the Eighth Joint National Committee (JNC8).Three randomized controlled trials (RCTs) were presented as evidence to support a blood pressure goal of <150/90 mm Hg for patients with diabetes. The Committee questioned the measure specification of a blood pressure measurement <140/90 mmhg (rather than 150/90 mmhg), but accepted the developer s justification that the specifications were based on the JNC8 panel members guideline recommendations. The Committee agreed that the empirical reliability and validity testing results were acceptable. Given the sufficient evidence, reliability, and validity of the measure, the Committee recommended the measure as suitable for continued endorsement. 5

6 0729: Optimal Diabetes Care (Minnesota Community Measurement): Endorsed Description: The percentage of adult diabetes patients who have optimally managed modifiable risk factors (A1c, blood pressure, statin use, tobacco non-use and daily aspirin or anti-platelet use for patients with diagnosis of ischemic vascular disease) with the intent of preventing or reducing future complications associated with poorly managed diabetes. Patients ages with a diagnosis of diabetes, who meet all the numerator targets of this composite measure: A1c less than 8.0, Blood Pressure less than 140 systolic and less than 90 diastolic, Statin use unless contraindications or exceptions, Tobacco-free (non-user) and for patients with diagnosis of ischemic vascular disease daily aspirin or antiplatelet use unless contraindicated. Measure Type: Composite; Level of Analysis: Clinician: Group/Practice; Setting of Care: Ambulatory Care: Clinician Office/Clinic; Data Source: Electronic Clinical Data: Electronic Health Record, Paper Medical Records This risk-adjusted all-or-none composite measure has been NQF-endorsed since The measure has 5 components: HbA1c <8.0%; blood pressure <140/90; statin use unless contraindications or exceptions, tobacco-free (non-user); and for patients with diagnosis of ischemic vascular disease, daily aspirin or antiplatelet use unless contraindicated. The specifications of the statin component of this measure have changed since the most recent endorsement of the measure due to changes in the clinical practice guidelines for cholesterol management that were released in November 2013 by the American College of Cardiology and the American Heart Association. In the earlier version of the measure, the statin component assessed reaching a target LDL <100; the revised version of this component assesses appropriate statin use. Evidence supporting this measure included evidence-based recommendations from the Institute for Clinical Systems Improvement, American College of Cardiology, and the U.S. Preventive Services Task Force. The Committee agreed that the evidence provided for each component supported a link to improved outcomes for diabetes patients. The Committee agreed that the results of the reliability and validity testing were acceptable. The measure is publicly reported and is used in payfor-performance and accreditation programs. Given the sufficient evidence, reliability, and validity of the measure, the Committee recommended the measure as suitable for continued endorsement. NQF received comments from 5 member organizations. These comments questioned the blood glucose threshold of less than 8%, criticized the composite approach, suggested a need for inclusion of sociodemographic factors in the risk-adjustment approach, and advised including additional detail regarding moderate or high intensity dosing of statin therapy in the measure. The Committee discussed each of the comments but did not change their recommendation for continued endorsement. During the appeals period, one organization submitted an appeal that was specific to the HbA1c component of the measure. The decision to re-endorse the measure was upheld by the Consensus Standards Approval Committee (CSAC) and the Executive Committee of the NQF Board of Directors. 6

7 Appendix A: Details of Measure Evaluation Measures Endorsed Rating Scale: H=High; M=Moderate; L=Low; I=Insufficient; NA=Not Applicable; Y=Yes; N=No 0061 Comprehensive Diabetes Care: Blood Pressure Control (<140/90 mm Hg) Submission Specifications Description: The percentage of patients years of age with diabetes (type 1 and type 2) whose most recent blood pressure level taken during the measurement year is <140/90 mm Hg. Numerator Statement: Patients whose most recent blood pressure level was <140/90 mm Hg during the measurement year. The outcome being measured is a blood pressure reading of <140/90 mm Hg, which indicates adequately controlled blood pressure. Adequately controlled blood pressure in patients with diabetes reduces cardiovascular risks and microvascular diabetic complications. Denominator Statement: Patients years of age by the end of the measurement year who had a diagnosis of diabetes (type 1 and type 2) during the measurement year or the year prior to the measurement year. See question S.9 Denominator Details for methods to identify patients with diabetes. Exclusions: Exclusions -Exclude patients who did NOT have a diagnosis of diabetes, in any setting, during the measurement year or the year prior to the measurement year. AND either: -A diagnosis of polycystic ovaries, in any setting, any time in the patient s history through December 31 of the measurement year, or -A diagnosis of gestational or steroid-induced diabetes, in any setting, during the measurement year or the year prior to the measurement year. Adjustment/Stratification: No risk adjustment or risk stratification Level of Analysis: Clinician : Group/Practice, Clinician : Individual, Health Plan, Integrated Delivery System Setting of Care: Ambulatory Care : Clinician Office/Clinic Type of Measure: Outcome Data Source: Administrative claims, Electronic Clinical Data, Electronic Clinical Data : Electronic Health Record, Electronic Clinical Data : Laboratory, Electronic Clinical Data : Pharmacy, Paper Medical Records Measure Steward: National Committee for Quality Assurance STANDING COMMITTEE MEETING [01/22/2015] 1. Importance to Measure and Report: The measure meets the Importance criteria (1a. Evidence, 1b. Performance Gap) 1a. Evidence: H-2; M-11; L-1; I-3; 1b. Performance Gap: H-15; M-2; L-0; I-0 7

