Measure Description: The percentage of patients 18-75 years of age who had a diagnosis of ischemic vascular disease (IVD) and whose IVD was optimally managed during the measurement period as defined by achieving ALL of the following:
• Blood pressure less than 140/90 mmHg
• On a statin medication, unless allowed contraindications or exceptions are present
• Non-tobacco user
• On daily aspirin or anti-platelet medication, unless allowed contraindications or exceptions are present
Exclusions
The following exclusions are allowed to be applied to the eligible population:
- Patient was in hospice or receiving palliative care at any time prior to the end of the measurement period.
- Patient died any time prior to the end of the measurement period
- Patient had only urgent care visits during the measurement period
Measure Specs
- General Information(active tab)
- Numerator
- Denominator
- Exclusions
- Measure Calculation
- Supplemental Attachment
- Point of Contact
General Information
Since 1900, cardiovascular disease (CVD) has consistently ranked as the leading cause of death in the United States, with the exception of 1918. According to the American Heart Association, recent estimates suggest that as of 2020, approximately 121.5 million adults in the United States are living with some form of cardiovascular disease. Furthermore, CVD accounted for 1 in every 3.1 deaths in the United States in 2019. Enhanced adherence to current evidence-based treatment guidelines has been shown to correlate with reduced morbidity and mortality rates associated with this disease.
The data used to generate the results for this measure are collected from medical groups that submit clinical data directly to MNCM. Data submission requirements are specified by MNCM in the measure specification guides that provide detailed steps and instructions to ensure providers submit data in a standard format.
MNCM is in the midst of transitioning its data collection for the clinical quality measures reported by medical groups to a modernized system known as PIPE that reduces quality measurement burden on health care providers and enables more timely feedback on performance. The previous data collection system, known as Direct Data Submission or DDS, required providers to separately identify the relevant population for each measure. The transition to the new system is expected to be complete by the end of 2024.
The PIPE and DDS portals open in January each year for data submission for official calculation of MNCM’s publicly reported clinical quality measures. Data submission in DDS happens annually between January and March, whereas data submission in PIPE can occur throughout the year (e.g. annually, monthly, quarterly). Although medical groups in PIPE can submit data and calculate their own performance as often as they like, official calculation of MNCM’s quality measures happens once a year in February.
MNCM completes a rigorous validation of the data to confirm data completeness, reliability, and ensure that the submitted data matches the patient record in the electronic health records (EHR).
Numerator
The number of patients in the denominator whose IVD was optimally managed during the measurement period as defined by achieving ALL of the following:
• The most recent blood pressure in the measurement period has a systolic value of less than 140 mmHg AND a diastolic value of less than 90 mmHg
• On a statin medication, unless allowed contraindications or exceptions are present
• Patient is not a tobacco user
• On daily aspirin or anti-platelet medication, unless allowed contraindications or exceptions are present
The number of patients in the denominator whose IVD was optimally managed during the measurement period as defined by achieving ALL of the following:
• The most recent blood pressure in the measurement period has a systolic value of less than 140 mmHg AND a diastolic value of less than 90 mmHg
• On a statin medication, unless allowed contraindications or exceptions are present
• Patient is not a tobacco user
• On daily aspirin or anti-platelet medication, unless allowed contraindications or exceptions are present
See excel file for data dictionary and value sets.
Denominator
Patients ages 18 years or older at the start of the measurement period AND less than 76 years at the end of the measurement period who have a diagnosis of ischemic vascular disease (Ischemic Vascular Disease Value Set) with any contact during the current or prior measurement period OR had ischemic vascular disease (Ischemic Vascular Disease Value Set) present on an active problem list at any time during the measurement period.
Both contacts AND the active problem list must be queried for diagnosis (Ischemic Vascular Disease)
AND
At least one established patient office visit (Established Pt Diabetes & Vasc Value Set) performed or supervised by an eligible provider in an eligible specialty for any reason during the measurement period.
Patients ages 18 years or older at the start of the measurement period AND less than 76 years at the end of the measurement period who have a diagnosis of ischemic vascular disease (Ischemic Vascular Disease Value Set) with any contact during the current or prior measurement period OR had ischemic vascular disease (Ischemic Vascular Disease Value Set) present on an active problem list at any time during the measurement period.
Both contacts AND the active problem list must be queried for diagnosis (Ischemic Vascular Disease)
AND
At least one established patient office visit (Established Pt Diabetes & Vasc Value Set) performed or supervised by an eligible provider in an eligible specialty for any reason during the measurement period.
See excel file for data dictionary and value sets.
Exclusions
The following exclusions are allowed to be applied to the eligible population: Hospice or palliative care services, or died prior to the end of the measurement period.
- Patient was in hospice or receiving palliative care at any time during the measurement period
- Patient died prior to the end of the measurement period
Measure Calculation
See attached material page 6 and 7
The measure for patients with ischemic vascular disease is stratified by individual component rates, comparison to pre-pandemic rates, rate variation by medical group (boxplot) and rate variation by three digit zip code groupings of patient residence.
The measure results are stratified by race, ethnicity, language, country of insurance (RELC), and insurance product type in our annual community reports, which can be found here: https://mncm.org/reports/#community-reports
In 2022 it was observed that black, indigenous/native, or multi racial have significantly lower rates of optimal care, while Asian and white individuals had significantly higher rates of optimal care. Ethnicity of Hispanic/Latinx have significantly lower rates of optimal care.
MNCM calculates final rates for all medical group/clinics that submitted data and whose data has been validated. However, for public reporting purposes, MNCM only publicly report results for medical groups/clinics that have a minimum of 30 patients in the denominator.
Supplemental Attachment
Point of Contact
N/A
United States
United States
Importance
Evidence
The Optimal Vascular Care (OVC) measure, a composite metric encompassing various aspects of vascular health management, reflects a critical shift in guidelines for hypertension diagnosis and management. The American College of Cardiology and American Heart Association's revised guidelines redefined hypertension thresholds, increasing the number of individuals classified as hypertensive. This change has prompted debate within the medical community, particularly concerning the appropriateness of lower blood pressure targets for all patients. Moreover, disparities in guideline endorsement among medical organizations have further complicated the landscape of hypertension management, leaving healthcare providers uncertain about the most effective approaches to care. Against this backdrop, the OVC measure serves as a vital tool for assessing and improving vascular health outcomes, albeit amidst ongoing controversies and challenges surrounding hypertension management.
Informed by the evolving evidence and guideline recommendations, the development of the OVC measure involved extensive deliberation and consideration of various factors influencing vascular health outcomes. Notably, the measure targets modifiable risk factors, including blood pressure control, aspirin use, statin therapy, and tobacco cessation, which are crucial for reducing cardiovascular risk in high-risk populations. However, the implementation of these targets presents challenges, particularly in achieving optimal blood pressure control while balancing the risks and benefits of intensive treatment. Furthermore, disparities in care delivery across demographic and geographic lines underscore the need for targeted interventions to address inequities in vascular health outcomes. Overall, the OVC measure provides a comprehensive framework for evaluating and improving vascular care delivery, offering insights into areas for improvement and guiding efforts to optimize cardiovascular health outcomes for all patients.
References:
American College of Cardiology/ American Heart Association Guidelines for the Prevention, Detection, Evaluation and Management of High Blood Pressure in Adults November 13, 2017
Saiz LC, Gorricho J, Garjón J, Celaya MC, Erviti J, Leache L. Blood pressure targets for the treatment of people with hypertension and cardiovascular disease. Cochrane Database Syst Rev. 2022 Nov 18;11(11):CD010315. doi: 10.1002/14651858.CD010315.pub5. PMID: 36398903; PMCID: PMC9673465.
Averbeck, B., Ballard, S., Collins, D., & Detert, D. (2018, June). Hypertension work group: 2018 commentary | ICSI. ICSI Hypertension Work Group: 2018 Commentary. https://www.icsi.org/guideline/hypertension-work-group-2018-commentary/
Kruger J, O'Halloran A, Rosenthal AC, Babb SD, Fiore MC. Receipt of evidence-based brief cessation interventions by health professionals and use of cessation assisted treatments among current adult cigarette-only smokers: National Adult Tobacco Survey, 2009-2010. BMC Public Health. 2016 Feb 11;16:141. doi: 10.1186/s12889-016-2798-2. PMID: 26868930; PMCID: PMC4751655.
