Author + information
- Published online February 9, 2016.
- Joseph P. Drozda Jr., MD, FACC, Chair, Writing Committee,
- T. Bruce Ferguson Jr., MD, FACC, FAHA, Writing Committee Member,
- Hani Jneid, MD, FACC, FAHA, FSCAI, Writing Committee Member,
- Harlan M. Krumholz, MD, SM, FACC, Writing Committee Member,
- Brahmajee K. Nallamothu, MD, FACC, Writing Committee Member,
- Jeffrey W. Olin, DO, FACC, FAHA, MSVM, Writing Committee Member and
- Henry H. Ting, MD, MBA, FACC, FAHA, Writing Committee Member
- ACC/AHA Performance Measures
- coronary artery disease
- health policy and outcome research
- myocardial infarction
- percutaneous coronary intervention
- peripheral arterial disease
- shared decision making
ACC/AHA Task Force on Performance Measures
Paul A. Heidenreich, MD, MS, FACC, FAHA, Chair
Nancy M. Albert, PhD, CCNS, CCRN, CCA, FAHA
Paul S. Chan, MD, MSc, FACC
Lesley H. Curtis, PhD
T. Bruce Ferguson, Jr., MD, FACC, FAHA
Gregg C. Fonarow, MD, FACC, FAHA
P. Michael Ho, MD, PhD, FACC, FAHA
Sean O’Brien, PhD
Andrea M. Russo, MD, FACC
Randal J. Thomas, MD, FACC, FAHA
Henry H. Ting, MD, MBA, FACC, FAHA
Paul D. Varosy, MD, FACC
Table of Contents
1. Introduction 560
1.1. Rationale for Update 560
1.2. Structure and Membership of the Writing Committee 560
1.3. Disclosure of Relationships With Industry and Other Entities 560
2. Methodology 561
2.1. Target Population and Care Period 561
2.2. Literature Review 561
2.3. Definition and Selection of Measures 561
3. 2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures 561
3.1. Gaps in Care 561
3.2. Broader Denominator (ASCVD)—Unique to This PM Set 562
4. General Discussion 564
4.1. Patient-Centered PMs and SDM 564
4.2. “Prescribed” Versus “Offered” 565
4.3. Prescription Versus Adherence 565
4.4. Exceptions and Exclusions 567
4.5. Method of Reporting 568
4.6. Limitations and Unintended Consequences 569
5. Future Directions 569
5.1. Improved Information Systems for Capturing Clinical Data 569
5.2. Measures of SDM and Shared Accountability 569
5.3. Conclusion and Summary 570
2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures: Performance Measure Set 572
Author Relationships With Industry and Other Entities (Relevant) 582
Peer Reviewer Relationships With Industry and Other Entities 583
2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures: Summary Analysis Table 587
American College of Cardiology (ACC)/American Heart Association (AHA) performance measures can serve as vehicles to accelerate appropriate translation of scientific evidence into clinical practice. Performance measures cover a subset of the most important recommended care practices and are considered appropriate for public reporting or use in pay for performance programs. Other measures of care that are not considered appropriate for public reporting or payment modification may be used as quality or test metrics for internal quality improvement. As defined by the ACC/AHA, quality metrics are those measures that have been developed to support self-assessment and quality improvement at the provider, hospital, and/or healthcare system level. These metrics may not meet all specifications of formal performance measures (1). In certain cases, an ACC/AHA performance measure writing committee may identify particular measures as quality metrics for the purposes of pilot testing with the potential of later promotion to performance measurement. Specific criteria for performance measures have been published (2,3) by the ACC/AHA and include an important gap in care and a clear path to improve care. Recently, value was added as an exclusion criterion (4), where a care practice deemed to be of poor value by an ACC/AHA guideline would not be considered as a performance measure. The ACC/AHA Task Force on Performance Measures has historically focused on process of care measures under the control of individual providers. However, writing committees may also create structural or outcome measures when they meet the ACC/AHA performance measurement criteria.
A goal of the ACC/AHA Task Force on Performance Measures is to rapidly create or update a performance measure when there are changes to a relevant ACC/AHA clinical guideline. Whenever possible, the ACC/AHA attempt to create relevant performance measures immediately following the publication of a guideline. However, the ACC/AHA believe that it is important to balance speed in measure development with a thorough review by stakeholders, content experts, and other interested parties using a public comment period. The goal is to have up-to-date and valid measures that can be used by all interested members of the healthcare system to evaluate and improve the quality of cardiovascular care.
Paul A. Heidenreich, MD, MS, FACC, FAHA
Chair, ACC/AHA Task Force on Performance Measures
The “2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures” Writing Committee (the writing committee) was charged with updating the current lipid performance measures (PMs) based on the new recommendations in the “2013 ACC/AHA Guideline on the Treatment of Blood Cholesterol to Reduce Atherosclerotic Cardiovascular Risk in Adults” (the Cholesterol Guideline) (5). In this measure set, the writing committee presents 5 PMs (Appendix A) 3 of which are intended for ambulatory settings and 2 for hospital (inpatient) settings. Four are revisions of lipid management measures appearing in 4 existing measure sets: “ACCF/AHA/ACR/SCAI/SIR/SVM/SVN/SVS 2010 Performance Measures for Adults With Peripheral Artery Disease” (6); “ACC/AHA 2008 Performance Measures for Adults With ST-Elevation and Non-ST-Elevation Myocardial Infarction” (7); “ACC/AHA/SCAI/AMA-PCPI/NCQA 2013 Performance Measures for Adults Undergoing Percutaneous Coronary Intervention” (8); and “ACCF/AHA/AMA-PCPI 2011 Performance Measures for Adults With Coronary Artery Disease and Hypertension” (9). These measure sets for percutaneous coronary intervention (PCI), coronary artery disease (CAD), peripheral artery disease (PAD), and ST-elevation myocardial infarction (STEMI)/non−ST-elevation myocardial infarction (NSTEMI) are summarized in Table 1. The fifth measure is new and applies to the population of patients with clinical atherosclerotic cardiovascular disease (ASCVD) as defined in the 2013 guideline (5).