8 Rationale: Evidence provided by the developer included a reference to five clinical practice guideline recommendations published in a February 2014 article in the Journal of the American Medical Association by the panel members appointed to the Eighth Joint National Committee (JNC8). Three randomized controlled trials (RCTs) were presented as evidence to support a blood pressure goal of <150/90 mm Hg for patients with diabetes. The goal of <140/90 mmhg is based on expert opinion of the JNC8 panel members and is consistent with the blood pressure goals articulated by the panel for the general population younger than 60 years. Both Committee members and the developers acknowledged there are no studies that directly support a blood pressure goal of <140/90 mm Hg for patients with diabetes. However, the Committee agreed that there is strong evidence that moderate lowering of blood pressure in patients with diabetes is associated with a reduction of cardiovascular risks and microvascular diabetic complications. Some Committee members suggested that there is no clinical or evidentiary support to distinguish between a threshold of <140/90 mm Hg vs. 140/90 mm Hg, noting that, operationally, clinicians tend to round up to 140/90 when not using digital cuffs if the reading is just slightly below that value. The developer explained that they based the measure threshold on the JNC8 panel members guideline recommendation of < 140/90 mm Hg for diabetic patients. HEDIS data presented by the developer for commercial, Medicaid, and Medicare health plans indicate average performance rates from 59% to 65% between Data presented by the developer for clinicians and practices participating in the NCQA Diabetes Recognition Program (DRP) indicate performance rates ranging from 80% to 81% for The Committee agreed that opportunity for improvement in performance of this measure remains. 2. Scientific Acceptability of Measure Properties: The measure meets the Scientific Acceptability criteria (2a. Reliability - precise specifications, testing; 2b. Validity - testing, threats to validity) 2a. Reliability: H-3; M-14; L-0; I-0 2b. Validity: H-6; M-8; L-0; I-3 Rationale: Developers provided results of reliability testing of the performance measure score from signalto-noise analysis of HEDIS and DRP data for the health plans and clinician levels of analysis, respectively. For health plans, the median reliability was 0.98 for commercial health plans, 0.97 for Medicaid health plans, and 0.95 for Medicare health plans; for clinicians, the median reliability was 0.6. The Committee agreed that the testing results demonstrate sufficient reliability. To demonstrate validity of the performance measure score, developers correlated the scores for this measure to scores from several other diabetes measures. For health plans, the correlations were moderate to high, statistically significant, and in the expected directions; for clinicians, the correlations were low but for most part in the expected directions. The developer also noted that face validity of the measure was assessed by three internal groups. The Committee agreed that the exclusions to the measure were appropriate. However, members raised concerns about the reliability and validity of the measure due to the potential for rounding (discussed more fully under evidence, above). The Committee also noted that often the intake blood pressure reading is what is entered into EHRs, but if a lower blood pressure is 8