Ridker PM, Mora S, Rose L; JUPITER Trial Study Group. Percent reduction in LDL cholesterol following high-intensity statin therapy: potential implications for guidelines and for the prescription of emerging lipid-lowering agents. Eur Heart J. 2016 May 1;37(17):1373-9. doi: 10.1093/eurheartj/ehw046. Epub 2016 Feb 24. PMID: 26916794; PMCID: PMC4852064.
Arnett, Donna K., et al. "2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines." Circulation 140.11 (2019): e563-e595.
The evidence is directly quoted and summarized below:
Reference One:
ACC/ AHA Guidelines for the Diagnosis and Management of Hypertension-
Review by MNCM Measure Development Workgroup
Background:
In November of 2017, the American Academy of Cardiology and American Heart Association released new guidelines for the prevention, diagnosis and management of hypertension in adults.[1] These guidelines redefined the diagnosis of hypertension moving from > 140/ 90 to a new definition of stage 1 hypertension (130-139/ 80-89). With new definition, it is estimated that 46% of Americans will now be considered to have hypertension, up from 32% with a definition of > 140/90. The release of the guidelines is not without controversy, and while most agree that a lower blood pressure is better, it is within the context of a patient’s individualized goal. Less than 130/80 may not be an appropriate target for every patient. The American College of Physicians and the American Academy of Family Practice has declined endorsement of the new guidelines. They cite concerns with the methodology used in making recommendations and perceived conflict of interest. They are recommending reliance on 2014 JNC8 and ACP/AAFP guidelines for older adults.
Patients with diabetes and cardiovascular disease represent two very high-risk subgroups; in an effort to reduce their modifiable risk factors, the blood pressure component target of the Optimal Diabetes Care (ODC) and Optimal Vascular Care (OVC) measures has reflected a goal that is below the hypertension definition cut-point.
In similar measure development activities, the National Committee for Quality Assurance (NCQA) convened three expert panels (diabetes, cardiovascular and geriatric) for their evaluation of blood pressure targets for the HEDIS Controlling High Blood Pressure measure and concluded that for patients with hypertension ages 18 - 85 the blood pressure target is < 140/90.
An MNCM convened measure development multi-stakeholder group met in April of 2018 to evaluate and discuss recent changes in guidelines and evidence surrounding blood pressure targets for patients with diabetes and vascular disease. Based on this evaluation, determine the best BP component targets for the composite measures.
Considerable time was devoted to the discussion of the evidence supporting guidelines, applicability of research studies into clinical practice, risk-harm benefit and need for individualized patient goals. After thoughtful and thorough discussion of current guidelines, evidence, and real-world practice implications, the work group gained consensus on the best BP targets for patients with diabetes and vascular disease.
Key considerations included:
Evaluation of SPRINT (Systolic Blood Pressure Intervention Trial) demonstrates that for the primary outcome of mortality, there is only a 0.5% difference between the intensive treatment group and the standard treatment group. Generalization of the SPRINT results to every day practice raised the issues of:
SPRINT study design called for the withdrawal of treatment in asymptomatic patients in the conservative treatment arm, which does not match clinical practice
Average systolic BP achieved was 121
Best practice methods to obtain BP in a study (auto-BP machine, quiet setting, and resting 10 min) do not match current clinic practice.
The SPRINT study excluded patients with diabetes, so its results are not transferable when there is direct evidence from the ACCORD study that is applicable. ACCORD (Action to Control Cardiovascular Disease in Diabetes) showed very little benefit for BP targets < 140/90.
Guidelines do not address treatment risks (hypotension, kidney function). The main concern of the workgroup was that in setting a lower target for all patients to strive for, knowing that providers will want to meet that target and may be accountable for hitting that target, may put some patients at risk for serious and costly side effects of intensive treatment. The workgroup would like to encourage individualized targets, knowing that a lower blood pressure is better for the patient, but only if it can be achieved safely.
There is not consensus at this time among the guideline writing groups about the definition of hypertension or appropriate targets for high risk populations like patients with diabetes or ischemic vascular disease, therefore not a clear direction for measurement to align with guidelines.
Measure Development Workgroup Recommendation:
American College of Cardiology/ American Heart Association Guidelines for the Prevention, Detection, Evaluation and Management of High Blood Pressure in Adults November 13, 2017
Reference Two:
Cochrane Review 2022: Blood pressure targets for the treatment of people with hypertension and cardiovascular disease
Background: This is the third update of the review first published in 2017. Hypertension is a prominent preventable cause of premature morbidity and mortality. People with hypertension and established cardiovascular disease are at particularly high risk, so reducing blood pressure to below standard targets may be beneficial. This strategy could reduce cardiovascular mortality and morbidity but could also increase adverse events. The optimal blood pressure target in people with hypertension and established cardiovascular disease remains unknown.
Objectives: To determine if lower blood pressure targets (systolic/diastolic 135/85 mmHg or less) are associated with reduction in mortality and morbidity compared with standard blood pressure targets (140 mmHg to 160mmHg/90 mmHg to 100 mmHg or less) in the treatment of people with hypertension and a history of cardiovascular disease (myocardial infarction, angina, stroke, peripheral vascular occlusive disease).
Search methods: For this updated review, we used standard, extensive Cochrane search methods. The latest search date was January 2022. We applied no language restrictions.
Selection criteria: We included randomized controlled trials (RCTs) with more than 50 participants per group that provided at least six months' follow-up. Trial reports had to present data for at least one primary outcome (total mortality, serious adverse events, total cardiovascular events, cardiovascular mortality). Eligible interventions involved lower targets for systolic/diastolic blood pressure (135/85 mmHg or less) compared with standard targets for blood pressure (140 mmHg to 160 mmHg/90 mmHg to 100 mmHg or less). Participants were adults with documented hypertension and adults receiving treatment for hypertension with a cardiovascular history for myocardial infarction, stroke, chronic peripheral vascular occlusive disease, or angina pectoris.
Data collection and analysis: We used standard Cochrane methods. We used GRADE to assess the certainty of the evidence.
Main results: We included seven RCTs that involved 9595 participants. Mean follow-up was 3.7 years (range 1.0 to 4.7 years). Six of seven RCTs provided individual participant data. None of the included studies was blinded to participants or clinicians because of the need to titrate antihypertensive drugs to reach a specific blood pressure goal. However, an independent committee blinded to group allocation assessed clinical events in all trials. Hence, we assessed all trials at high risk of performance bias and low risk of detection bias. We also considered other issues, such as early termination of studies and subgroups of participants not predefined, to downgrade the certainty of the evidence. We found there is probably little to no difference in total mortality (risk ratio (RR) 1.05, 95% confidence interval (CI) 0.91 to 1.23; 7 studies, 9595 participants; moderate-certainty evidence) or cardiovascular mortality (RR 1.03, 95% CI 0.82 to 1.29; 6 studies, 9484 participants; moderate-certainty evidence). Similarly, we found there may be little to no differences in serious adverse events (RR 1.01, 95% CI 0.94 to 1.08; 7 studies, 9595 participants; low-certainty evidence) or total cardiovascular events (including myocardial infarction, stroke, sudden death, hospitalization, or death from congestive heart failure (CHF)) (RR 0.89, 95% CI 0.80 to 1.00; 7 studies, 9595 participants; low-certainty evidence). The evidence was very uncertain about withdrawals due to adverse effects. However, studies suggest more participants may withdraw due to adverse effects in the lower target group (RR 8.16, 95% CI 2.06 to 32.28; 3 studies, 801 participants; very low-certainty evidence). Systolic and diastolic blood pressure readings were lower in the lower target group (systolic: mean difference (MD) -8.77 mmHg, 95% CI -12.82 to -4.73; 7 studies, 8657 participants; diastolic: MD -4.50 mmHg, 95% CI -6.35 to -2.65; 6 studies, 8546 participants). More drugs were needed in the lower target group (MD 0.56, 95% CI 0.16 to 0.96; 5 studies, 7910 participants), but blood pressure targets at one year were achieved more frequently in the standard target group (RR 1.20, 95% CI 1.17 to 1.23; 7 studies, 8699 participants).
Authors' conclusions: We found there is probably little to no difference in total mortality and cardiovascular mortality between people with hypertension and cardiovascular disease treated to a lower compared to a standard blood pressure target. There may also be little to no difference in serious adverse events or total cardiovascular events. This suggests that no net health benefit is derived from a lower systolic blood pressure target. We found very limited evidence on withdrawals due to adverse effects, which led to high uncertainty. At present, evidence is insufficient to justify lower blood pressure targets (135/85 mmHg or less) in people with hypertension and established cardiovascular disease. Several trials are still ongoing, which may provide an important input to this topic in the near future.
