New research shows substantially different death rates for ischemic heart disease (IHD) and chronic heart failure (CHF) across the Veteran Affairs (VA) healthcare system, even after adjustment for a wide range of health factors.
Among nearly 1 million veterans at 138 VA medical centers, risk-adjusted annual mortality rates for IHD ranged from 5.5% to 9.4% (P < .001).
The gap was even wider for CHF, where risk-adjusted mortality rates varied from 11.1% to 18.9% (P < .001).
Risk-adjusted mortality rates were significantly higher than the national average at 21% of VA hospitals for IHD and 25% for CHF. On the other hand, they were significantly lower than the national average for IHD and CHF at 26% and 27% of VA hospitals, respectively, according to the study, published May 16 in JAMA Cardiology.
“It wasn’t easy actually to put hospitals in a bucket and say the hospitals that are in urban areas are really struggling and the hospitals in the suburbs are thriving. That wasn’t the case,” lead author, Peter Groeneveld, MD, Philadelphia VA Medical Center, Pennsylvania, told theheart.org | Medscape Cardiology. “It’s sort of like: all politics is local; well, all healthcare is local too.”
Prior studies have shown the VA system provides the same or better quality of medical and surgical care compared with healthcare systems in the private sector. Credit often goes to its early and intense focus since the 1990s on standardizing processes-of-care measures and incentives across the system. The problem is that by requiring all VAs to be close to identical in terms of management for chronic disease, it is not necessarily going to be responsive to individual local populations, which couldn’t be more different, for example, at his VA in West Philadelphia vs that by his mother in White River Junction, Vermont, Groeneveld said.
“The national standards and national quality program in many ways have constrained the VA to not be as creative on the local level and adopt a quality program to the specific needs of their specific veteran population, and I think that is one of the central messages of the paper,” he said.
The analysis included 930,079 veterans with IHD and 348,015 with CHF receiving care from April 2010 to September 2014 at 138 of the VA’s 144 medical centers.
Veterans with IHD were predominantly white men, with 82% having hypertension, 49% diabetes, 23% chronic pulmonary disease, and 17% heart failure. Veterans with CHF were slightly more likely to be women and/or not white, with 84% having hypertension, 61% diabetes, 36% chronic pulmonary disease, and 26% chronic kidney disease.
The mean age in both cohorts was 77 years, and 266,406 veterans were members of both cohorts. The ratio of patients with HF to patients with IHD varied from 18% to 52% across VA hospitals.
The unadjusted annual mortality rate was 7.4% in the IHD cohort and 14.5% in the HF cohort.
Mean composite IHD-CHF standardized survival scores were higher in the East (z score, 0.51) and Midwest (z score, 0.53) than in the West (z score, –0.29) and South (z score, –0.58) regions (P < .001).
The 30-Day Elephant in the Room
A comparison with other published quality measures, however, found no association between a hospital’s annual IHD mortality rate and rate of 30-day acute MI posthospitalization mortality (R2 = 0.01; P = .35) and only a weak correlation between its annual CHF mortality rate and rate of 30-day CHF posthospitalization mortality (R2 = 0.16; P < .001).
Commenting to theheart.org | Medscape Cardiology, Gregg Fonarow, MD, director of the Ahmanson-UCLA Cardiomyopathy Center, Los Angeles, California, said, “The degree of variation in mortality is surprising, but this really puts a focus on this issue of the very narrow focus that CMS [Centers for Medicare & Medicaid Services] has taken to overly weight 30-day rehospitalizations — to ignore largely quality of care in outcomes in the ambulatory care settings for heart failure and ischemic heart disease from meaningful quality and outcome metrics. It’s really way too narrow, short-sighted, and not meaningful.”
Within the VA setting itself, many of the conventional process-of-care metrics are superficial or inadequate, he said. For example, physicians can have a patient on a trivial dose of β-blockers and still receive full credit; newer, more important medications, such as sacubitril/valsartan (Entresto, Novartis), are not included; and statins as a yes/no variable fails to capture whether patients are adherent or on high-intensity statin therapy.
“The real point is really needing meaningful, discriminatory, up-to-date process-of-care measures and providing appropriate, timely risk-adjusted outcome measures back to centers,” Fonarow said. “Because the number of lives that we could be talking about here just in the VA and in those centers that were performing below the median for annual survival, we’re talking about tens of thousands of additional lives that could be saved each year.”
In an accompanying editorial, Paul Heidenreich, MD, VA Palo Alto Health Care System, California, writes that the lack of correlation between 30-day outcomes and annual mortality rates “would imply that something other than quality of care was being measured.”
“Perhaps the most important question is how hospitals can use these data to improve…. Without measures of process of care known to improve mortality, it is not clear how hospital leaders can intervene to lower the mortality rate,” he said.
Heidenreich proposes three steps going forward: enhanced electronic data sharing to allow tracking of patients across health systems; additional health services research to learn how to best measure and attribute outcomes; and increased analytical expertise in hospital executive suites, deans’ offices, and boardrooms to use the research findings.
“Perhaps a chief evidence officer (under a new acronym, “CEVO”) can be hired who will guide leadership to consider measure validity, reliability, and, most importantly, uncertainty,” he said. “Decisions in the face of uncertain quality metrics will always need to be made, but if the data limitations are acknowledged, then we will be better stewards of resources applied to improve quality of care.”
The research was sponsored by a grant from the Veterans Affairs’ Health Services Research and Development Service. The authors and Heidenreich have reported no relevant financial relationships. Fonarow reports serving as associate editor for JAMA Cardiology; being a member of the steering committee for Get With the Guidelines; and consulting for Novartis, Amgen, Janssen, and Bayer.
JAMA Cardiol. Published May 16, 2018. Full text, Editorial
Follow Patrice Wendling on Twitter: @pwendl. For more from theheart
Tidak ada komentar:
Posting Komentar