9 observed later in the visit, this value may not be recorded in the EHR (or perhaps is entered only in the physician s notes). The developer explained that the most recent reading in the measurement year is used, and if there are multiple readings in one day or visit, the lowest systolic and the lowest diastolic readings can be reported by the provider. However, the Committee noted that this level of selection would be difficult to implement in EHRs. 3. Feasibility: H-14; M-3; L-0; I-0 (3a. Clinical data generated during care delivery; 3b. Electronic sources; 3c.Susceptibility to inaccuracies/ unintended consequences identified 3d. Data collection strategy can be implemented) Rationale: The Committee agreed that given the measure is in use, it is feasible to collect the data. Members also agreed that the data elements are generated as part of the care delivery process. 4. Usability and Use: H-12; M-5; L-0; I-0 (Used and useful to the intended audiences for 4a. Accountability and Transparency; 4b. Improvement; and 4c. Benefits outweigh evidence of unintended consequences) Rationale: The Committee agreed that the measure is in use in several accountability programs, including public reporting of health plan data. One member mentioned the lack in improvement in health plan performance over time; however, another member noted that physicians in the DNR have substantially higher rates than the health plans, suggesting that further improvement should be possible. The Committee voiced no concerns about potential unintended consequences of the measure. 5. Related and Competing Measures This measure is a competing measure to measure 0729: Optimal Diabetes Care, a composite measure that includes blood pressure control as one of its components. NQF staff asked the Committee to discuss whether there is justification for continued endorsement of this individual measure if the composite retains endorsement. The Committee discussed the pros and cons of endorsing both the individual measure and the composite measure. The Committee ultimately agreed that, while the composite measure is useful to assess patient-centric performance across a variety of clinical areas, endorsement of this individual measure also can be beneficial, particularly for users who want to focus on the blood pressure control components specifically or for those who have data collection constraints and cannot use the composite. The Committee therefore recommended continued endorsement of both the individual measure and the composite measure. Standing Committee Recommendation for Endorsement: Y-14; N-3 6. Public and Member Comment Comments received: One commenter supported the Committee s recommendation for endorsement of this measure. 9

10 7. Consensus Standards Approval Committee (CSAC) Review (May 12, 2015): Y-15; N-0; A-0 CSAC Decision: Approved for continued endorsement 8. Board of Directors Review: Yes (June 15, 2015) Board Decision: Ratified for continued endorsement 0729 Optimal Diabetes Care Submission Specifications Description: The percentage of adult diabetes patients who have optimally managed modifiable risk factors (A1c, blood pressure, statin use, tobacco non-use and daily aspirin or anti-platelet use for patients with diagnosis of ischemic vascular disease) with the intent of preventing or reducing future complications associated with poorly managed diabetes. Patients ages with a diagnosis of diabetes, who meet all the numerator targets of this composite measure: A1c less than 8.0, Blood Pressure less than 140 systolic and less than 90 diastolic, Statin use unless contraindications or exceptions, Tobacco-free (non-user) and for patients with diagnosis of ischemic vascular disease daily aspirin or antiplatelet use unless contraindicated. Please note that while the all-or-none composite measure is considered to be the gold standard, reflecting best patient outcomes, the individual components may be measured as well. This is particularly helpful in quality improvement efforts to better understand where opportunities exist in moving the patients toward achieving all of the desired outcomes. Please refer to the additional numerator logic provided for each component. Numerator Statement: Patients ages 18 to 75 with diabetes who meet all of the following targets from the most recent visit during the measurement year: A1c less than 8.0, Blood Pressure less than 140/90, Statin Use if no contraindications/ exceptions, Tobacco non-user and Daily aspirin or anti-platelets for patients with diagnosis of ischemic vascular disease use unless contraindicated. Denominator Statement: Patients ages 18 to 75 with diabetes who have at least two visits for this diagnosis in the last two years (established patient) with at least one visit in the last 12 months. Exclusions: Valid exclusions include patients who only had one visit to the clinic with diabetes codes during the last two years, patients who were pregnant, died or were in hospice or palliative care, or a permanent resident of a nursing home during the measurement year. Adjustment/Stratification: Level of Analysis: Clinician : Group/Practice Setting of Care: Ambulatory Care : Clinician Office/Clinic Type of Measure: Composite Data Source: Electronic Clinical Data, Electronic Clinical Data : Electronic Health Record, Paper Medical Records Measure Steward: MN Community Measurement STANDING COMMITTEE MEETING [01/22/2015] / [01/28/2015] 10