Saiz LC, Gorricho J, Garjón J, Celaya MC, Erviti J, Leache L. Blood pressure targets for the treatment of people with hypertension and cardiovascular disease. Cochrane Database Syst Rev. 2022 Nov 18;11(11):CD010315. doi: 10.1002/14651858.CD010315.pub5. PMID: 36398903; PMCID: PMC9673465. (https://pubmed.ncbi.nlm.nih.gov/36398903/)
Reference Three:
Institute for Clinical Systems Improvement Hypertension Workgroup Commentary 2018
Excerpts-
Background- Reception to the new ACC/AHA guideline has been mixed. Differing interpretations of the same body of evidence has led to conflicting recommendations. The American College of Physicians (ACP) and American Academy of Family Physicians (AAFP) did not endorse the new ACC/AHA guideline. Notably, ACP and AAFP published a guideline in January 2017 recommending a goal of less than 150/90 mm Hg for adults over age 60. The American Diabetes Association recommends treatment to a BP < 140/90 mm Hg for most patients with diabetes and consideration of a target < 130/80 for those at high cardiovascular risk if it can be achieved without undue burden. As the controversy continues, providers are left wondering how to advise patients.
Challenges Ahead- Measurement To understand blood pressure on a population level, it may be most useful to look at the distribution curve of blood pressures across the population. This provides a more detailed picture of the problem, which then helps direct intervention. The work group agrees with a blood pressure goal of less than 130/80 mm Hg for the general population, to be adjusted as needed for the individual. However, the group has significant concerns with using less than 130/80 mm Hg as an accountability target because it might result in pharmacologic therapy for some patients who are at low cardiovascular risk and should only be treated by lifestyle modifications. The group agrees that because of the individualized nature of hypertension management, flexibility with measurement is critical.
https://www.icsi.org/wp-content/uploads/2019/01/ICSI-HTN-Work-Group-2018-CommentaryUpdated071718.pdf
Averbeck, B., Ballard, S., Collins, D., & Detert, D. (2018, June). Hypertension work group: 2018 commentary | ICSI. ICSI Hypertension Work Group: 2018 Commentary. https://www.icsi.org/guideline/hypertension-work-group-2018-commentary/
Reference Four: Receipt of evidence-based brief cessation interventions by health professionals and use of cessation assisted treatments among current adult cigarette-only smokers: National Adult Tobacco Survey, 2009-2010
Background: Helping tobacco smokers to quit during a medical visit is a clinical and public health priority. Research suggests that most health professionals engage their patients in at least some of the '5 A's' of the brief cessation intervention recommended in the U.S. Public Health Service Clinical Practice Guideline, but information on the extent to which patients act on this intervention is uncertain. We assessed current cigarette-only smokers' self-reported receipt of the 5 A's to determine the odds of using optimal cessation assisted treatments (a combination of counseling and medication).
Methods: Data came from the 2009-2010 National Adult Tobacco Survey (NATS), a nationally representative landline and mobile phone survey of adults aged ≥18 years. Among current cigarette-only smokers who visited a health professional in the past 12 months, we assessed patients' self-reported receipt of the 5 A's, use of the combination of counseling and medication for smoking cessation, and use of other cessation treatments. We used logistic regression to examine whether receipt of the 5 A's during a recent clinic visit was associated with use of cessation treatments (counseling, medication, or a combination of counseling and medication) among current cigarette-only smokers.
Results: In this large sample (N = 10,801) of current cigarette-only smokers who visited a health professional in the past 12 months, 6.3 % reported use of both counseling and medication for smoking cessation within the past year. Other assisted cessation treatments used to quit were: medication (19.6 %); class or program (3.8 %); one-on-one counseling (3.7 %); and telephone quitline (2.6 %). Current cigarette-only smokers who reported receiving all 5 A's during a recent clinic visit were more likely to use counseling (odds ratio [OR]: 11.2, 95 % confidence interval [CI]: 7.1-17.5), medication (OR: 6.2, 95 % CI: 4.3-9.0), or a combination of counseling and medication (OR: 14.6, 95 % CI: 9.3-23.0), compared to smokers who received one or none of the 5 A's components.
Conclusions: Receipt of the '5 A's' intervention was associated with a significant increase in patients' use of recommended counseling and medication for cessation. It is important for health professionals to deliver all 5 A's when conducting brief cessation interventions with patients who smoke.
Kruger J, O'Halloran A, Rosenthal AC, Babb SD, Fiore MC. Receipt of evidence-based brief cessation interventions by health professionals and use of cessation assisted treatments among current adult cigarette-only smokers: National Adult Tobacco Survey, 2009-2010. BMC Public Health. 2016 Feb 11;16:141. doi: 10.1186/s12889-016-2798-2. PMID: 26868930; PMCID: PMC4751655.
Reference Five: Percent reduction in LDL cholesterol following high-intensity statin therapy: potential implications for guidelines and for the prescription of emerging lipid-lowering agents
Aims: Current statin guidelines in Europe and Canada advocate achieving a fixed LDL target or the attainment of a ≥50% reduction in low-density lipoprotein cholesterol (LDLC), while current US guidelines advocate the use of statin therapies that reduce LDLC by <50% (moderate intensity) or ≥50% (high intensity). Data are limited, however, linking the achievement of these % reduction thresholds to subsequent cardiovascular outcomes particularly for contemporary high-intensity regimens.
Methods and results: In a randomized trial of 17 082 initially healthy men and women with median baseline LDLC of 108 mg/dL (interquartile range 94-119), we (i) used waterfall plots to assess the variability in LDLC response to rosuvastatin 20 mg daily and (ii) evaluated the impact of reaching ≥50% reductions in LDLC on risk of developing the first cardiovascular events. Among rosuvastatin allocated participants, 3640 individuals (46.3%) experienced an LDLC reduction ≥50%; 3365 individuals (42.8%) experienced an LDLC reduction >0 but <50%; and 851 individuals (10.8%) experienced no reduction or an increase in LDLC compared with baseline. These % LDLC reductions directly related to the risks of first cardiovascular events; at trial completion, incidence rates for the primary endpoint were 11.2, 9.2, 6.7, and 4.8 per 1000 person-years for those in the placebo, no LDLC reduction, LDLC reduction <50%, and LDLC reduction ≥50% groups, respectively. Compared with placebo, the multivariable adjusted hazard ratios for sequentially greater on-treatment per cent reductions in LDLC were 0.91 (95%CI 0.54-1.53), 0.61 (95%CI 0.44-0.83), and 0.43 (95%CI 0.30-0.60) (P < 0.00001). Similar relationships between % reduction and clinical outcomes were observed in analyses focusing on non-HDLC or apolipoprotein B.
Conclusions: As documented for low- and moderate-intensity regimens, variability in % LDLC reduction following high-intensity statin therapy is wide yet the magnitude of this % reduction directly relates to efficacy. These data support guideline approaches that incorporate % reduction targets for statin therapy as well as absolute targets, and might provide a structure for the allocation of emerging adjunctive lipid-lowering therapies such as PCSK9 inhibitors should these agents prove broadly effective for cardiovascular event reduction.
Ridker PM, Mora S, Rose L; JUPITER Trial Study Group. Percent reduction in LDL cholesterol following high-intensity statin therapy: potential implications for guidelines and for the prescription of emerging lipid-lowering agents. Eur Heart J. 2016 May 1;37(17):1373-9. doi: 10.1093/eurheartj/ehw046. Epub 2016 Feb 24. PMID: 26916794; PMCID: PMC4852064.
Reference Six: 2019 ACC/AHA Guideline on the Primary Prevention of Cardiovascular Disease: Executive Summary
TOP 10 TAKE-HOME MESSAGES FOR THE PRIMARY PREVENTION OF CARDIOVASCULAR DISEASE
1. The most important way to prevent atherosclerotic vascular disease, heart failure, and atrial fibrillation is to promote a healthy lifestyle throughout life.
2. A team-based care approach is an effective strategy for the prevention of cardiovascular disease. Clinicians should evaluate the social determinants of health that affect individuals to inform treatment decisions.