1.1 Rationale for the Update
To ensure that ACC/AHA PMs for cardiovascular disease fulfill their intended purposes, remain relevant, and are fully aligned with current clinical practice guidelines, the ACC/AHA Task Force on Performance Measures (the Task Force) requires a transparent and consistent process that will allow focused updates to individual PMs when needed. This may occur when new guideline recommendations are released, when the Task Force receives feedback from end users of the measures about critical implementation problems, or when unintended adverse consequences associated with implementation of the measure(s) are detected. The current writing effort used the Cholesterol Guidelines’ recommendations (5), which are significantly different from those of the prior Adult Treatment Panel III guidelines and emphasize administration of high-intensity statin therapy instead of achievement of low-density lipoprotein cholesterol (LDL-C) targets.
1.2 Structure and Membership of the Writing Committee
The members of the writing committee included clinicians specializing in interventional cardiology and general cardiology, as well as persons with expertise in development of guidelines and development, implementation, and testing of PMs. Chairs for each of the previously published PMs (Table 1) were selected for the current writing effort.
1.3 Disclosure of Relationships With Industry and Other Entities
The Task Force makes every effort to avoid actual, potential, or perceived conflicts of interest that could arise as a result of relationships with industry or other entities (RWI). Detailed information on the ACC/AHA policy on RWI can be found online. All members of the writing committee, as well as those selected to serve as peer reviewers of this document, were required to disclose all current relationships and those existing within the 12 months before initiation of this writing effort. ACC/AHA policy also requires that the writing committee co-chairs and at least 50% of the writing committee have no relevant RWI.
Any writing committee member who develops new RWI during his or her tenure on the writing committee is required to notify staff in writing. These statements are reviewed periodically by the Task Force and members of the writing committee. Author and peer reviewer RWI relevant to the document are included in the appendixes (see Appendix B for relevant writing committee RWI and Appendix C for relevant peer reviewer RWI). Additionally, to ensure complete transparency, the writing committee members’ comprehensive disclosure information, including RWI not relevant to the present document, is available as an online supplement. Disclosure information for the Task Force is also available online.
The work of the writing committee was supported exclusively by the ACC and AHA without commercial support. Members of the writing committee volunteered their time for this effort. Meetings of the writing committee were confidential and attended only by committee members and staff from the ACC and AHA.
The development of PM systems involves identification of a set of measures targeting a specific patient population observed over a particular period. To achieve this goal, the Task Force has outlined a set of mandatory sequential steps (2). The following sections outline how these steps were applied by the present writing committee.
2.1 Target Population and Care Period
The target population for the ASCVD PM reflects the ACC/AHA Cholesterol Guidelines (5) population and consists of patients ages 18 to 75 years. In the focused update of the 4 existing lipid PMs, the target population consists as well of patients ages 18 to 25 years, representing a change from the age range previously specified in each measure set. This change was felt to be necessary in order to maintain consistency with the Class of Recommendation I, Level of Evidence A recommendation in the guidelines for treatment of patients with clinical ASCVD. Additionally, the writing committee developed exclusion criteria for the measures where appropriate in order to further specify the target population.
2.2 Literature Review
The writing committee used the Cholesterol Guidelines (5) as a primary source for deriving the measures. The writing committee also carried out a literature review to assess contemporary gaps in care.
2.3 Definition and Selection of Measures
The writing committee focused on developing these measures against the ACC/AHA attributes of PMs. Each measure was constructed in a way to ensure it was evidence based, desirable in regard to measure selection, feasible to implement, and consistent with accountability (Table 2). After the peer review and public comment period, the writing committee reviewed and discussed the comments and made further refinements in the measure set. The writing committee evaluated the potential measures against the ACC/AHA attributes of PMs (Table 2) to reach consensus on which measures should be advanced for inclusion in the final measure set; the Summary Analysis Table (Appendix D) captures this evaluation process. The majority of the writing committee believed that the 5 measures in the set fulfilled the PM attributes.
3 2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures
3.1 Gaps in Care
Each of the original writing committees that developed the measures revised in this update identified secondary prevention performance gaps in the patient populations that were the subjects of their measure sets. There is evidence that these gaps are ongoing, although the published studies deal primarily with prescription of medication.
A study from the REACH (Reduction of Atherothrombosis for Continued Health) Registry found that only 83% of ambulatory patients with known ASCVD were receiving lipid-lowering agents (11). A prospective study by Rabus and colleagues of 73 patients with angiographically diagnosed CAD found that only 44% received prescriptions for statins (12). Reports from the National Cardiovascular Data Registry PINNACLE Registry of ambulatory patients with CAD revealed that only 66.5% (103,830 of 156,145) were receiving optimal medical therapy (OMT), including statins (13), that 77.8% (30,160 of 38,775) were prescribed statins (14), and that uninsured patients were 6% less likely to receive lipid-lowering therapy (15). Additionally, the study by Maddox and colleagues found substantial variation in prescription patterns by practice site (13,15). Shah and colleagues reported that, among 292 patients from Olmstead County, MN, with incident acute myocardial infarction (MI), only 44% were still taking statins 3 years after their infarction (16). Interestingly, a study by Borden and colleagues (17) involving patients in the National Cardiovascular Data Registry CathPCI Registry failed to show any significant improvement in the prescription of OMT after PCI following publication of the results of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) study, which had demonstrated no incremental advantage of PCI over OMT on outcomes other than angina-related quality of life in stable CAD. Among all 467,211 patients (173,416 before [37.1%] and 293,795 after [62.9%] the COURAGE trial) who met the study criteria, the use of OMT at discharge following PCI before and after the COURAGE trial was 63.5% (95% confidence interval, 63.3% to 63.7%) and 66.0% (95% confidence interval, 65.8% to 66.1%), respectively (p < 0.001).