11 1. Importance to Measure and Report: The measure meets the Importance criteria (1a. Evidence, 1b. Performance Gap) 1a. Evidence: H-5; M-11; L-0; I-1; 1b. Performance Gap: H-15; 1d. Composite Quality Construct and Rationale: H-4; M-7; L-4; I-1 Rationale: For all but one of the components included in this composite (tobacco-free), the developer presented recommendations from the 2014 clinical practice guidelines developed by the Institute for Clinical Systems Improvement (ICSI), which were based on a systematic review of evidence that was graded either high or moderate. Additional evidence-based recommendations from the American College of Cardiology and U.S. Preventive Services Task Force also were presented. Committee members agreed that the evidence supports the relationship between each component and desired health outcomes. Data provided by the developer indicate that for 2014, only 38.9% of diabetic patients in Minnesota met all five component targets from the composite measure. Committee members agreed that although performance on some of the components is quite high, overall performance indicates opportunity for improvement. Although some Committee members voiced concern over the all-or-none structure of the measure, others agreed that a more comprehensive measure that focuses on management of multiple risk factors is needed. The Committee agreed that the developer description of the quality construct, rationale, and aggregation and weighting approach is explicitly articulated and logical. 2. Scientific Acceptability of Measure Properties: The measure meets the Scientific Acceptability criteria (2a. Reliability - precise specifications, testing; 2b. Validity - testing, threats to validity) 2a. Reliability: H-9; M-7; L-0; I-0; 2b. Validity: H-1; M-10; L-4; I-1; 2d. Composite: H-1; M-10; L-4; I-1 Rationale: Committee members noted that the specifications of the statin component of this measure have changed since the most recent endorsement of the measure due to changes in the ACC/AHA clinical practice guidelines on cholesterol management released in November, In the earlier version of the measure, the statin component assessed reaching a target LDL threshold of < 100 mg/dl; the revised version of this component assesses statin use. Committee members questioned whether the measure assesses if a patient is on the appropriate statin dose. Developers clarified that the measure does not consider the statin dose but assesses only whether a patient is on a statin. Members also questioned the age range of for the statin component of the measure. The developer clarified that for patients years of age, this component is applicable only if the patient has ischemic vascular disease or a very high LDL level, in accordance with the ACC/AHA guidelines. The developer clarified that the level of analysis for the measure is clinician groups (not individual clinicians), and also noted that multiple clinics may form a clinician group. They also clarified that the measure does not require having a minimum of 30 patients. Developers presented results of signal-to-noise reliability testing of the performance measure score. They clarified that the beta-binomial method was used for the reliability testing because the composite score itself is a binary (yes/no) measure. Members noted that although the 11

12 reliability was quite high for most clinician groups, it was lower than 0.7 for some clinician groups. To demonstrate validity of the performance measure score, developers examined the association between the scores for this measure with the scores from the Optimal Vascular Care measure (NQF #0076), hypothesizing that clinician groups likely provide similar quality of care to different patients who also require management of multiple risk factors. The R 2 value from this analysis was The developers also described several steps occurring during the data submission process as demonstration of empirical validity testing at the data level element. Developers also clarified that the measure is risk-adjusted for three factors (insurance type, age group, and diabetes type) and noted that the risk-adjustment strategy was developed using data from all clinicians in Minnesota. However, one member expressed some concern that the only adjustment for sociodemograhic status is insurance type. Developers clarified that other potential risk factors that were considered were not statistically significant and thus were not included in the risk-adjustment model. Several Committee members voiced concern about holding physicians accountable for the patient s tobacco use, as some see actual tobacco use (as opposed to efforts for tobacco cessation) as out of the control of the clinician. However, another member referred to data showing that physicians can influence their patients to stop tobacco use. Developers also noted that statewide, they have seen an approximate 2.5% increase in tobacco-free patients in Minnesota. One Committee member noted the need for clarity about potential adverse effects related to statin use. Another member referenced the flow diagram provided by the developer that details several contraindications for statin use, while another member echoed the importance of the potential for adverse reactions when making treatment decisions. After developers clarified the performance rates for each of the components, Committee members questioned whether the aspirin component (performance rate =99.5% in MN) is needed in the composite. Developers noted that while this component may be "topped out" in MN, this happened over a four-year period of focus on this component. They also referenced a New England Journal of Medicine article that found a 34.8% performance rate nationally in the primary care setting. Finally, they noted that performance on this component across ACOs is, on average, 75.3%. 3. Feasibility: H-7; M-4; L-4; I-0 (3a. Clinical data generated during care delivery; 3b. Electronic sources; 3c.Susceptibility to inaccuracies/ unintended consequences identified 3d. Data collection strategy can be implemented) Rationale: The measure data can be collected through electronic clinical data and paper records. One Committee member noted that, due to the number of components included in the composite, the data collection effort for this composite measure may be intensive. Developers stated that submission of this measure by all clinician groups in MN is mandated by the state. While they acknowledged that MN has many large practices that use EHRs, small practices even those who still use paper medical records are able to submit data on this measure. The developers did, however, acknowledge the data collection burden for the new statin component if a patient has not been prescribed a statin (i.e., identifying exceptions due to contraindications). 12