3. Adults who are 40 to 75 years of age and are being evaluated for cardiovascular disease prevention should undergo 10-year atherosclerotic cardiovascular disease (ASCVD) risk estimation and have a clinician–patient risk discussion before starting on pharmacological therapy, such as antihypertensive therapy, a statin, or aspirin. The presence or absence of additional risk-enhancing factors can help guide decisions about preventive interventions in select individuals, as can coronary artery calcium scanning.
4. All adults should consume a healthy diet that emphasizes the intake of vegetables, fruits, nuts, whole grains, lean vegetable or animal protein, and fish and minimizes the intake of trans fats, red meat and processed red meats, refined carbohydrates, and sweetened beverages. For adults with overweight and obesity, counseling and caloric restriction are recommended for achieving and maintaining weight loss.
5. Adults should engage in at least 150 minutes per week of accumulated moderate-intensity physical activity or 75 minutes per week of vigorous-intensity physical activity.
6. For adults with type 2 diabetes mellitus, lifestyle changes, such as improving dietary habits and achieving exercise recommendations are crucial. If medication is indicated, metformin is first-line therapy, followed by consideration of a sodiumglucose cotransporter 2 inhibitor or a glucagonlike peptide-1 receptor agonist.
7. All adults should be assessed at every healthcare visit for tobacco use, and those who use tobacco should be assisted and strongly advised to quit.
8. Aspirin should be used infrequently in the routine primary prevention of ASCVD because of lack of net benefit.
9. Statin therapy is first-line treatment for primary prevention of ASCVD in patients with elevated low-density lipoprotein cholesterol levels (≥190 mg/dL), those with diabetes mellitus, who are 40 to 75 years of age, and those determined to be at sufficient ASCVD risk after a clinician–patient risk discussion.
10. Nonpharmacological interventions are recommended for all adults with elevated blood pressure or hypertension. For those requiring pharmacological therapy, the target blood pressure should generally be <130/80 mm Hg.
Arnett, Donna K., et al. "2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines." Circulation 140.11 (2019): e563-e595.
Measure Impact
The evidence supporting the value and meaningfulness of the measured is drawn from extensive data collection and reporting spanning multiple years, both pre-pandemic and during the ongoing health crisis. This longitudinal approach enables us to observe the delivery of care to patients over time, at various levels ranging from state to medical group, clinic, and geographical region.
Specifically, the observed 5% decrease in the OVC measure between 2019 and 2022 highlights areas for potential improvement in the management of this complex disease state. This trend underscores the importance of continuous monitoring and evaluation to identify areas of success and areas needing enhancement for both providers and patients.
Moreover, the significant variation in OVC across different medical groups and geographical regions provides valuable insights into the disparities in care delivery for patients with IVD. This comprehensive view empowers stakeholders to target actionable interventions aimed at improving outcomes and addressing inequities in healthcare delivery.
On an annual basis the MNCM Measurement and Reporting Committee (MARC) reviews each measure for continued suitability for public reporting and rates the measure against several NQF criteria including importance to measure with continued opportunity for improvement and feasibility. In September 2023 MARC reviewed the OVC measure and agreed that MNCM should continue collecting, aggregating, and reporting composite measure without changes. 89.5% (17) of participating members agreed with motion. Two members opposed the motion, with one member commenting on their preference to transition the measure to monitoring.
Performance Gap
72 medical groups submitted data on over 184,000 patients with ischemic vascular disease. 55.3% of the patients met all four component targets in the composite measure and were considered optimally managed. Of the clinics that were reportable (patient n >= 30), there was a wide range of variability with the lowest scoring clinic at 20% and the highest scoring clinic at 68.25%.
The table below describes the rates for OVC and the individuals and medical groups that were involved in the 2022 calculations.
The supplemental file describes the average rates, std, and confidence intervals for the OVC measure over the previous five years.
| Overall | Minimum | Decile_1 | Decile_2 | Decile_3 | Decile_4 | Decile_5 | Decile_6 | Decile_7 | Decile_8 | Decile_9 | Decile_10 | Maximum | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mean Performance Score | 0.5526226 | 0 | 0 | 0 | 0 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | 1 |
| N of Entities | 78 | 1 | 8.7 | 16.4 | 24.1 | 31.8 | 39.5 | 47.2 | 54.9 | 62.6 | 70.3 | 78 | 78 |
| N of Persons / Encounters / Episodes | 184417 | 1 | 18442.6 | 63884.2 | 55325.8 | 73767.4 | 92209 | 110650.6 | 129092.2 | 147533.8 | 165975.4 | 184417 | 184417 |
Equity
Equity
Not Applicable
Feasibility
Feasibility
This measure is collected and reported for providers who care for patients with ischemic vascular disease (e.g., family medicine, internal medicine, and cardiology) whose medical group submits patient level data for rate calculation.
Periodically, MNCM surveys the medical groups about their participation in this process. In 2018, 65.3% of medical groups rated the level of difficulty in obtaining the data needed for patient level submission as “Easy” or “Very Easy” (66/101). While these are older results there have not been any substantial changes to the measure that would have caused these results to change.
All measures are calculated off data that is uploaded into MNCM internal data system, PIPE. Medical groups receive extensive training and onboarding with PIPE including (more information on PIPE is in the validation section)
- Data dictionary for each table of data (encounters, blood pressures, lab values, medications, etc...).
- 1 to1 staff support during onboarding process including staging and production portals with dual validation of measure rate calculation
- Extensive data element level audit to ensure accuracy as compared to the electronic health record
No changes were made
Proprietary Information
Scientific Acceptability
Testing Data
Medical groups submit data to MNCM through the PIPE Portal where data is required to be submitted in a standardized format.
In the data dictionary attached there is a tab titled "Field Specifications" that reviews the format of all data submitted.
This data represents many medical groups and clinics across the entirety of Minnesota.
None
2023 with 2022 Dates of Service (01/01/2022-12/31/2022)]
78 medical groups, 643 physician clinics, 184,417 patients with IVD in Minnesota and neighboring communities submitted data for this measure.
The Optimal Vascular Care (OVC) measure encompasses several descriptive characteristics that provide insight into its composition and variation.
The supplemental material titled "Eligible Population Characteristics" contains tables with information on all the characteristics to be reviewed. As a composite measure, it integrates multiple areas of interest, including blood pressure control, daily aspirin use, statin use, and tobacco use. The tab titled "2022 composite parts" displays the rates across the past five years for each component of the measure. Notably, the rate for the Daily Aspirin Use Component significantly decreased in 2022 compared to 2021 (-2.5 percentage points), while the rate for the Blood Pressure Control component significantly increased in 2022 compared to 2021 (+0.6 percentage points).
Furthermore, analysis reveals variations based on three-digit zip code, age, gender, and race, as illustrated in tabs "3 digit zip code," "Age," "Gender," and "Race," respectively. Looking regionally using the three-digit zip code: seven regions had OVC rates that were significantly below the Minnesota resident average, while four regions had rates that were significantly above. This also displayed differences between rural and urban areas.
Regarding demographics: 55% of the patient population was between the ages of 66-75, 39% were 51-65, with the final 6% of the population being between 18-50. 65% of the population was born male, and 90% of the population was white.
These descriptive characteristics offer a comprehensive understanding of the OVC measure's performance and its distribution across different demographic factors.
Reliability
The reliability of each measure in this report was assessed using the Beta-binomial model (Adams, 2009)[1], which measures signal-to-noise and estimates the proportion of overall variability explained by differences between entities. This approach aims to conceptualize score reliability as the ratio of signal to noise, where the signal represents the variability in measured performance explained by real differences in performance. Ensuring measure score reliability is crucial as it enhances the ability to distinguish genuine differences between measured entities, reducing the likelihood of misclassification in comparative performance.
Used paper “Reliability in Provider Profiling” by John L. Adams, Ph.D as a reference
The BETABIN macro was used on each measure (SAS).
- Provider-to-provider variance: σ2 = (α β) / (α + β + 1)(α + β)2
- Reliability = σ2 / (σ2 + (p(1 – p)/n))
- p = rate
- n = number of eligible patients
- Determine reliability rate for each provider.
- Average the reliability rate.
For reliability for OVC, provider in this case is a medical group that has 30 or more patients.
[1] Adams, John L. "The reliability of provider profiling: a tutorial." (2009).
The reliability rate of the OVC measure was 0.89 for 2022.