The extent to which statin treatment is initiated as a result of a shared decision-making (SDM) process between patient and clinician has not been systematically assessed but is likely small. Additionally, there is minimal information available about the statin doses being used in practice, although there are reasons for concern that many patients are being undertreated. It is common practice among clinicians to use the smallest dose of a medication necessary to achieve a therapeutic target and minimize the chance of adverse effects. Even PMs such as those revised in this focused update exclude the requirement for therapy in patients who have achieved goals. A study of 38,775 patients in the PINNACLE Registry by Arnold and colleagues revealed findings consistent with undertreatment of patients with CAD (14). They found that 6,573 (17.0%) patients were not receiving any lipid-lowering therapy. Cholesterol levels were available for 3,365 of these patients, 1,794 (53.3%) of whom had LDL-C levels <100 mg/dL, consistent with clinicians either failing to treat or discontinuing lipid-lowering therapy when patient LDL-C levels met the previously recommended therapeutic target. In a secondary analysis of the Translational Research Investigating Underlying Disparities in Acute Myocardial Infarction Patients’ Health Status (TRIUMPH) study, only 23% of 4,271 patients discharged alive following an acute MI were on maximal statin therapy, with substantial variability across hospitals (18). Finally, a study in the Get With The Guidelines Registry of 65,396 patients with acute coronary syndromes (ACS) who were discharged with lipid-lowering agents found that only 38.3% were discharged with intensive lipid-lowering therapy (19). An editorial challenged measure developers to track the use of effective drug therapy, including dose (20). The current measures are designed therefore not only to promote the use of statins as recommended by the Cholesterol Guidelines (5) but also to emphasize the importance of high-intensity dosing.
3.2 Broader Denominator (ASCVD)—Unique to This PM Set
The target population for the ASCVD PM includes women and men between 18 and 75 years of age who have clinical ASCVD, which includes the following: ACS, history of MI, stable or unstable angina, coronary (including PCI) or other arterial revascularization, stroke, transient ischemic attack, or PAD. Although this patient population seems heterogeneous, it encompasses a variety of patients who all share presumed atherosclerosis (21) as a common pathophysiology. Atherosclerosis is a chronic diffuse disease involving a myriad of arterial beds with intermittent acute clinical manifestations, predominantly occurring as a result of superimposed thrombosis, plaque progression, spasm, embolism, or a combination of the above. Other pathophysiological processes can contribute to the creation of stenosis or aneurysms in the arterial circulation; however, atherosclerosis remains the most common pathophysiology.
Patients with clinical ASCVD represent 1 of 4 major groups identified by the writing committee of the Cholesterol Guidelines (5). For these patients, treatment with a 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitor, commonly known as a statin, is clearly beneficial. According to the Cholesterol Guidelines (5), patients with clinical ASCVD were identified by using the inclusion criteria from randomized clinical trials (RCTs) in secondary prevention. In addition, the potential for reduction of risk for ASCVD with statins in these patients clearly exceeds the potential for adverse effects (22).
The Cholesterol Treatment Trialists provided a comprehensive assessment of the benefits observed with statins (23). They undertook meta-analyses of individual participant data from 26 RCTs and demonstrated reduction in all-cause mortality, which was largely attributable to significant reductions in deaths due to CAD and other cardiac causes (23). The majority of studies in the aforementioned report included patients with known ASCVD. Of the 26 RCTs included, 5 trials (39,612 subjects, all of whom had CAD) compared more versus less intensive statin regimens. The trials demonstrated that more intensive regimens produced a highly significant 15% further reduction in major vascular events, driven by reductions in coronary death or nonfatal MI, coronary revascularization, and ischemic stroke (23). The investigators also found no significant effects observed on deaths due to cancer or other nonvascular causes or on cancer incidence, even at low LDL-C concentrations (23).
The aforementioned report was a meta-analysis of RCTs (23). Concerns about the quality and quantity of safety reporting in RCTs have been raised previously, and many researchers find the reporting of risks in RCTs to be largely inadequate (24–26). Data from RCTs should generally be supplemented by evidence from effectiveness studies to inform best clinical practice (27).
On extensive examination of clinical studies from the literature, the current clinical evidence does not support the notion that titrating lipid therapy to achieve proposed low LDL-C levels is beneficial or safe. Conversely, compelling evidence supports near-universal empirical statin therapy for patients at high cardiovascular risk regardless of their LDL-C levels (28). Thus, many argued to abandon the paradigm of treating patients to LDL-C targets and instead replace it with a more tailored treatment approach (i.e., personalized care), which aims not only to improve patient outcomes but also reduce harms and costs caused by overtreating patients at low risk (29,30).
For patients with ACS, which includes unstable angina, NSTEMI, and STEMI, the general period of assessment is the inpatient hospitalization or related emergency department visit. For other patients (non-ACS patients), the PM is intended to assess the care for patients at the practitioner level in an ambulatory care setting for the primary purpose of quality improvement. For these non-ACS patients, the outpatient care period is defined as the care provided in an outpatient setting within the time under evaluation, which is usually 12 months.
There are important potential exceptions for routine initiation of statin treatment. For primary prevention, the Cholesterol Guidelines expert panel determined that, despite the high level of risk for cardiovascular disease in patients with a higher New York Heart Association class of heart failure or receiving hemodialysis, the available evidence suggests that initiation of statin therapy might not achieve a significant risk reduction (31–33). In recognizing this, the expert panel made no recommendations about the initiation or discontinuation of statins in these populations, allowing for physician judgment in individual patients (5). Additional exceptions and exclusions related to secondary prevention measures are discussed in a separate section of this report.