13 4. Usability and Use: H-5; M-7; L-4; I-0 ((Used and useful to the intended audiences for 4a. Accountability and Transparency; 4b. Improvement; and 4c. Benefits outweigh evidence of unintended consequences) Rationale: Committee members noted that the measure is publicly reported and is used in pay-forperformance and accreditation programs. Performance is slowly increasing across the state of Minnesota, suggesting quality of care may be improving. Data submitted by the developer demonstrate relatively consistent improvement of performance in MN from the years Committee members agreed that this composite measure is patient-centric and acknowledged the importance of using a comprehensive measure that assesses performance of reducing multiple risk factors. Some Committee members expressed concern that the measure could incent some providers to "cherry-pick" patients or make their practices less hospitable to certain patients or certain subgroups of patients (the tobacco-free component of the measure was a particular concern). 5. Related and Competing Measures This measure is a competing measure to the following measures o 0061: Comprehensive Diabetes Care: Blood Pressure Control (<140/90 mm Hg). o 0575: Comprehensive Diabetes Care: Hemoglobin A1c (HbA1c) control (<8%) NQF staff asked the Committee to discuss whether there is justification for continued endorsement of the individual measures if the composite retains endorsement. The Committee discussed the pros and cons of endorsing both individual measures and the composite measure. The Committee ultimately agreed that while the composite measure is useful to assess patientcentric performance across a variety of clinical areas, endorsement of individual measures also can be beneficial, particularly for users who want to focus on certain components of the composite or those who have data collection constraints and cannot use the composite. The Committee therefore recommended continued endorsement of both the individual measures and the composite measure. Standing Committee Recommendation for Endorsement: Y-13; N-4 6. Public and Member Comment Comments received: Two commenters raised concern over the glucose control component of the composite, referencing the National Action Plan for Adverse Event Prevention, which was released in August, The National Action Plan states that the blood glucose threshold of <8% for patients <75 years of age does not conform to glycemic control guidelines from the American Diabetes Association, Department of Veterans Affairs, Department of Defense, and American Geriatrics Society (i.e., by excluding certain patients such as those with limited life expectancy or with certain co-morbid conditions, or by stratifying according to medication type). Developer response: According to the Institute for Clinical Systems Improvement (ICSI) 2014 Guidelines for Diabetes there is high quality evidence and a strong recommendation in support 13

14 of an A1c target of less than 8.0. Excerpt from the guideline is as follows: Algorithm Annotation #4- Glycemic Control and A1c Goals. Recommendation: A clinician should personalize goals with patients diagnosed with T2DM to achieve glycemic control with a hemoglobin A1c < 7% to < 8% depending on individual patient factors. Benefits: Achieving near-normal glycemic control lowers risk of diabetes microvascular complications such as retinopathy, nephropathy and amputations. Achieving A1c of 6.9 to 7.9% may also significantly reduce macrovascular complications based on Steno-2 and UKPDS data. Quality of Evidence: High Strength of Recommendation: Strong. Measurement does not and should not preclude good clinical judgement; however the measure development work group believes that a target of < 8.0 is reasonable and supported by guidelines. Our measure does have an upper age limit cut-off of 75 years and we allow exclusions for death, permanent nursing home resident or patients who are receiving hospice or palliative care services. Two commenters were critical of the composite measure itself, citing concern that use of the composite measure could mask the individual care processes that most need improvement. Developer response: While it is true that the measure is reported at the composite level, the individual components and the associated rates are available to the medical groups for better understanding their rates and for use in quality improvement to know which areas have opportunity for improvement. MNCM and the measure development work group firmly believe that achieving the intermediate physiological outcome targets related to blood pressure and glycemic control in addition being tobacco free and use of daily aspirin and statins where appropriate are the diabetic patient s best mechanisms of avoiding or postponing long term complications associated with this chronic condition which affects millions of Americans. Measuring providers separately on individual targets is not as patient centric as a measure that seeks to reduce multiple risk factors for each patient. Diabetic patients are more likely to reduce their overall risk and maximize health outcomes by achieving several intermediate physiological targets. Two commenters noted that documenting HbA1c levels >8% but less than 9% cannot be done using CPT-II coding, necessitating need for medical chart review. Developer response: A point of clarification, these measure do not rely on CPTII codes for numerator compliance, nor are they indicated anywhere in our measure specification. Measure specifications focus on the electronic health record as a source of clinical information for calculating numerator compliance; actual A1c values are utilized in the case of the A1c target. Additionally, 80 to 90% of all the clinics in MN are reporting this information from their electronic health records without the need for additional chart abstraction. One commenter suggested a need for including sociodemographic factors in the risk-adjustment approach. Developer response: Our risk adjustment model does include insurance product, which is a proxy for socioeconomic status. During the process of measure development, the expert panel discusses potential variables for risk adjustment that are important to consider for the measured population. For this measure, variables that are available for evaluation include gender, age, zip, race/ethnicity, country of origin, primary language, insurance product, diabetes type, depression and ischemic vascular disease. The potential risk adjustment variables are then evaluated for appropriate inclusion in the model based on a t value outside the range of -2.0 and Currently, the variables that have demonstrated acceptable properties are insurance product, age bands (18-25, 26-50, and 65 to 75) and diabetes type (1 or 2). Race/ethnicity has been collected for this measure in MN for the past few years, but has now reached a level of 14