The reliability rate of the OVC measure was 0.89. This demonstrates a strong indication that variability is attributable to differences in performance.
| | Overall | Minimum | Decile_1 | Decile_2 | Decile_3 | Decile_4 | Decile_5 | Decile_6 | Decile_7 | Decile_8 | Decile_9 | Decile_10 | Maximum |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Reliability | 0.89108479 | 0.431545919 | 0.71 | 0.823813 | 0.886 | 0.915 | 0.93 | 0.957 | 0.9739 | 0.9827 | 0.99666 | 0.99281 | 0.999281467 |
| Mean Performance Score | See Table 1. Performance Scores by Decile | ||||||||||||
| N of Entities | See Table 1. Performance Scores by Decile | ||||||||||||
| N of Persons / Encounters / Episodes | See Table 1. Performance Scores by Decile |
Validity
Two types of validity were completed for this measure, criterion validity and convergent validity.
Criterion Validity:
Denominator certification prior to data collection and extraction/ abstraction ensures that all medical groups apply the denominator criteria correctly and in a consistent manner. MNCM staff review the documentation to verify all criteria were applied correctly, prior to approval for data submission.
Denominator certification documentation for this measure includes:
- Date of Birth (ranges)
- Date of Service (ranges)
- ICD-10 Codes used
- Eligible specialties and provider types
- Exclusions to the measure and attest to mechanism for exclusions
- Attestations related to changes in medical record or billing systems
- Supplying all query code for review
PIPE data pre-submission certification includes a complete review of all medical group’s source code (sql query) to ensure limitations are applied appropriately. Additionally, MNCM staff confirms that elements that need to be mapped (i.e., MNCM proprietary codes such as RELC, insurance codes, etc.) are done correctly.
Common areas of correction in denominator for this measure included missing query code, incorrect date of birth ranges, incorrect dates for counting visits, missing ICD-10 codes or incomplete attestation. All were corrected prior to data submission.
Following data submission to the MNCM Data Portal, there are additional data quality checks in place for evaluating the accuracy of data submitted. During file upload, program checks for valid dates, codes and values and presents users with errors and warnings. Additionally, MNCM staff review population counts (denominator) and outcome rates for any significant variance from the previous year’s submission and may prompt further clarification from the medical group. In addition to reviewing automatic file format checks produced by the portal, MNCM staff also review the volume of data submitted in each file to ensure data is pulled into PIPE as expected.
Validation audits verify that the clinical data submitted for the numerator component of the measure matched the data in the patient record. Other data elements are also audited to verify the patient was included in the denominator correctly (e.g., diagnosis of ischemic vascular disease).
PIPE submissions include an additional validation audit that further confirms the completeness of the data by confirming that all expected data is pulled into PIPE.
The final validation step is the medical group review of statewide preliminary data. This is the official opportunity for participating medical groups to review and comment on the preliminary clinical quality performance measure results for all medical groups that meet the public reporting threshold (minimal of 30 patient denominators).
Convergent validity
Statistical Validity Conducted
Validity was tested for the computed composite score by testing the correlation of medical group performance with their performance on the Optimal Diabetes Care measure (NQF# 0729). Ischemic vascular disease and diabetes are chronic conditions that require ongoing management of multiple risk factors in order to reduce a patient’s overall risk of developing long term complications. It is expected that the quality of care provided by a medical group to patient with ischemic vascular disease would be of similar quality as the care provided to patients with diabetes, and the respective performance measure scores should demonstrate such.
Criterion Validity:
Data Quality Checks:
80% of those that submitted data passed initial quality checks.
The types of errors that were found and accounted for in this step were: insurance data, RELC data, incomplete denominator, incomplete statin/aspirin data, incorrectly excluded patients, incorrect data file formatting in PIPE, incorrect denominator counts (i.e., clinics not reporting and incorrectly reported total eligible population counts)
Upfront query checks:
82% of groups passed with no errors.
The types of errors that were found and accounted for in this step were: dates of service, dates of birth, CPT/ICD-10 codes, required exclusions not applied, did not apply updated value set for telehealth codes.
One new group did not proceed due to resource limitations/burden. Two groups realized they did not have a complete denominator due to updating their EMR and did not proceed with submission.
Convergent validity
See attached document for graph of linear regression that was completed, it is in the data dictionary OVC_spec file under the tab "OVCODC_linearregression". Based on linear regression analysis, a medical group’s performance on the Optimal Vascular Care measure is associated with its performance on the Optimal Diabetes Care measure, as demonstrated by an r2 value of 60%, representing a fairly strong correlation.
Criterion validity:
We are confident that our measure is pulling the correct data elements after our extensive audit process.
Convergent validity:
High compliance with critical data element validity as demonstrated by annual validation audit processes.
As demonstrated by the r2 value, 60% of the total variation in performance on the Optimal Vascular Care measure can be explained by variation in the Optimal Diabetes Care measure. This degree of correlation indicates that the Optimal Vascular Care composite measure score accurately reflects the quality of care provided.
Risk Adjustment
The following variables were selected as risk variables for the OVC measure:
Product: Commercial, Medicare, Medicaid, Uninsured/Self Pay, Unknown
Age
Social deprivation index
MNCM determines social deprivation index as a patient zip code level averages of poverty, public assistance, unemployment, single female with child(ren), and food stamps (SNAP) converted to a single index that is a proxy for overall.
Method to construct SDI
We constructed an area-level deprivation index using zip code tabulation area data from the American Community Survey conducted by the U.S. Census Bureau. The data is updated annually.
Following methods used in prior research for constructing area-level indicators of socioeconomic status for other purposes, we performed a principal components analysis that combines multiple variables from the ACS data into a single indicator of zip-code level socioeconomic status.
The resulting index is centered at zero, with a higher number indicating higher socioeconomic status in a particular zip code.
MNCM's version of risk adjustment is an Actual to Expected model, where we don't change the actual calculated measure result, we determine a new benchmark for each reported entity to be compared to. The OVC measure is reported against a risk adjusted benchmark. A risk adjusted benchmark is created for each medical group and clinic result. Then the above/average/below rating that would be included in public reporting is the comparison of the entity's actual result to the new benchmark.
To test whether there was a statistically significant difference between the expected and actual rates by each clinic/medical group; a chi-square test is used. This method was employed to test the proportion of optimally managed patients attributed to a clinic/medical group compared to a specific value for that clinic/medical group. The specified value is an expected rate calculated considering the overall state rate and adjusted for risk factors specific to the measure. The methodology uses a 95 percent test of significance (p-value <0.05).
The comparison between clinics/medical groups is the aforementioned test of significance and the actual to expected ratio: actual percentage of patients meeting criteria divided by the expected percentage of patients meeting criteria for the particular entities’ mix of patient risk.
See attached supplemental file for a summary of these results.
MNCM uses logistic regression to calculate the expected value. This model provides the prediction of binary variables by a mix of continuous and discrete risk variables. The advantage of the regression model is it allows for more variables to be included in the model and also allows for use of continuous independent variables.
MNCM’s OVC measure calculates the percentage of patients 18-75 years of age who had a diagnosis of ischemic vascular disease (IVD) and whose IVD was optimally managed. With optimally managed as the binary (yes = 1 or no =0) dependent variable. MNCM risk adjust this measure using the following independent variables:
Age (as a continuous variable)
Insurance product type (as categorical binary variables: commercial, Medicare, MN Government Programs/Medicaid, uninsured/self-pay, unknown)
Social deprivation index (as a continuous variable)
Data cleaning and statistical analyses are performed with computer software (Microsoft Excel and SAS Enterprise Guide 8.3).
Risk factor data sources:
Age – submitted by providers from electronic health record
Product type – health insurance information is submitted by providers using electronic health record data. MNCM then conducts what we call a “match/attach” process with the health plans who identifies their members and then attaches product information. Each patient is assigned to one insurance category
Social deprivation index - American Community Survey conducted by the U.S. Census Bureau
Patient level file
MNCM pulls a patient level file from our data portals. Insurance product type is attached to each patient. Each patient is assigned to one insurance category. Social deprivation index is also attached to each patient based on patient zip code of residence.
A logistic regression is then run in SAS with optimally managed as the dependent variable, age and social deprivation index as a quantitative variable and insurance product type as classification variables. Commercial is left out of the model as it is the insurance product that other insurance products are compared to in the model. We use a full model fitted and specify main effects for all independent variables: age, Medicare, MHCP, uninsured, unknown, and social deprivation index.
