Author Affiliation: Department of Cardiovascular Medicine, University of Florida, Gainesville.
Improving quality of health care delivery has been a major focus since the publication of the Institute of Medicine report “Crossing the Quality Chasm.”1 While concerns about funding and access to quality health care have not been resolved, there is overwhelming agreement that the foundation of efforts to improve care is predicated on outcomes measurement.2
Since the 1980s, 2 professional societies, the American College of Cardiology (ACC) and the American Heart Association (AHA), have collaborated on guidelines for cardiovascular disease (CVD). The methodology is a rigorous, systematic, peer-review process of scientific evidence to develop documents that guide practice.3 Performance measures developed for CVD are the logical progression following implementation of guidelines as an indicator of quality of delivered care. Similar rigorous methodology for performance measures was developed by the ACC/AHA Task Force on Performance Measures in 2000.2 Categorized as either “performance measures” or “quality measures,” performance measures are designated as appropriate for both quality improvement and external reporting and quality measures are those appropriate for quality improvement and not for external reporting until further validation and testing. Performance measures are complex and, in the case of cardiac risk factors such as hypertension and lipid levels, have to address whether patients are “treated” and whether they achieve “control.” All measures have limitations and pose challenges to ensure that they are accurate reflections of intended outcomes, ie, measurement of blood pressure (BP) control. Reporting of performance measures is crucial to ensure that care is appropriately evaluated, modified, and improved to ensure that patients receive the highest levels of quality care.
In this issue of the Archives, findings on performance of 879 Department of Veterans Affairs (VA) medical centers and outpatient clinics for hypertension management in diabetic patients are presented. In this analysis of 713 790 patients treated during 2009-2010, Kerr et al4 have developed a performance measure for BP that is linked to clinical action measures (systolic BP [SBP] <140/diastolic BP [DBP] <90 mm Hg; or SBP <150 mm Hg /DBP <65 mm Hg; or SBP <150 mm Hg and receiving ≥3 or more moderate-dose BP medications; or an appropriate clinical action within 90 days). This methodology is important because it allows for acknowledgment of reaching a specific risk factor level, while it also attributes credit for working to achieve a goal over time rather than a single point in time. The authors also specified a marker of potential hypertension overtreatment (SBP <130 mm Hg/DBP <65 mm Hg and either receiving ≥3 BP medications, starting a new BP medication class within 90 days, or receiving an increase in BP medication dose within 90 days of the index BP) to define a group of patients who may have received aggressive, potentially risky treatment.
One important aspect of this Commentary is to acknowledge the VA for its systemwide approach, commitment to quality care, and forward thinking to implement an electronic medical record system that permits analysis of care across their very large system using both clinical and pharmacy data. In the report by Kerr et al,4 94% of diabetic patients met the clinical action measure for BP (82% had a BP <140/90 mm Hg and 12% had a BP ≥140/90 mm Hg but appropriate management). The attainment of such a large percentage of patients who achieved performance measures is an outstanding accomplishment.
That being said, there are concerns with the performance measure definitions used in this report. The issues noted herein illustrate concerns regarding the complexity of developing performance measures and their applications. While intended use is always a function of quality improvement, there may be unintended potential use as scorecards for health systems and individual practitioners in a more punitive way. The first concern is that the performance measure does not match available guideline recommendations at the time the index BPs were collected. The available hypertension and diabetes guidelines were the Seventh Report of the Joint National Committee (JNC7) hypertension guidelines published mid 20035 and the 2009 American Diabetes Association (ADA) guidelines.6 Blood pressure goals for diabetic patients in JNC7 were SBP lower than 130 mm Hg/DBP lower than 80 mm Hg, and the ADA recommended the same. Neither guideline specified a lower threshold; however, the ADA reported scarcity of data for lower goals, noting that the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial was in process and might further define care. However, ADA 2012 recommendations7 for BP list an SBP lower than 130 mm Hg as appropriate for most diabetic patients (level of evidence C), and based on patient characteristics and response to therapy, higher/lower SBP targets may be appropriate (level of evidence B), with a DBP lower than 80 mm Hg (evidence level B). Kerr et al,4 based on expert consensus and review of emerging data that lower targets may increase risk in diabetic patients, set the performance measure of SBP lower than 140 mm Hg/DBP lower than 90 mm Hg, but, in fact, this would constitute undertreatment if the guideline recommendations were the basis of the performance measure. The “slippery slope” here is developing metrics at a future point in time and assessing practitioner compliance to a standard that was not in place at the time the measurement was collected. This methodology may, as in this case, misrepresent care as inappropriate, or over or underreport compliance. If the data are reevaluated based on the guidelines for SBP lower than 130 mm Hg/DBP lower than 80 mm Hg, the percentage of patients achieving the measure would presumably be less than reported. This illustrates potential for achieving a variety of outcomes depending on the measure definition. While no one would argue achieving a goal of SBP lower than 140mm Hg/DBP lower than 90 mm Hg in more than 700 000 patients is not a tremendous accomplishment, they did not meet the 2009 guideline targets.
Of additional concern is the definition of overtreatment. While overtreatment and polypharmacy are very important concerns for patient safety,8 the definitions are based on the authors' critical critique of current research involving management of hypertension in patients with diabetes, but not the guidelines in place at the time of measurement. With no guidance as to a lower threshold other than epidemiological (for BP) and other risk factor (for low-density lipoprotein cholesterol) data that “lower is better,”9 one could argue that lower BPs were an indicator of better care, not “overtreatment.” Because there is no reporting of increased risk and no clear understanding of the complexity of comorbidities, a blanket statement that 10% of the VA population may be overtreated creates a negative impression of care that might not be true. One could argue that there was no overtreatment based on current guidelines but better attainment of recommended goals and that centers that were more successful at BP control exceeded the metric.
These issues call the question, what is a reasonable expectation for updating of guidelines and performance measures? Professional societies such as the ACC, AHA, ADA, and others are to be commended for their efforts in developing and updating guidelines, responding to the rapid influx of practice changing research by producing “focused updates” in a timely fashion. What is clearly lagging are JNC8 and National Cholesterol Education Program guidelines. Science and medicine are iterative in nature and move forward as new data emerge to challenge and advance current knowledge. As a result of fast-paced scientific discovery, guidelines and performance measure processes must be nimble so that practitioners can be guided with the most current data and performance can be measured appropriately. The reporting of performance measures is important, and the development of tightly linked clinical measures as those by Kerr et al4 is an important step forward in evaluating the complexities of management for hypertension and serve as a model for other measures. However, reporting and evaluating performance measures must ensure that performance is linked to guidelines at the time of performance.
Correspondence: Dr Handberg, Department of Cardiovascular Medicine, University of Florida, 1600 Archer Rd, PO Box 100277, Gainesville, FL 32610 (email@example.com).
Published Online: May 28, 2012. doi:10.1001/archinternmed.2012.2261
Financial Disclosure: None reported.
Handberg E. How Do Guidelines Impact Measures of Performance? Can They Keep Up?Comment on “Monitoring Performance for Blood Pressure Management Among Patients With Diabetes Mellitus”. Arch Intern Med. 2012;172(12):945-946. doi:10.1001/archinternmed.2012.2261