15 reliability in which it can be evaluated for its impact. MNCM continues to review variables and their impact on the measure and part of its measure risk adjustment strategy. One commenter suggested the need for additional detail regarding moderate or high intensity in the description of statin use for the measure. Developer response: The measure development work group thoroughly discussed the pros and cons of specifying a certain dose of the statin medication and based on the following factors ultimately decided to not specify a dose of moderate or high intensity for numerator compliance: 1) data burden for practices, 2) controversy and burden surrounding the CV risk calculator, 3) ICSI 2014 Diabetes Guideline recommendations for measurement and 4) cardiology work group member s believe that there is some benefit for some patients who can only tolerate a lower intensity dose. Committee response: During its review of the individual measure assessing HbA1c<8% in the spring of 2014 (#0575), the Committee considered the clinical practice guideline recommendations from American Diabetes Association (2013), American Geriatric Society (2003), VA/DOD (2010), and American Association of Clinical Endocrinologists (AACE) (2011). During their discussion of that measure, members specifically noted that for some patients (e.g., frail elderly patients, those with limited life expectancy,) HbAc1 values slightly above 8% might be reasonable and that target HbA1c values for such patients should be individualized. In their more recent evaluation of the composite measure (#0729) in January 2015, the Committee considered the2014 clinical practice guideline recommendations from the Institute for Clinical Systems Improvement, which suggest a target threshold of <7% - <8%, depending on patient factors. The Committee acknowledged that the <8% threshold may not be appropriate for all patients but they agreed that the 8.0% cutoff was a reasonable target for a national healthcare performance measure and that 100% performance on the measure is not expected. The Committee also noted that measure #0729 includes an the upper age limit of 75 and excludes patients who died, are permanent nursing home residents, or are receiving hospice or palliative care services, which addresses at least some of the concerns voiced by the commenters. Finally, the Committee strongly recommended development of performance measures that assess occurrence and severity of hypoglycemia in the outpatient setting. Use of an all-or-none scoring approach does not hinder providers from tracking performance of the individual components of the composite and instituting appropriate improvement initiatives. Moreover, use of such composite measures is a patient-centric approach that allows providers to assess their success in reducing multiple patient risk factors across a variety of clinical areas. The Committee acknowledged NQF's recent policy change that has lifted the prohibition against including sociodemographic factors in risk-adjustment for outcome, resource use, and other quality measures. The Committee agreed that insurance type, which is included in the riskadjustment model for this measure, can be considered a proxy for sociodemographic status. The Committee also accepted the developer's explanation that other potential sociodemographic factors were considered for inclusion in the measure's risk-adjustment approach but ultimately were not included because they were not statistically significant. 7. Consensus Standards Approval Committee (CSAC) Review (May 12, 2015): Y-15; N-0; A-0 CSAC Decision: Approved for continued endorsement 15

16 8. Board of Directors Review: Yes (June 15, 2015) Board Decision: Ratified for continued endorsement 9. Appeals The Veterans Health Administration (VHA) lodged an appeal on this measure. The appeal was specific to the HbA1c component of the measure. The following points were included in the appeal: Application of the <8% threshold will expose patients years of age to the risk of adverse drug events, especially, but not limited to, hypoglycemia, without significant health benefits. Reliance upon the Institute for Clinical Systems Improvement (ICSI) guidelines has resulted in a compromised evidence review that is not consistent with the position of the DHHS National Action Plan for Prevention of Adverse Drug Events. The appellants argue that the ICSI guidelines presented by the developers are not as explicit in addressing older, more vulnerable populations and in identifying risk factors that justify A1c values greater than 8%. The exclusions specified in the measure (death, permanent nursing home placement, receipt of hospice or palliative care services) are not consistent with American Diabetes Association, American Geriatric Society, and Veterans Administration/Department of Defense guidelines, which recommend multiple exclusion criteria for a <8% target, including life expectancy less than 5 years, risk factors for hypoglycemia, prior hypoglycemia, food insufficiency, issues of health literacy and numeracy, etc. Use of insurance type as one of the factors included in the risk-adjustment approach cannot adequately address research findings indicating that low health literacy and low socioeconomic status is associated with risk of hypoglycemia among those who have insurance. The appellant also noted that food insufficiency is an increasingly recognized issue. Concerns about the accuracy of A1c testing results Lack of a balancing measure for high-risk patients Summary of the Developer Response: The developer addressed each of the issues raised by the appellant. They specifically included information about the blood glucose control guidelines from the American Diabetes Association, the American Geriatrics Society, and the Veterans Administration/Department of Defense. They also provided a brief discussion of the evolution of the measure as clinical practice guidelines changed, as well as information about their experience with the measure in Minnesota over the last decade. NQF Response: Both the VHA and CMS raised concern over the glucose control component of the composite during the public and member commenting period, and both referenced the National Action Plan for Adverse Event Prevention. In their discussion of the comments during the postcomment call, the Endocrine Standing Committee acknowledged that the <8% threshold may not be appropriate for all patients, but agreed that the 8.0% cutoff was a reasonable target for a national healthcare performance measure and that 100% performance on the measure is not expected. The Committee also briefly discussed the exclusions to the measure and recommended development of measures to assess occurrence and severity of hypoglycemia in the outpatient setting. The Committee did not discuss testing accuracy or the measure's riskadjustment approach during the post-comment call. 16