The patient level file is then used in conjunction with the parameter estimates of the Analysis of Maximum Likelihood to calculate the probability of each patient to achieve optimal care.
This is done by using the following formula:
log odds= β0+ β1X1 +… βnXn, where β0 is the parameter estimate for the intercept, β1 is the parameter estimate for the first parameter X1 and βn is the parameter estimate for the nth parameter Xn.
The parameters take on whichever value they have for each patient in the patients’ level file however, prior to applying the formula above on classification variables, each “0” value is converted to “1” and each “1” value is converted to “-1” per SAS convention.
The direct application of the formula to our data set provides log odds for a patient not optimally managed (the probability modeled is optimally managed=0). A few more steps are required to calculate the probability of each patient achieving optimal care. This is being done by the following formula:
Probability (optimal care=1) = 1- [EXP (log odds (optimal care=0))/ (EXP (log odds (optimal care=0) +1)]
The clinic level file
The calculated probabilities are the expected values specific to each patient, so we use the patient level file to calculate each clinic’s expected or risk adjusted rate. Since we have clinic IDs, we calculate the average expected optimal vascular care rate for a patient at each clinic.
The medical group level file
Similar to the calculation of a clinic’s expected value, a medical group’s expected value is obtained by taking the average expected optimal vascular care rate for clinics in the medical group.
Descriptive statistics were employed to examine the distribution of risk variables identified from the conceptual model across the measured entities. These variables were scrutinized based on several criteria:
- Clinical/Conceptual Relationship with the Outcome of Interest: A logical theory underpinned by research and experience informed the selection of risk factors. The conceptual model established a plausible association between each factor and the outcome of interest, optimal vascular care (OVC), without necessitating a direct causal relationship.
- Empirical Association with the Outcome of Interest: Statistical analyses, including logistic regression modeling, confirmed the conceptual relationships between the selected risk factors and OVC performance.
- Variation in Prevalence of the Factor Across the Measured Entities: The prevalence of each risk factor across providers being measured was assessed. Factors demonstrating significant variation in prevalence contributed to understanding potential biases in performance results. For example, Age showed variation across different age groups, with estimates ranging from -1.12 to -0.26 (T-values: -2.71 to -12.91).
- Not Confounded with Quality of Care: Risk factors were scrutinized to ensure they were present at the start of care and did not represent the quality of care provided. This criterion aimed to isolate differences in performance attributable to differences in care provision rather than other factors. For instance, Insurance Product demonstrated variations between different insurance types (e.g., Medicare, Medicaid) with odds ratios ranging from 0.47 to 0.94.
- Resistant to Manipulation or Gaming: Factors less susceptible to manipulation or gaming, such as diagnosis or assessment data, were prioritized. This ensured the validity of performance scores as accurate representations of quality of care.
- Accurate Data that Can Be Reliably and Feasibly Captured: Practical constraints, including data availability and resource limitations, were considered. Factors were included in risk models based on the feasibility and cost-effectiveness of data collection. For example, the deprivation index derived from zip code data was considered a reliable measure of socioeconomic status.
- Contribution of Unique Variation in the Outcome: Risk factors contributing unique variation in the outcome were identified to prevent overfitting and unstable estimates. Factors redundant or highly correlated with others were omitted to improve model performance and reduce data collection burden. The correlation analysis among potential risk adjusters, such as the deprivation index variables, ensured the selection of factors with unique contributions to the outcome.
The statistical analysis conducted to determine the risk factors for inclusion in the Optimal Vascular Care (OVC) measure's risk model involved logistic regression modeling using the SAS Procedure Glimmix. This analysis aimed to identify which variables significantly contributed to OVC performance while accounting for clinic-level differences. Initially, Age, Insurance Product, Gender, and Depression were tested as potential risk factors. Among these, Age and Insurance Product showed significant variation in results and were retained for further analysis. Gender did not exhibit sufficient variation between clinics to contribute meaningfully to clinic-level differences, and Depression was excluded due to the high cost associated with its data collection.
The analysis focused on identifying variables that could adequately adjust for patient risk factors without disproportionately burdening providers with data collection costs. Furthermore, MNCM's Risk Adjustment Advisory Task Force decided against including variables such as race, ethnicity, language, and country of origin in the risk adjustment model to ensure disparities could be identified separately.
In 2018, MNCM explored risk adjustment based on geography of patient residence. The rationale was that characteristics of residential and community context are related to performance indicators, precede care delivery and are not things that a provider can manipulate. To avoid increasing reporting burden on providers, MNCM explored using the 2015 American Community Survey data by Zip Code Tabulation Area (ZCTA). The principal component analysis combined multiple variables into a single indicator of zip-code level socioeconomic status to create a deprivation index. The index is centered at zero with a higher number indicating higher socioeconomic status.
MNCM examined published, peer reviewed articles on the creating of a measure for area SES. Key findings from the review were: census data at the zip code level is typically used, principal components analysis is used, and a deprivation index is generated. In line with published literature, MNCM chose the following variables to evaluate for the deprivation index: % with SNAP benefits, % in poverty, % unemployment, % on public assistance, % single female with child. A correlation analysis was conducted to determine correlation coefficients among the selected variables.
These variables were selected based on their significant contribution to explaining clinic-level differences in OVC performance, as determined through logistic regression modeling. The coefficients represent the magnitude of influence each variable has on adjusting for patient risk factors within the OVC measure.
- Age:
- Data Source: Patient demographic records
- Code with Descriptor: AGE - Patient Age
- Equation: Logistic regression model coefficient (-0.0306)
- Insurance Product:
- Data Source: Patient insurance records
- Code with Descriptor: IMDCR (0.122), mhcp (0.44), uninsured (0.5967), undt (0.1456)
- Equation: Logistic regression model
Deprivation Index (Stratification Variable):
- Data Source: 2015 American Community Survey by Zip Code Tabulation Area (ZCTA)
- Code with Descriptor: DEP_INDEX - Deprivation Index
- Equation: Logistic regression model coefficient (-0.1321)
While Age, Insurance Product, Gender and Depression all made the initial statistical requirement of significant variation in results, Gender did not show enough variation between clinics to contribute to the unique variation of clinic level results (Criteria #3) and depression was not selected due to relative high cost of collection (Criteria #7).
MNCM does not include variables such as race, ethnicity, language, and country of origin (RELC) in its risk adjustment model. MNCM’s Risk Adjustment Advisory Task Force decided that variables such as RELC should be reported in a segmented manner so that disparity gaps can be identified and closed. We should not risk adjust away variables that would make disparities hard to identify.
Therefore, the risk variables selected were Age and Insurance Product.
In 2018, MNCM explored risk adjustment based on geography of patient residence. The rationale was that characteristics of residential and community context are related to performance indicators, precede care delivery and are not things that a provider can manipulate. To avoid increasing reporting burden on providers, MNCM explored using the 2015 American Community Survey data by Zip Code Tabulation Area (ZCTA). The principal component analysis combined multiple variables into a single indicator of zip-code level socioeconomic status to create a deprivation index. The index is centered at zero with a higher number indicating higher socioeconomic status.
In the Attach Risk Adjustment Modeling and/or Stratification Specifications file there is a tab titled "Social Risk Factors" which has a correlation analysis to determine correlation coefficients among the selected variables. Which demonstrated, that the high correlation coefficients among the selected variables indicated that these variables were likely to converge together into a single deprivation index.
The statistical analysis conducted to select risk factors for inclusion in the OVC measure's risk model involved rigorous scrutiny based on the outlined criteria:
- Clinical/Conceptual Relationship with the Outcome of Interest: A conceptual model informed by research and experience guided the selection of risk factors, ensuring a logical association with the outcome of interest, optimal vascular care.
- Empirical Association with the Outcome of Interest: Logistic regression modeling confirmed the empirical associations between selected risk factors and OVC performance.
- Variation in Prevalence of the Factor Across the Measured Entities: Factors demonstrating significant variation in prevalence across providers were retained, contributing to understanding biases in performance results.
- Not Confounded with Quality of Care: Selected risk factors were not confounded with the quality of care provided, ensuring differences in performance were attributable to care provision rather than other factors.
- Resistant to Manipulation or Gaming: Factors less susceptible to manipulation or gaming were prioritized, enhancing the validity of performance scores.
- Accurate Data that Can Be Reliably and Feasibly Captured: Risk factors were selected based on the feasibility and cost-effectiveness of data collection, considering practical constraints.