Historically, the Task Force has developed separate sets of PMs in discrete patient populations, including patients with STEMI and NSTEMI (7), PAD (6), CAD (9), and those undergoing PCI (8). These separate seminal documents, each inclusive of a PM pertaining to statin therapy in its corresponding population, may generally be more useful in specialty care quality improvement programs. Although the writing committee is adhering to this philosophy in revising the lipid PMs for each of these 4 specific populations, it is taking a novel approach in creating a new PM that applies to the much broader population of patients with ASCVD. This PM is concordant with the Cholesterol Guidelines (5), which was based on evidence from RCTs and their meta-analyses showing risk reduction among the variety of patients with clinical ASCVD, including those with ischemic cerebrovascular events. The writing committee believes that primary care clinicians and specialists concerned with secondary prevention of ASCVD will find these new PMs easy to use in the clinical setting. The 5 updated measure sets are summarized in Table 3.
4 General Discussion
4.1 Patient-Centered PMs and SDM
The recommendation to initiate statins for secondary prevention is based on strong evidence in which benefit far exceeds risk (34). However, better patient outcomes are realized only if patients agree with, act on, and adhere to the recommendation for 5 to 10 years. The importance of clinician-patient discussions about statin therapy is specifically emphasized in the Cholesterol Guidelines (5). Among patients who are prescribed statins for secondary prevention, most initiate treatment, but up to half discontinue statins at 1 to 2 years’ follow-up (16,35). Therefore, a PM that represents only the number of patients prescribed statins in the numerator divided by the number of patients eligible to receive statins for secondary prevention in the denominator is inadequate and does not reflect quality of care. Rather, a measure that reflects the proportion of patients who participated in SDM would promote patient participation in the treatment plan, potentially increasing adherence to guideline-recommended care and improving patient-centered outcomes.
SDM: What Is It?
The SDM approach aims to promote a process whereby patients and clinicians together make a choice about treatments that incorporates 2 perspectives: 1) clinicians recommending treatments based on strong evidence in which benefit exceeds risk, and 2) patients deliberating on how treatments fit with their preferences, values, and personal context (36). In this framing, the clinician is the expert on evidence-based medicine and guidelines, and the patient is the expert on his or her preferences, values, and personal context. SDM mitigates the power differential between these 2 experts and acknowledges that both perspectives contribute equal weight to decision making. By incorporating patient preferences, values, and personal context to decision making, clinicians strengthen their implementation of evidence-based medicine and guidelines in a patient-centered manner to improve outcomes that matter to patients. A PM that integrates patient values, preferences, and personal context with evidence-based medicine and guidelines is novel and changes the focus from recommending and prescribing statins based on strong evidence to promoting choice by an informed patient whether or not to initiate statins.
SDM: Why Do It?
SDM has often been framed as an approach to curb overuse of expensive or unnecessary treatments. However, clinicians and healthcare organizations should bear the responsibility for curbing overuse, that is, recommending treatments where benefit does not exceed risk. The rationale for SDM includes patient safety, patient engagement, patient experience, and ethical principles. The patient safety reasoning arises from the notion that a misdiagnosis of a patient’s medical condition leads to unnecessary and unwanted tests and treatments associated with harm and cost. Similarly, misdiagnosis of a patient’s preferences, values, and personal context can lead to unnecessary and unwanted treatments. The patient engagement principle poses the question “What would the patient choose if the patient knew what the clinician knows?” When patients are offered the opportunity to participate in SDM, the majority prefer this approach, and patient satisfaction and experience with care improve (37). The ethical justification for SDM is based on the principle of autonomy that patients should be empowered to make informed decisions about their health.
SDM: How to Do It?
The path forward includes advancing how clinicians engage patients in decision making, developing tools to promote and facilitate SDM, and measuring that SDM occurred (38,39). Clinicians need to embrace the concept that evidence-based medicine and guidelines alone are not sufficient to make a recommendation or decision; rather, the evidence has to be considered from the viewpoint of what matters to individual patients. Hence, the clinical encounter transforms from one where the clinician strives to convince the patient of the “right answer” to one where the clinician and patient collaborate, deliberate, and arrive at the “best answer” that fits patient preferences, values, and context. Decision aids are one type of tool used during clinical encounters and have been shown to increase knowledge transfer to patients about personalized benefit and risk, improve patient involvement in decision making, and reduce decisional conflict (40). Decision aids can be implemented at the point of care to promote SDM, including the choice of whether or not to initiate a statin for the next 5 to 10 years to lower cardiovascular risk (http://statindecisionaid.mayoclinic.org). Measuring the occurrence of SDM is a developing science, and in the scenario of clinicians recommending statins for secondary prevention, potential measures to assess if SDM occurred include: 1) Does the patient know his or her personalized cardiovascular risk? 2) Was a statin offered to reduce risk? 3) What decision did the patient make about whether or not to initiate statins?
4.2 “Prescribed” Versus “Offered”
A true measure of SDM would assess the process of imparting information on statin therapy, including benefits, harms, and alternative approaches, along with the patient “outcomes” of that discussion. The possible outcomes of SDM on statin therapy would be that the
1. Patient agrees to initiate statin therapy and gets a prescription
2. Patient declines to initiate statin therapy
3. Patient is undecided and will continue deliberations
Given the short time frame available for completing this focused update, it was determined that developing a measure of SDM that could be implemented inclusive of its key components was not possible, and the decision was made to defer this task to the writing committees of each of the individual PM sets at the times of the next full revisions. At the same time, the writing committee determined that it would be important to put forward measures that more closely approximated SDM than do measures of prescription only.
Prescription-only measures have the disadvantage that they assess an action (prescribing medication) that is completely under a provider’s control but one that can be performed without any participation by the patient. In other words, prescription-only measures reflect none of the patient outcomes of SDM.