17 Consensus Standards Approval Committee (CSAC) Review (September, 2015): On September 8, 2015, the CSAC convened to discuss the appeal. The CSAC voted to uphold endorsement for the measure (with 92% approval). CSAC members acknowledged the appellants concerns, but remained supportive of endorsement, noting that the measure adequately reflects a synthesis of the guidelines and can reasonably accommodate individualized care for older patients who may not meet the 8% threshold. Additionally, CSAC members expressed satisfaction that the issues raised were addressed by both the developer and the Endocrine Standing Committee. Board of Directors Executive Committee (October, 2015): While the Executive Committee acknowledged the concerns raised in the appeal, it upheld the CSAC decision and voted unanimously to uphold endorsement of the measure. Members noted the exclusions that are applied to the measure and recognized the measure has been publicly reported in Minnesota since 2007 without any concerns having been raised by providers. However, they also noted the need to carefully consider older age groups and more frail individuals during measure development and called for development of balancing measures (e.g., hypoglycemia in the outpatient setting). Measures Withdrawn from Consideration One measure previously endorsed by NQF was not re-submitted for maintenance of endorsement in this evaluation cycle. The following measure is being retired by the developer and thus endorsement will be removed: Measure Measure Steward Reason for withdrawal Pharmacy Quality Alliance 0546: Diabetes: Appropriate Treatment of Hypertension PQA is retiring this measure due to the new guidelines published in JAMA by the panel members appointed to the Eighth Joint National Committee (JNC8). 17

18 Appendix B: Endocrine Portfolio Use in Federal Programs Neither of the two measures evaluated in this cycle of the project is currently used in federal quality improvement programs. 18

19 Appendix C: Project Standing Committee and NQF Staff STANDING COMMITTEE William Golden, MD, MACP (Co-Chair) Arkansas Medicaid Little Rock, AR James Rosenzweig, MD (Co-Chair) RTI International Boston, MA Robert Bailey, MD Janssen Scientific Affairs, LLC Venice, FL Tracey Breen, MD North Shore-LIJ Department of Medicine New Hyde Park, NY William Curry, MD, MS Pennsylvania State University College of Medicine Hershey Park, PA Vicky Ducworth The Boeing Company Charleston, SC R. James Dudl, MD Kaiser Permanente San Diego, CA Ingrid Duva, RN, PhD Veterans Administration Norcross, GA Starlin Haydon-Greatting, MS, BSPharm, FAPhA Illinois Pharmacists Association Springfield, IL Ann Kearns, MD, PhD Mayo Clinic Rochester, MN M. Sue Kirkman, MD University of North Carolina Chapel Hill, NC Anne Leddy, MD, FACE Glouchester-Mathews Free Clinic Moon, VA 19

20 Grace Lee, MD Virginia Mason Medical Center Seattle, WA Laura Makaroff, DO Health Resources Services Administration Rockville, MD Anna McCollister-Slipp Galileo Analytics Washington, DC Patricia McDermott, RN Aetna, Inc. St. Charles, IL Janice Miller, DNP, CRNP, CDE Thomas Jefferson University Philadelphia, PA Claudia Shwide-Slavin, MS, RD, DC-ADM, CDE American Association of Diabetes Educators (AADE) New York, NY Janet Sullivan, MD Westchester Medical Center Hawthorne, NY William Taylor, MD Department of Population Medicine, Harvard Medical School Boston, MA NQF STAFF Helen Burstin, MD, MPH Chief Scientific Officer Marcia Wilson, PhD, MBA Senior Vice President Karen Johnson, MS Senior Director Kathryn Streeter, MS Senior Project Manager Kaitlynn Robinson-Ector, MPH Project Analyst 20