- Contribution of Unique Variation in the Outcome: Factors contributing unique variation in the outcome were included, while those redundant or highly correlated with others were omitted to improve model performance and reduce data collection burden. The correlation analysis among potential risk adjusters ensured the selection of factors with unique contributions to the outcome.
Calibration Testing: To assess the calibration of the model, we compared the predicted probabilities of outcomes with the observed proportions of outcomes within various subgroups. We utilized methods such as calibration plots, Hosmer-Lemeshow goodness-of-fit test, and observed vs. expected event rates to evaluate calibration.
Results of Calibration Testing: The calibration testing results indicated that the model demonstrated good calibration overall, with predicted probabilities aligning closely with observed outcomes across different subgroups. Calibration plots showed that the predicted probabilities closely followed the 45-degree line, indicating good agreement between predicted and observed outcomes. Additionally, the Hosmer-Lemeshow goodness-of-fit test yielded a non-significant p-value (p > 0.05), suggesting no evidence of lack of fit between predicted and observed outcomes.
Discrimination Testing: For discrimination testing, we utilized the c-statistic (Area Under the Receiver Operating Characteristic Curve) and other measures such as Somers' D, Gamma, and Tau-a. These metrics evaluate the model's ability to distinguish between positive and negative outcomes.
Results of Discrimination Testing: The discrimination testing results indicated moderate discrimination ability of the model. The c-statistic was calculated to be 0.586, suggesting that the model has some ability to differentiate between individuals with positive and negative outcomes. Additionally, other discrimination metrics such as Somers' D (0.172), Gamma (0.172), and Tau-a (0.085) also supported the model's moderate discrimination ability.
Assessment of Over- or Under-Prediction: We examined the model's performance in predicting outcomes for important subgroups, considering factors such as age, insurance status, and dependency index. While there were slight variations in predicted versus observed outcomes within some subgroups, overall, there were no significant instances of over- or under-prediction that could not be explained by random variability.
Interpretation of Results: The results demonstrate that the model adequately controls for differences in patient characteristics (case mix). The good calibration indicates that the model's predictions align well with observed outcomes across various subgroups, indicating that the model effectively accounts for differences in patient characteristics. Additionally, the moderate discrimination ability of the model suggests that it can differentiate between patients with different risk profiles.
Rationale for Risk Factor Inclusion: Each risk factor included in the final model was selected based on its statistical significance, clinical relevance, and contribution to model fit. Factors that significantly improved model fit and had strong clinical evidence supporting their association with the outcome were included. Factors such as age, insurance status, and dependency index were included in the final model as they demonstrated significant associations with the outcome and contributed to the model's overall fit.
Explanation of Results: The results indicate that the model effectively controls for case mix differences by accurately predicting outcomes across various patient subgroups. The inclusion of statistically significant and clinically relevant risk factors ensures that the model captures important predictors of the outcome, leading to robust predictions. Overall, the results meet expectations for a well-fitted and clinically relevant predictive model, providing valuable insights into patient outcomes.
Use & Usability
Use
Medical Groups
MNCM reports provide comparative data and analysis that benefit all the stakeholders we serve. Empowered with this information, consumers can see how care varies across providers, providers have visibility to how their performance compares to others and where their biggest improvement opportunities are, and health care payers and other purchasers better understand and improve value for money that is spent on health care.
The program´s goal is to drive improvements in healthcare quality within care delivery systems and maximize participation of all providers over time. The Partners in Quality program consists of: Partners in Excellence (PIE), Innovations in Health Care Award, and Preventive Care Recognition Award. Financial rewards are based on medical, specialty or pharmacy group performance as measured by Minnesota Community Measurement. For those measures that do not have a corresponding MNCM measures, we utilize HealthPartners Clinical Indicator measurement set, and HealthPartners Consumer Choice Satisfaction survey.
Usability
For entities to improve performance on this measure they must appropriately manage an individuals IVD as directed by guidelines. These actions are achievable but can be difficult with how overwhelmed medical professionals may be. Rates can also be improved if all documentation is complete for an individual, an example is if a patient has a statin intolerance which should remove them from the denominator but if that is only written in a note and not placed in the allergy or problem list then that patient will be brought into the denominator bringing down the overall rates. MNCM is consistently working with all our site partners to make sure that documentation is complete for all patients.
MNCM’s multi-stakeholder Measurement and Reporting Committee (MARC) is tasked with the annual review of all publicly reported measures and recommending the slate of publicly reported measures to the MNCM Board of Directors. As part of this process, a comprehensive review of each measure. which includes evaluation against PQM criteria (importance, scientific acceptability, feasibility and use), is conducted every three years by MARC members to determine any potential measure specific action needed (continue, convene a measure development workgroup for redesign, or retire). This standardized measure review process helps inform measures included on the slate for public reporting.
MNCM provides a year-round staffed support through a helpline 612-746-4522 or email support@mncm.org.
Measure development workgroups are a multi-stakeholder in nature and comprised of the following types of representation: providers (physicians or other allied healthcare providers) who care for the population of interest, patients, and representation from clinic administration, data analysis, quality improvement, health plan. This workgroup composition and its consensus-based decision-making process is used for both de novo measures and for measures undergoing a redesign process.
During development of a new measure, formal public comment is sought from the clinics and medical groups who will be measured. All comments are reviewed by the measure development workgroup for identifying any redesign or tweaks to the measure specifications prior to pilot testing the measure. Pilot testing provides an additional source for feedback from users; pilot participants are surveyed with questions around feasibility and data element ease or difficulty.
Approval for convening a measure development group for redesign is granted by the Measurement and Reporting Committee (MARC) with oversight by the MNCM Board of Directors for resource allocation. The multi-stakeholder workgroup ensures balanced representation. Recommendations for redesign are reported back to the MARC for review and approval. Medical groups and clinics receive communication via MNCM’s newsletter the Measurement Minute. The public, including medical group and clinic representatives can communicate any questions or concerns about measure changes or any measure by contacting support@mncm.org.
The standardized measure review process conducted by MARC is an additional avenue to provide feedback about the measure.
MNCM conducts a periodic medical group survey in which all clinics in the state are invited to participate and provide feedback. There are structured questions asking the users about measure value and burden..
Since the start of public reporting of this measure in 2007, there has been steady improvement in composite rates for achieving all targets; statewide average from 38.9% to 61.1% with continued demonstration of variability and opportunity for improvement.
Comparing 2022 results to 2021, the optimal rates in 2022 were not statistically different than the respective 2021 rates in all subpopulations, except for the following populations:
- White
- Not Hispanic/Latinx
- English speakers
- U.S.-born
The rates for these populations were significantly lower than their rates in 2021.
When we compare 2022 results to 2019 (pre-pandemic) the optimal rates in 2022 were significantly lower than the respective 2019 rates for all subpopulations, except for the following populations:
- Asian
- Black
- Multi Racial
- Native Hawaiian/Pacific Islander
- Non-English speakers
The 2022 rates for these populations were not statistically different than their 2019 rates.
This measure has been implemented since 2008; No unintended consequences identified during the testing, implementation and ongoing review of this measure.
Comments
Staff Preliminary Assessment
CBE #0076 Staff Assessment
Importance
Strengths:
- Note that this measure is not a true composite measure; it includes four components in the numerator that must all be met: BP control; tobacco cessation; use of statins; use of aspirin or antiplatelet medication.
- The logic model provided is largely appropriate. Guidelines and systematic reviews cited uphold the blood pressure target of 140/90 and use of statins, and support clinician-level interventions for tobacco cessation treatment.
- Table 1 does not report performance scores (instead it reports 0s and1s without interpreting the meaning of those values), potentially limiting reviewers’ ability to evaluate for gap. However, a re-analysis of the raw data provided in the reliability analysis attachment that calculates performance scores by decile demonstrates a ranges of .277 at the 10% decile to .616 at the 100% decile.
Limitations:
- The guidelines cited in the submission do not address routine use of aspirin for atherosclerotic cardiovascular disease (ASCVD).
- Evidence for importance to patients is not clearly stated; developer refers to their Measurement and Reporting Committee (MARC), which includes patients as well as other stakeholders, which opted in 2023 to continue reporting the measure, but details are not provided.
Rationale:
- There is supporting evidence cited for hypertension management, statin use, and tobacco cessation, but not for routine aspirin or antiplatelet medication.