The writing committee thought that, at the least, some indication of the outcome of a patient declining a statin prescription should be included in the numerator. An example of reporting such a PM is presented in Figure 1. The ACC/AHA PM methodology states that if the provider documents the patient’s refusal of medication, the patient should be removed from the denominator of the measure as an exception for patient reasons (2). In these instances, the patient is removed from both numerator and denominator. This approach does accommodate SDM. However, the writing committee decided to construct measures whereby patients with exceptions for patient reasons would be retained in both numerators and denominators in order to “give credit” for SDM rather than have these efforts simply disappear from measures should the patient decline medication. The combination of “prescribed” and “exception for patient reasons” was termed “offered” and was considered to be a surrogate for the patient outcomes of agreement and decline and to represent an initial step away from the current state of measuring prescribing toward more comprehensive and direct measures of SDM. The methodology of capturing patients with exceptions in the numerator was actually used in the construction of the prior PAD statin measure so that patients with LDL-C levels <100 mg/dL who had medical or personal reasons for not being prescribed a statin were retained in the numerator (6). Additionally, a strength of using patient exceptions in this manner is that it does not require capturing any new data elements for the measures but is simply an alteration of measure construction using the same measure components.
The writing committee also thought that it was important for the measures to reflect the Cholesterol Guidelines’ recommendation that patients with clinical ASCVD be offered high-intensity statin therapy and receive moderate-intensity statin when high-intensity statin therapy is contraindicated according to each manufacturer’s prescribing information or when characteristics predisposing the patient to statin-associated adverse effects are present (5). Again, it was determined that the statin dose should also be subjected to SDM. As such, patients “offered” high-intensity statin therapy are included in the numerators and denominators of the revised measures, as are patients who have documented medical exceptions (2) for not being offered a high-intensity statin and who are “offered” a moderate-intensity statin instead. Patients with medical exceptions to moderate-intensity statin therapy are excluded both from the numerators and denominators. High-intensity and moderate-intensity statins are defined according to the dosing tables included in the Cholesterol Guidelines (5).
4.3 Prescription Versus Adherence
The writing committee also considered measures of patient adherence to statin medications but ultimately concluded that, as with prescription-only measures, such measures would not be optimal for assessing provider performance and improving quality of care. Measures of adherence suffer from the same inadequacy as measures of prescription in that they ignore patient preferences. Furthermore, adherence is a different process, occurring after the patient has agreed to a treatment plan as a result of SDM. Measures of adherence oversimplify complex human behaviors, and when used in provider incentive and public reporting programs, put the onus for adherence entirely on the clinician. Additionally, measuring adherence is quite difficult, particularly when using clinical data from the electronic health record (EHR).
Several components ultimately determine whether a patient derives maximum benefit from medications that decrease cardiovascular event rates. Clinicians must prescribe the medication at an optimal dose and the patient must take the medication as instructed on a long-term basis. There is still an opportunity to improve persistence in taking statins among patients with known cardiovascular disease. This is especially true for patients with PAD: only 33% were using statins within 3 months of incident diagnosis, whereas 37% were using statins 18 months after incident diagnosis (41).
Ensuring adherence is not simply a matter of writing a prescription and advising the patient to take the medication (42). It has been estimated that at 2 years’ follow-up, one-half of patients are no longer taking statins (42). The discontinuation rate is highest in patients with asymptomatic chronic diseases such as hypertension and hypercholesterolemia (42). In those who do not discontinue the medication completely, studies have demonstrated that patients take fewer than 50% of the doses prescribed (42). Nonadherence to taking medication is an important public health consideration, affecting health outcomes and overall healthcare costs (43). The concept of shared accountability as related to long-term adherence and PMs is important to achieve the ultimate goal of improved patient outcomes and quality of life. Shared accountability must include the healthcare team, the healthcare system, and the patient (44). At the same time, it remains important that the clinician and patient have ongoing discussions about statin therapy and the reasons for less than optimal adherence if such is found to be the case.
Intentional and Unintentional Nonadherence
1. Cost of medication
2. Patient cannot afford the co-pay
3. Unclear label instructions
4. Patient forgetfulness
5. Adverse effects from medication that patient is too embarrassed to discuss with doctor
6. Patient does not like the idea of having to take medication
7. Patient does not understand the importance of a given medication for a condition for which he or she has no symptoms
8. Patient–practitioner relationship is suboptimal
9. Polypharmacy and complexity of regimen
It is important to understand the concept of “unintentional nonadherence,” which is defined in the Institute of Medicine report Health Literacy: A Prescription to End Confusion as “the degree to which individuals have the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions (47).” Unintentional nonadherence is thought to be a passive process on the part of the patient and may involve a lack of understanding of physical problems, resulting in an inability to follow treatment instructions, impaired manual dexterity, poor eyesight, or forgetfulness. In a recent systematic review and meta-analysis, it was noted that there is a statistically significant relationship between health literacy and medication adherence; however, the magnitude of effect was small when compared with other causes of nonadherence, such as medication beliefs and cost (48). One of the most important aspects of long-term medication adherence is to review the medication list; ask about adverse effects, cost, and adherence; and discuss barriers to adherence at every office visit (44). Finally, the shared accountability of all of those involved in the prescription process, including the patient, is critical to the success or failure of long-term medication adherence (44).
Medication adherence is dependent on a complex interrelationship between the disease and demographic and socioeconomic factors; beliefs of the patient and the patient’s family and friends; the medication itself; and the healthcare system (patient−care provider interaction, access to care and care organizations) (49). It is readily apparent that an individual clinician should not be solely accountable for a patient’s adherence to medication, because many factors related to nonadherence are out of the healthcare provider’s control, although there are measures the clinician can take to increase the likelihood of adherence. These include the quality of the patient−care provider relationship, SDM, and avoidance of overly complex and expensive medication regimens.
A Cochrane Database systematic review concluded that some interventions were nominally effective in the short term. However, it was found that current methods of improving adherence for chronic health problems are mostly complex and not very effective (43). These complex solutions to adherence included combinations of more convenient care, information, counseling, reminders, self-monitoring, reinforcement, family therapy, psychological therapy, mailed communications, crisis intervention, manual telephone follow-up, and other forms of additional supervision or attention (43). Even with the most effective methods to increase adherence, the magnitude of effect was quite small. Clearly, much more research is needed to determine the causes of nonadherence and to develop patient- and systemwide strategies to reduce the likelihood that the patient will stop taking medication shown to have beneficial effects on overall health.