21 Appendix D: Measure Specifications 0061 Comprehensive Diabetes Care Blood Pressure Control (<140/90 MM HG) STATUS Submitted STEWARD National Committee for Quality Assurance DESCRIPTION The percentage of patients years of age with diabetes (type 1 and type 2) whose most recent blood pressure level taken during the measurement year is <140/90 mm Hg. TYPE Outcome DATA SOURCE Administrative claims, Electronic Clinical Data : Electronic Clinical Data, Electronic Clinical Data : Electronic Health Record, Electronic Clinical Data : Pharmacy, Paper Medical Records. This measure is based on administrative claims and medical record documentation collected in the course of providing care to health plan patients. NCQA collects the Healthcare Effectiveness Data and Information Set (HEDIS) data for this measure directly from Health Management Organizations and Preferred Provider Organizations via NCQA s online data submission system. When collected by clinicians for the NCQA Diabetes Recognition Program, data is collected using the DRP Data Collection Tool (DCT) LEVEL Clinician : Individual, Clinician : Group/Practice, Health Plan, Integrated Delivery System SETTING Ambulatory Care : Clinician Office/Clinic TIME WINDOW The measurement year (12 month period) NUMERATOR STATEMENT The outcome being measured is a blood pressure reading of <140/90 mm Hg, which indicates adequately controlled blood pressure. Adequately controlled blood pressure in patients with diabetes reduces cardiovascular risks and microvascular diabetic complications. NUMERATOR DETAILS ADMINISTRATIVE Use automated data to identify the most recent blood pressure reading taken during an outpatient visit or non-acute inpatient encounter during the measurement year. The patient is 21

22 numerator compliant if the blood pressure reading is <140/90 mm Hg. The patient is not numerator compliant if the blood pressure is =140/90 mm Hg, if there is no blood pressure reading during the measurement year or if the reading is incomplete (e.g. the systolic or the diastolic level reading is missing). If there are multiple blood pressures on the same date of service, use the lowest systolic and the lowest diastolic blood pressure as the representative blood pressure. Organizations that use the CPT Category II codes to identify numerator compliance must search for all codes in the following value sets and use the most recent codes during the measurement year to determine numerator compliance for both the systolic and diastolic levels: VALUE SET / NUMERATOR COMPLIANCE Systolic Less than 140 Value Set / Systolic compliant Systolic Greater Than/Equal to 140 Value Set / Systolic not compliant Diastolic Less than 80 Value Set / Diastolic compliant Diastolic Value Set / Diastolic Compliant Diastolic Greater Than/Equal to 90 Value Set / Diastolic Not Compliant MEDICAL RECORD The organization should use the medical record that it uses to collect data for other diabetes care indicators such as the HbA1c <8 mg/dl indicator. If the organization does not collect data for other diabetes care indicators, it should use the medical record of the provider that manages the patient s diabetes. If that medical record does not contain a blood pressure, the organization may use the medical record of another primary care provider or specialist from whom the patient receives care. To determine if blood pressure is adequately controlled, the organization must identify the representative blood pressure following the steps below. Identify the most recent blood pressure reading noted during the measurement year. DO NOT include blood pressure readings that meet the following criteria: -Taken during an acute inpatient stay or an ED visit. -Taken during an outpatient visit which was for the sole purpose of having a diagnostic test or surgical procedure performed (e.g., sigmoidoscopy, removal of a mole). -Obtained the same day as a major diagnostic or surgical procedure (e.g., stress test, administration of IV contrast for a radiology procedure, endoscopy). -Reported by or taken by the patient. Identify the lowest systolic and lowest diastolic blood pressure reading from the most recent blood pressure notation in the medical record. If there are multiple BPs recorded for a single date, use the lowest systolic and lowest diastolic BP on that date as the representative BP. The systolic and diastolic results do not need to be from the same reading when multiple readings are recorded for a single date. The patient is not numerator compliant if the BP does not meet the specified threshold or is missing, or if there is no BP reading during the measurement year or if the reading is incomplete (i.e., the systolic or diastolic level is missing). DENOMINATOR STATEMENT Patients years of age by the end of the measurement year who had a diagnosis of diabetes (type 1 and type 2) during the measurement year or the year prior to the measurement year. See question S.9 Denominator Details for methods to identify patients with diabetes. 22