Author + information
- Christopher P. Cannon, MD∗ ()
- Division of Cardiovascular Medicine, Department of Medicine, Brigham and Women's Hospital, Harvard Clinical Research Institute, Boston, Massachusetts
- ↵∗Reprint requests and correspondence:
Dr. Christopher P. Cannon, Harvard Clinical Research Institute, 930 Commonwealth Avenue, 3rd Floor, Boston, Massachusetts 02215.
Quality improvement has become a standard component of clinical care over the past decade—and it has led to dramatic improvements in care. Looking back, in 2001 Fonarow et al. (1) reported that the use of statins following myocardial infarction was a very disappointing 31.7%. This and other observations of very low use of therapies well documented to improve outcomes—such as aspirin, statins, and beta-blockers—spurred the development of large-scale registries and quality-improvement programs to increase the use of these life-saving agents. Great successes have been seen with these programs, which include CHAMP (Cardiac Hospitalization Atherosclerosis Management Program) (2), GAP (Guidelines Applied in Practice) (3), Get With the Guidelines (4), CRUSADE (Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementation of the ACC/AHA Guidelines) (5), GRACE (Global Registry of Acute Coronary Events) (6), ACTION (Acute Coronary Treatment and Intervention Outcomes Network) Registry (7), and REACH (Reduction of Atherothrombosis for Continued Health) Registry (8). Over time, use of each of these agents has improved to >90%.
Although such “report cards” have helped, they currently monitor only the use of therapies proven effective >20 years ago, and as such, they do not appear to keep pace with the advances in medical therapeutics. This has prompted many to ask: “How should we improve monitoring of quality?” One approach is to add quality metrics to monitor the use of newer therapies. However, objections can be raised to quickly adding new therapies. Some believe that the appropriateness of the new agents might not apply to all patients in clinical practice, even within the approved indication, and that as such, some “clinical experience” should be gained before new agents are added to our quality reports. Others point out that while some new therapies are shown to be more effective than standard therapies, they usually cost more, and thus the lack of use may not be due to a medical reason or to a lack of quality care, but to a simple cost issue. One would not want a report of a lack of use of a newer, expensive agent to be misconstrued as a failure of the physician's quality of care if it were due to cost constraints of the patient or insurance company.
Another approach has been to look at composite measures that include all of the initial quality metrics, whereby one would monitor how many patients received all of the guideline-recommended therapies. That is, what proportion of patients (without contraindications to any one of the agents) received all 4 classes of drugs (antiplatelet, beta-blocker, statin, and angiotensin-converting enzyme inhibitor [ACEI] or angiotensin receptor blocker [ARB])? As one might expect, the rates of compliance with the quality metrics fall sharply from >90% for individual metrics to 40% to 60% for the combined metrics, (9) exposing incomplete care, even at hospitals that seemed to provide very high levels of quality. For example, in a patient with diabetes, glycemia may be well controlled with a hemoglobin A1c <7%, but is the patient receiving aspirin, a statin, and an ACEI/ARB? Similarly, in a patient with hypertension, blood pressure (BP) may be controlled, but is the patient receiving a statin (which has been shown to improve outcomes) (10)? Thankfully, after these composite measures are added to the reports generated for hospitals that participate in registries and quality-improvement initiatives, the rates of “defect-free care” improve (11).
In this issue of the Journal, Arnold et al. (12) have taken things 1 step further, introducing the dose of these medications as a new quality measure. They note that, in the clinical trials in which therapies have been shown to be beneficial, the doses of the therapies used are generally high (so as to maximize their pharmacologic and then clinical effect), yet in everyday clinical practice, the doses used are often lower.
Arnold et al. (12) analyzed data on 6,748 patients with acute myocardial infarction discharged from 31 hospitals in the United States between 2003 and 2008. The data were collected at discharge and at 1-year follow-up in 2 registries: PREMIER (Prospective Registry Evaluating Myocardial Infarction: Events and Recovery) and TRIUMPH (Translational Research Investigating Underlying Disparities in Acute Myocardial Infarction Patients' Health Status). In these populations, the use of beta-blockers, statins, and ACEI/ARBs was >87% for each class of medications, suggesting excellent quality of care.