4.4 Exceptions and Exclusions
The detailed description of exceptions is central to the characterization of the PM for the treatment of blood cholesterol to reduce the risks of ASCVD in adults with established disease. In developing exceptions and exclusions for these measures, the writing committee followed the ACC/AHA PM development methodology (2,3), which is concordant with that of the American Medical Association−Physician Consortium for Performance Improvement (now PCPI) (50). Notably, measure exceptions are based predominantly on clinical judgment, individual patient characteristics, or patient preferences. On the other hand, exclusions are used in circumstances not requiring clinical judgment to factor into the decision making (e.g., patients who died, left against medical advice, or were discharged to hospice). In accordance with the ACC/AHA PM development methodology (3), the writing committee maintains that if a patient has a potential contraindication but does in fact receive treatment, then that patient should be included in both the numerator and denominator of the PM. This approach recognizes that contraindications may be relative and rewards clinicians for exerting clinical judgment and successfully fulfilling the PM (by offering a medication when the clinician believes the benefits outweigh the risks).
The writing committee recognizes that there are justifiable reasons for patients not receiving a service that is the subject of a process PM. According to the ACC/AHA report “New Insights Into the Methodology of Performance Measurement” (3), exceptions are used in these instances because the data from these patients should still be captured for the purposes of internal quality improvement analyses. Exceptions can be related to medical reasons (e.g., the patient’s allergic history, concern for potential adverse drug interaction, and intolerance to therapy), patient reasons (e.g., patient preference, social or religious reasons, economic reasons), or system reasons (e.g., lack of available resources, insurance coverage/payer-related limitation). The writing committee has provided for exceptions from the denominator for medical reasons only and does not allow exceptions for system reasons. The writing committee concluded that the frequency of events that could be interpreted as system reasons for not offering a statin would be very low and would not likely have a material impact on measure results, making it unnecessary to make provisions for them in measure construction. (As noted previously, patients with exceptions for patient reasons are retained in the denominator.) Additionally, the writing committee has chosen to use only exceptions for most patients with ASCVD but has also made exclusions for patients with STEMI and NSTEMI and those undergoing PCI in concordance with the methodology previously used in the PM documents for these patients (7,8).
The ACC/AHA PM methodology (2,3) advocates that all patients who receive the treatment (e.g., statin) should be included in the numerator and denominator of the measure and that the assessment of the documented contraindications to therapy will be undertaken only among the remaining patients who did not receive it. As previously noted (7), some contraindications are relative or temporary or both and may resolve between the time of documentation and provision of therapy. This approach will likely help eliminate false exclusions of patients who are ultimately appropriately treated and further decrease the burden of data abstraction.
The writing committee endorses the recommendations in the ACC/AHA PM methodology report (3) that clinicians document the specific reasons for exclusions and exceptions in the patient’s health records for purposes of optimal patient management, future research, and audit-readiness; to identify practice patterns and opportunities for quality improvement; and for accountability purposes. The sources of data where exclusions and exceptions can be sought include all administrative data and claims, prospective flow sheets, and patient health records (both electronic and hard copy [paper]). The writing committee maintains that the abstractor may wish to pay close attention to the clinician’s documentation. In the patient health record, any of the above care providers must still link the specific reason reported for the nonuse of statins for the documentation to count as a reason for not offering moderate- to high-intensity statin (as an example: “Patient receiving hemodialysis; no statins needed”).
The writing committee has revised the exceptions and exclusions for the lipid measures from the prior PM documents (6–9). Most importantly, it has decided not to exclude patients with a known LDL-C level <100 mg/dL, as was done specifically in the “ACC/AHA 2008 Performance Measures for Adults with ST-Elevation and Non-ST-Elevation Myocardial Infarction” (7). This is in alignment with the Cholesterol Guidelines (5), which do not recommend an approach of treat-to-cholesterol target but rather the use of fixed doses of cholesterol-lowering drugs (specifically statins) to reduce the risk of ASCVD. This is also concordant with the evidence from RCTs showing that ASCVD events are reduced by using maximally tolerated statins in those patients shown to benefit (23). Notably, statin dosage increases occurred in only a few RCTs with the intent of maximizing therapy, but these were not true tests of defining optimal goals for LDL-C, because not all persons in the statin treatment groups received drug therapy titrated to achieve a specific LDL-C, nor were specific treatment targets compared (5).
Denominator Exclusions (Not Included in Numerator or Denominator)∗:
• Patient died
• Patient left against medical advice
• Patient discharged to hospice or for whom a comfort measures−only order is documented
• Patient transferred to another hospital for inpatient care
Denominator Exceptions (Not Included in Numerator or Denominator):
• Documentation of medical reason(s) for not prescribing a statin (e.g., allergy, intolerance to statin[s], hepatic failure, hemodialysis, heart failure, other medical reasons)
4.5 Method of Reporting
In selecting a methodology for measures that calls for inclusion of patients in the numerator for 1 of 2 reasons, that is, patients for whom statins were prescribed and patients for whom statins were not prescribed for patient reasons, the writing committee recognizes the need for more visibility for both components of the numerator. The writing committee therefore recommends that reporting of these measures include the proportion of patients for whom statins were prescribed and the proportion for whom statins were not prescribed due to exceptions for patient reasons. In this construct, the percentage of patients offered statins remains the PM for accountability purposes, whereas reporting the additional data provides full transparency for the components of the measure along with information useful to the care team in understanding their performance (Figure 1).
4.6 Limitations and Unintended Consequences
The writing committee recognizes that a different approach to medication PMs from that used with most prior such measures is presented in this paper. As such, at this time conclusions cannot be drawn about the effectiveness of these measures in helping to close the performance gaps described above. Furthermore, as with all PMs, the potential for unintended consequences exists for these measures and may be greater because of the different methodology being used. For instance, when PMs are used in accountability programs, the risk of “gaming” by clinicians is always present. Clinicians could, for example, convince patients not to take high-intensity statins for fear of adverse effects and still receive “credit” under the methodology described herein. The method for reporting these measures described in this paper should mitigate the potential for gaming, but it remains to be seen whether or not that will be the case. Although the writing committee is convinced that these measures will assist clinicians in improving patient care, careful evaluations of benefits and unintended consequences should be carried out, particularly when they are used in pay-for-performance and public reporting programs.