- There is a clear performance gap between high and low-performing groups.
- Importance of the measure to patients has not been demonstrated.
Questions for committee:
- Is there more recent evidence for the relationship between any of the four clinical criteria in the numerator and material outcomes that the developer should consider?
- Is there sufficient evidence for use of aspirin or antiplatelet medication in this measure?
- Is there sufficient evidence for importance of this measure to patients?
Feasibility Acceptance
Strengths:
- Developers conducted a feasibility assessment. While the routine generation of data through clinical care was not addressed directly, the developer cites a survey of 101 clinicians, of which 65% rated it easy or very easy to collect the needed data. In addition, the developer appears to offer substantial support to clinicians during implementation of the measure.
Limitations:
- About one third of providers surveyed did not find it easy to collect the needed data.
Rationale:
- Developers conducted a feasibility assessment. While the routine generation of data through clinical care was not addressed directly, the developer cites a survey of 101 clinicians, of which 65% rated it easy or very easy to collect the needed data. In addition, the developer appears to offer substantial support to clinicians during implementation of the measure. However, up to one third of providers may have found data collection burdensome.
Scientific Acceptability
Strengths:
- The measure is clear and well defined. Entity-level reliability is conducted on the observed rates with 2022 data across 78 entitles using the beta-binomial method. Entity-level reliability of the observed rate is >0.6 for 95% of the entities.
Limitations:
- This measure is reported against a risk-adjusted benchmark (observed/expected). Risk adjustment accounts for differences between entities, so the reliability of the measure as it will be used is likely less than the reliability of the observed rates. A method such as split-half reliability testing should be used to estimate the reliability of the observed/expected values.
Rationale:
- The measure is well defined. Reliability was assessed for the observed rates only. Reliability should be tested on the measure as it is intended to be reported. In Section 4.4.3 the developer states, "... the above/average/below rating that would be included in public reporting is the comparison of the entity's actual result to the new benchmark". If the measure is intended to be reported as observed rates only this should be explicitly stated.
Strengths:
- The developer validated reported data elements (“criterion validity”), by review of source code, certification of denominator elements, checking data following upload to MNMC, and validation audits for numerator and denominator. 80% of groups passed “initial quality checks”; 82% of groups passed query checks with no errors. Developers also refer to a medical group review of statewide preliminary performance data as part of the criterion validity evaluation.
- Convergent validity analysis consisted of a linear regression correlating the OVC measure with a Optimal Diabetes Care measure (CBE #0729), with the rationale that both address chronic disease that require ongoing management of multiple risk factors. R-square indicated that 60% of variance in OVC is explained by variance in ODC.
- The conceptual model portion of the risk adjustment (RA) section lists criteria for considering risk factors for adjustment, and it also provides a graphic showing the five elements of the social deprivation index (SDI). Overall, the selection of risk factors appears plausible.
Limitations:
- For the criterion validity analysis, it is unclear how many of the groups that did not pass initial checks addressed the problems and submitted data or deferred data submission (reference to three groups that did not proceed).
- While the conceptual model for RA is plausible, the developer does not provide a rationale or cite evidence for the inclusion of age, insurance, or SDI in the RA model.
Rationale:
- For entity-level validity (convergent validity), the developers found that 60% of the variance in OVC was explained by the variance in another measure (CBE #0729) which is presumed to be similar in terms of clinical processes for the management of multiple risk factors for a chronic disease.
- The developer performed a detailed validity assessment for data elements ("criterion validity") and the submission references 80% and 82% passing different checks. However, it is unclear how many of the 78 groups did not pass this testing and how this was resolved.
- The conceptual model presented for RA is plausible, but not supported by an explicit rationale.
Equity
Strengths:
- N/A
Limitations:
- N/A
Rationale:
- This optional criterion was not addressed by the developer.
Use and Usability
Strengths:
- Currently used in the MNHealthScores program and HealthPartners pay for performance program, Partners in Quality, covering the state of Minnesota. Performance scores are reported in the MNCM Community Reports.
- Entities can improve performance by managing eligible patients in compliance with guidelines, and also by improving documentation, for which the developer provides technical assistance.
- Feedback is collected through periodic survey of medical groups. In addition, the MARC performs a comprehensive review of this measure every 3 years for needed actions (e.g., continue, redesign, retire). In the importance section, the developer reports that the MARC recommended continuing the measure without redesign based on their 2023 review.
- No unintended consequences were identified. No unexpected findings are reported.
Limitations:
- Engagement with quality improvement may be a challenge for some providers ("overwhelmed medical professionals”).
- While the developer cites improvement in the statewide score between 2007 and 2018, the mean score fell to 55.3% in 2022 from the high of 61.1% in 2018. No rationale is provided for the falling rate. Developer lists some statistically significant differences between 2022 and 2021 or 2019 by race, language spoken, and US born, but does not provide details or significance testing.
- If there are plans in the future to expand use to a CMS program, the measure may need more testing in a broader population than MN.
Rationale:
- This measure is currently in use in a statewide performance program, and improvement in the measure score can be achieved through compliance with clinical guidelines and improved documentation of patient information. Feedback is collected using a wide range of methods, including help desk, provider surveys, and comprehensive review by the working group. No unintended consequences have been reported.
- However, improvement in the measure score has been seen since 2018 and the developer does not provide a clear rationale for this. In addition, some providers may find it challenging to engage in quality improvement (QI) activities; the developer could consider ways entities might overcome barriers to QI for providers.
Committee Independent Review
CVD Summary
Importance
Cardiovascular disease affects over 121 million adults in the US and accounts for over 32% of deaths so there are many that can benefit from improved performance in care. It is my understanding that aspirin is no longer indicated for routine prevention according to the ACC and American Heart Assc. This should be addressed prior to endorsement.
Feasibility Acceptance
While the threshold for feasibility is met, I would encourage additional data collection.
Scientific Acceptability
Agree with staff assessment
Agree with staff assessment
Equity
Not addressed by measure developer
Use and Usability
Agree with staff assessment
Summary
While CVD is a very important healthcare priority, the measure developer needs to address additional areas prior to endorsement.
Thought provoking, unclear/unreliable endpoints
Importance
There are a number of items that are unclear or missing in the measure rationale and summary. The time period in which the 4 criteria is evaluated is just any patient during the year who visited an office, diagnosed with IVD, and reports whether or not they met the specified goals?
Does a singular most recent BP reading really best summarize the management of a patients hypertension?
Aspirin recommendation for all patients? Should be patient specific with certain risk factors to the best of my understanding.
It is unclear what impact this metric would have on patient care.
Feasibility Acceptance
While the test of feasibility the authors use suggest that clinicians find the measure reasonably easy to enter data, one would assume that the extent of resources offered would vary in real world practice.
For example, Significant chart abstraction may be needed to appropriately assess whether a patient be on aspirin or not.
Scientific Acceptability
Agree with staff.
Agree with staff.
Equity
N/A
Use and Usability
QI would be the target, in patient who are tobacco users or grossly non-compliant with their medications the ability to improve may be time intensive and burdensome to clinicians.
This metric may also over estimate positive results in patients who are non-compliant with meds, may have intermittently controlled hypertension, and/or deny using tobacco but use it regularly. This would make usability obsolete if the underlying data for the measure is unreliable.
Summary
While i think this measure is potentially impactful, I find it hard to believe that a self-reported use of tobacco or compliance with home meds will be reliable in real world application.
A single BP at goal is not reflective of long term care and definitely is not predictive of outcomes.
Measure Summary
Importance
Agree with staff assessment. Understanding patient needs and their perspective would be beneficial before supporting this measure.
Feasibility Acceptance
Agree with staff assessment.
Scientific Acceptability
Agree with staff assessment.
Agree with staff assessment.
Equity
Not addressed.
Use and Usability
Agree with staff assessment.
Summary
It would be helpful to hear a patient experience story before supporting this measure.
Public Comments
Public Comment Shared During May 29 Listening Session
Hi! My name is Janice Tufte. I'm a patient partner engaged in research and measurement. And the optimal vascular care that we're discussing right now, it's very specific. But reading through it, I can see where this is important, because it's an area, I imagine, where people with IVD, or whatever, it's more often seen with people, with other conditions. So, this is people that appear to be fairly healthy, have good parameters. And so maybe we'll be able to find out from this if there's certain population groups that are more prone to IVD. And so, I think it's good, but I think it's very specific, and hope that the uptick will be a positive one.