But, as one might expect, the investigators found low levels of use at “goal” doses: only one-third of patients received doses of beta-blockers, statins, and ACEI/ARBs that were at least 75% of the target dose in the pivotal trials. In addition, during follow-up, there was up-titration of the dose in only about 25% of patients. As a result, at 12 months after myocardial infarction, only 12%, 26%, and 32% of eligible patients were receiving the target doses of beta-blockers, statins, and ACEI/ARBs, respectively.
This low percentage seems as bad as the early registry data on simple use of these agents (at any dose) from a decade ago. In an effort to offer high-quality care, physicians typically check to ensure that all appropriate, guideline-recommended medications are used, but they perhaps have not paid as much attention to the doses of these medications, and thus this paper is a real eye-opener.
One strength of the analysis is that there is good rationale for encouraging the use of higher doses. For statins (13) and ACEIs (14), head-to-head, randomized trials have documented superior efficacy with the higher versus lower doses. A second strength is that the investigators excluded patients whose systolic BP was <110 mm Hg, in whom it would seemingly be difficult to up-titrate the dose of an antihypertensive medication.
A caution, though, is the “calibration” of the metric. They have selected a seemingly generous cut point of “goal” dose—75% of the planned target dose of the medication. However, this level of compliance was not achieved even in clinical trials. For example, in 1 recent ARB trial, the mean achieved dose was 72% of the target dose (15). That is, only 50% of the patients reached 72% of the target dose. Thus, to set a benchmark of wanting to see 100% patients reach this “goal” dose is not really fair, considering that it was not possible to achieve that level of compliance in the setting of a carefully conducted clinical trial emphasizing up-titration to the target dose. Thus, we have to be careful in setting expectations and must do further work on the exact benchmark of these quality measures. One could imagine, first, a careful survey of the actual trials (likely requiring asking the primary investigators to query their trial database and to report the proportion of patients in the actual trials who achieved different percentage levels of the target dose). Then, one would have a good benchmark to begin to compare others in routine practice.
A second issue is that measuring this aspect of quality in future registries will require more data from registries—not just the simple presence of a medication on a medication list, but also the dose, and then BP, heart rate, and lipid values. These data should be available from electronic medical records, but the validity of these measures as passively collected in electronic medical records is not well defined.
Thus, we have here an important new area of research in quality measures—we now move beyond a “yes/no” metric to look at the dose of each class of drugs to ensure full use of the evidence-based therapy. More work is needed to establish appropriate “goal” doses and benchmarks and to ensure that data collected from electronic medical records are valid for analysis. Nonetheless, this paper takes a major step forward and opens up a new way to try to maximize the benefit of these life-saving therapies for our patients.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. Cannon has received research grants/support from Accumetrics, AstraZeneca, CSL Behring, Essentialis, GlaxoSmithKline, Merck, Regeneron, Sanofi, and Takeda; honoraria for serving on the advisory boards of Alnylam, Bristol-Myers Squibb, Lipimedix, and Pfizer (but funds donated to charity); and has served as a clinical advisor for and holds equity in Automedics Medical Systems.
- American College of Cardiology Foundation
- Fonarow G.C.,
- French W.J.,
- Parsons L.S.,
- Sun H.,
- Malmgren J.A.
- Fox K.A.,
- Goodman S.G.,
- Klein W.,
- et al.
- Roe M.T.,
- Messenger J.C.,
- Weintraub W.S.,
- et al.
- Kumar A.,
- Fonarow G.C.,
- Eagle K.A.,
- et al.
- Sever P.S.,
- Dahlof B.,
- Poulter N.R.,
- et al.
- Lewis W.R.,
- Ellrodt A.G.,
- Peterson E.,
- et al.
- Arnold S.V.,
- Spertus J.A.,
- Masoudi F.A.,
- et al.
- Packer M.,
- Poole-Wilson P.A.,
- Armstrong P.W.,
- et al.