5 Future Directions
5.1 Improved Information Systems for Capturing Clinical Data
Measures of provider-patient interaction, including SDM, require data from clinical sources such as EHRs and clinical registries and cannot be derived from insurance claims. Unfortunately, challenges remain for implementation of such measures in clinical data sets. EHRs have a relatively constrained number of discrete data fields related to provider-patient encounters that can be used for calculating measures. In addition, significant variation in data definitions exists among registries and EHRs and among instances of ostensibly the same EHR, making comparisons among providers difficult. Finally, and of particular importance to measures of SDM, patient-reported data are rarely captured in EHRs or registries.
Registries have typically obtained discrete data through the use of case report forms, but completion of such forms is labor intensive and not a feasible solution in the nonresearch clinical setting. Checklists and “smart forms” are often used in EHRs to collect key data elements for PMs, particularly measures of patient education and counseling. These techniques have proved to be less than satisfactory, because checking a box that SDM has occurred communicates little of what that process included. Additionally, smart forms require additional work by the provider during the patient encounter and are therefore used only sporadically.
As was the case with the original measures on which they are based, the revised lipid measures were designed with the current reality of clinical information systems in mind. For truly robust measures of SDM to be feasible, significant improvements in collecting clinical data will be required. Efforts to capture data in the course of care, as opposed to requiring providers to enter data, are being made by a variety of health information technology companies and by registries such as the National Cardiovascular Data Registry PINNACLE Registry, which employs a system integration software utility (51) to extract elements from the EHR that are necessary for measures calculation. Structured data capture whereby data from the EHR would be collected as discrete elements has been recognized by the Office of the National Coordinator for Health Information Technology as a priority and is the subject of a major initiative (52). An example of a “structured format” is a standardized provider office note that contains discrete fields for key data elements potentially including those necessary for assessing SDM.
Methods for capturing patient-reported data for comparative effectiveness research are also being investigated, with significant interest in the use of EHR patient portals for that purpose (53). Patient-reported data collected in this way would be useful as well in assessing SDM from the patient’s point of view. Additionally, integration of EHR-derived data with national registries would provide a mechanism for implementation of standard data definitions, enabling valid comparisons and analyses employing common data models.
5.2 Measures of SDM and Shared Accountability
SDM has been a core concept that the writing committee has considered in the construction of these PMs for statin therapy in ASCVD. It is the writing committee’s rationale for the numerator to include the number of patients offered a statin as defined. This outcome reflects both patients who declined a statin as well as those for whom a prescription was provided. However, measuring whether or not SDM has occurred during a clinical encounter is a developing science and pushes the boundaries of the new field of shared accountability measures. In particular, determining whether high-quality SDM has been provided depends on understanding the process. Ideally, this involves the extent to which a patient 1) has obtained knowledge of the risks and benefits of a decision, and 2) has demonstrated engagement in the decision-making process through deliberations that consider the patient’s preferences, values, and capacity for action and agreement for long-term adherence. Future approaches to address PMs in this area will require a better understanding of both of these issues, perhaps allowing for addressing situations where patients remain undecided and continue deliberations with their providers.
In addition to measuring whether a statin was offered to reduce risk for patients with ASCVD, potential measures to assess the quality of SDM may include: 1) measuring if the patient knows his or her personalized cardiovascular risk; 2) assessing whether the patient understands the available options and their risks and benefits; 3) asking patients about interactions with their provider; and 4) determining if high-quality decision tools were incorporated into the clinical encounter. Future studies in these areas will need to determine that reliable and valid PMs may be constructed under this framework of shared accountability measures that potentially require surveying the patient about these domains. The examples listed above also must be broadened to address additional challenges in measurement of longitudinal adherence and across multiple care settings (hospital, outpatient clinic, home), as well as development of PMs and incentives that also target patients directly rather than providers alone, because having participated in SDM and having agreed to take a statin, the patient assumes responsibility for following through on this commitment or bringing any concerns back to the provider. PMs should also measure the performance of other stakeholders who share accountability for adherence, including insurance companies, policy makers (benefit design and care coordination programs), and accountable care organizations.
5.3 Conclusion and Summary
The writing committee believes that these new PMs closely support the new ACC/AHA Cholesterol Guidelines, will assist providers in providing better care to their patients with improved outcomes, and represent an advance in medication measures because they promote SDM and appropriate dosing. At the same time, the writing committee recognizes that much remains to be done to develop truly robust measures of SDM as well as measures of shared accountability for medication adherence.
American College of Cardiology
Patrick T. O’Gara, MD, FACC, President
Shalom Jacobovitz, Chief Executive Officer
William J. Oetgen, MD, MBA, FACC, Executive Vice President, Science, Education, and Quality
Lara Slattery, MHS, Senior Director, ACC Scientific Reporting
Jensen S. Chiu, MHA, Team Lead, Quality Measurement
Laura L. Ritzenthaler, PA, MBA, Associate, PINNACLE Registry
Amelia Scholtz, PhD, Publications Manager, Clinical Policy and Pathways
Penelope Solis, JD, Associate, Clinical Measurement
American College of Cardiology/American Heart Association
Naira Tahir, MPH, Associate, Clinical Measurement
American Heart Association
Elliott Antman, MD, FAHA, President
Nancy Brown, Chief Executive Officer
Rose Marie Robertson, MD, FACC, FAHA, Chief Science and Medical Officer
Gayle R. Whitman, PhD, RN, FAHA, FAAN, Senior Vice President, Office of Science Operations
Melanie B. Turner, MPH, Science and Medicine Advisor, Office of Science Operations
Jody Hundley, Production Manager, Scientific Publications, Office of Science Operations
Appendix A 2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures: Performance Measure Set
Appendix C Peer Reviewer Relationships With Industry and Other Entities—2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures
Appendix D 2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures: Summary Analysis Table
Endorsed by the American Academy of Family Physicians, American Association of Cardiovascular and Pulmonary Rehabilitation, American Geriatrics Society, American Society of Health-System Pharmacists, Association of Black Cardiologists, Inc., Preventive Cardiovascular Nurses Association, Society for Cardiovascular Angiography and Interventions, Society for Cardiovascular Magnetic Resonance, Society for Vascular Medicine, Society for Vascular Nursing, Society for Vascular Surgery, Society of Cardiovascular Computed Tomography, Society of Interventional Radiology, and Society of Thoracic Surgeons.
This document was approved by the American College of Cardiology Board of Trustees in November, 2014 and the American Heart Association Science Advisory and Coordinating Committee in October, 2014.
The American College of Cardiology requests that this document be cited as follows: Drozda JP Jr, Ferguson TB Jr, Jneid H, Krumholz HM, Nallamothu BK, Olin JW, Ting HH. 2015 ACC/AHA focused update of secondary prevention lipid performance measures: a report of the American College of Cardiology/American Heart Association Task Force on Performance Measures. J Am Coll Cardiol 2016;67:558–87.
Between July 8, 2014, and July 22, 2014, the document underwent a 15-day peer review period. Between July 15, 2014, and August 5, 2014, the document underwent a 21-day public comment period.
This article has been copublished in Circulation: Cardiovascular Quality and Outcomes.
Copies: This document is available on the World Wide Web sites of the American College of Cardiology (acc.org) and the American Heart Association (my.americanheart.org). For copies of this document, please contact the Elsevier Reprints Department via fax (212) 633-3820 or e-mail .
Permissions: Multiple copies, modification, alteration, enhancement, and/or distribution of this document are not permitted without the express permission of the American College of Cardiology. Requests may be completed online via the Elsevier site (http://www.elsevier.com/authors/obtaining-permission-to-re-use-elsevier-material).
↵∗ Apply specifically to STEMI/NSTEMI and PCI measure sets.
- American College of Cardiology Foundation and the American Heart Association, Inc.
- Bonow R.O.,
- Masoudi F.A.,
- Rumsfeld J.S.,
- et al.
- Spertus J.A.,
- Eagle K.A.,
- Krumholz H.M.,
- et al.
- Spertus J.A.,
- Bonow R.O.,
- Chan P.,
- et al.
- Anderson J.L.,
- Heidenreich P.A.,
- Barnett P.G.,
- et al.
- Stone N.J.,
- Robinson J.G.,
- Lichtenstein A.H.,
- et al.
- Olin J.W.,
- Allie D.E.,
- Belkin M.,
- et al.
- Krumholz H.M.,
- Anderson J.L.,
- Bachelder B.L.,
- et al.
- Nallamothu B.K.,
- Tommaso C.L.,
- Anderson H.V.,
- et al.
- Drozda J.,
- Messer J.V.,
- Spertus J.,
- et al.
- Normand S.L.,
- McNeil B.J.,
- Peterson L.E.,
- et al.
- Kumar A.,
- Fonarow G.C.,
- Eagle K.A.,
- et al.
- Maddox T.M.,
- Chan P.S.,
- Spertus J.A.,
- et al.
- Arnold S.V.,
- Spertus J.A.,
- Tang F.,
- et al.
- Smolderen K.G.,
- Spertus J.A.,
- Tang F.,
- et al.
- Shah N.D.,
- Dunlay S.M.,
- Ting H.H.,
- et al.
- Arnold S.V.,
- Kosiborod M.,
- Tang F.,
- et al.
- Roger V.L.
- Kashani A.,
- Phillips C.O.,
- Foody J.M.,
- et al.
- Nallamothu B.K.,
- Hayward R.A.,
- Bates E.R.
- Krumholz H.M.,
- Hayward R.A.
- Hayward R.A.,
- Krumholz H.M.
- Elwyn G.,
- Dehlendorf C.,
- Epstein R.M.,
- et al.
- Ting H.H.,
- Brito J.P.,
- Montori V.M.
- Coylewright M.,
- Branda M.,
- Inselman J.W.,
- et al.
- Subherwal S.,
- Patel M.R.,
- Kober L.,
- et al.
- Haynes R.B.,
- Ackloo E.,
- Sahota N.,
- et al.
- Peterson E.D.,
- Ho P.M.,
- Barton M.,
- et al.
- Bailey S.C.,
- Oramasionwu C.U.,
- Wolf M.S.
- Nielsen-Bohlman L.,
- Panzer A.M.,
- Kindig D.A.
- Zhang N.J.,
- Terry A.,
- McHorney C.A.
- ↵Physician Consortium for Performance Improvement. PCPI position statement: specification and categorization of measure exclusions: recommendations to PCPI work groups: Available at: http://www.ama-assn.org/ama/pub/physician-resources/physician-consortium-performance-improvement/about-pcpi.page?. Accessed January 6, 2015.
- ↵American College of Cardiology. NCDR PINNACLE Registry. Available at: https://www.ncdr.com/WebNCDR/pinnacle/PINNACLErequestinfo. Accessed January 6, 2015.
- ↵Office of the National Coordinator for Health Information Technology. Standards and interoperability framework: Available at: Available at http://www.siframework.org/. Accessed April 17, 2014.
- ACC/AHA Task Force on Performance Measures
- Table of Contents
- 1 Introduction
- 2 Methodology
- 3 2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures
- 4 General Discussion
- 5 Future Directions
- Appendix A 2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures: Performance Measure Set
- Appendix B Author Relationships With Industry and Other Entities (Relevant)—2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures
- Appendix C Peer Reviewer Relationships With Industry and Other Entities—2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures
- Appendix D 2015 ACC/AHA Focused Update of Secondary Prevention Lipid Performance Measures: Summary Analysis Table