Patients admitted to high-volume hospitals for acute myocardial infarction (AMI) have reduced 30-day mortality compared with those treated at lower-volume hospitals, although above a certain threshold the effect of volume on outcome tapers off. Teaching hospitals and those with revascularization capabilities—whether percutaneous or surgical—also produce superior results, according to research published in the March 25, 2010, issue of the New England Journal of Medicine.
To explore the relationship between hospital volume and mortality, researchers led by Joseph S. Ross, MD, MHS, of the Mount Sinai School of Medicine (New York, NY), examined claims data for Medicare recipients who were hospitalized between 2004 and 2006 for AMI, heart failure, and pneumonia. Based on annual condition-specific volumes averaged over 3 years, hospitals were placed into quartiles ranging from low volume (first and second quartiles) to medium (third quartile) and high volume (fourth quartile). Those that had fewer than 10 cases per year were excluded from the analysis.
Effect of Volume Significant but Tapers Off
During the study period, there were 734,972 admissions for AMI at 4,128 hospitals, of which half were categorized as low volume, a quarter as medium volume, and a quarter as high volume. Patients admitted to large-volume hospitals were younger and more likely to have undergone PCI or CABG in the previous year compared with those treated at low-volume hospitals (P ≤ 0.01). The mean annual hospitalization rates for AMI were:
While there was much heterogeneity in mortality within each volume category, increased hospital volume was significantly associated with a reduced 30-day death rate for AMI patients (risk-adjusted OR 0.89; 95% CI 0.88-0.90; P < 0.001). The association was attenuated as volume increased; at low-volume hospitals, increasing the annual number of admissions by 100 would result in a 20% reduction in mortality, whereas the same increase would reduce mortality by 10% at medium-volume hospitals and by 4% at high-volume hospitals.
Notably, once a hospital’s annual volume reached a threshold of 610 AMI patients, adding 100 more patients per year would no longer significantly decrease mortality. The volume threshold differed according to whether the hospital was an academic medical center or able to provide revascularization (either CABG or PCI). At teaching hospitals, the estimated volume threshold was 260 patients, compared with 629 for non-teaching hospitals. At revascularization-capable hospitals, the volume threshold amounted to 432, compared with 586 for those that did not perform such procedures.
For admissions related to pneumonia (n = 1,418,252) or heart failure (n = 1,324,287), volume also had an effect on outcome. However, both conditions had lower volume thresholds than did AMI. In heart failure patients, higher volume reduced 30-day mortality by 9% (risk-adjusted OR 0.91; 95% CI 0.90-0.92; P < 0.001), and the threshold was 500 patients. In pneumonia patients, mortality was lowered 5% (risk-adjusted OR 0.95; 95% CI 0.94-0.96; P < 0.001), with a threshold of 210. Teaching hospitals and revascularization-capable hospitals continued to produce superior results in these cohorts.
Findings Inform Policy, Daily Practice
The results have clear policy implications, the paper reports. For instance, policy makers may attempt to increase volume at the lowest-volume hospitals. “Perhaps [this could be implemented] by ensuring that small hospitals are not located within close proximity to one another,” they write. “This could be accomplished through state certificate-of-need regulations or critical-access-hospital programs.”
In a telephone interview, Dr. Ross told TCTMD that the “idea that greater volume leads to better outcomes has been widely demonstrated in studies of procedural and surgical care.” However, it was not known whether that type of association would be seen with medical conditions that involve more team-based care.
“There are many more moving parts, and nurses do more of the care. It’s unclear whether you would get that same efficiency of scale,” Dr. Ross said, adding that while there have been a few volume-related studies of AMI care, this is the first to look at this issue in heart failure and pneumonia.
One particularly interesting observation made by the study is the “idea that bigger isn’t necessarily better,” Dr. Ross commented, adding that this is particularly true for pneumonia. “Certainly, the smaller volume hospitals seem to have, on average, slightly worse outcomes, but the marginal benefit that you get from being at a higher volume hospital goes away [at a certain point].”
The key will be to tease out what the higher-performing small hospitals do differently, he continued. “Are they really good at identifying patients who will do better if they get transferred to a place that has revascularization capacity? Are they better at stabilizing [patients] and taking care of people? Do they reallocate their nurses to make sure the patients are being followed up, or taking the right medications?” Dr. Ross elaborated. “That is where we have the most to learn, looking at these smaller hospitals and figuring out what they do so well.”
William W. O’Neill, MD, of the University of Miami (Miami, FL), told TCTMD that the study carries messages for multiple audiences.
“For the public, it’s just a word of caution about low-volume hospitals” and their tendency to have poorer outcomes, Dr. O’Neill said in a telephone interview. And for physicians and administrators who currently practice at low-volume centers, he added, the results are a reminder to pay attention to mortality. “There are many small-volume hospitals that were getting really good results. [The study’s] not a blanket indictment of all these places, it’s just that if you go to a smaller hospital, the results could be very variable,” Dr. O’Neill explained. “Doctors and administrators practicing there have to be really cautious and ever vigilant about their results.”
While 30-day mortality rates hovered around 20% in low-volume hospitals, he pointed out that some of centers in this category had death rates surpassing 60%. “In those places that get mortality greater than 60%, you really have to ask whether they should even be allowed to take care of MI patients,” Dr. O’Neill asserted.
More Low-Volume Hospitals with Cardiology Programs
Speaking in a telephone interview with TCTMD, Tony G. Farah, MD, of Allegheny General Hospital (Pittsburgh, PA), commented that these issues have become particularly relevant as increasing numbers of small hospitals have initiated cardiac care services. Based on his own experience working with such hospitals, he was unsurprised that they tended to have poorer outcomes.
For one, hospitals that have low volume or are located in underserved areas can have difficulty recruiting cardiologists. This is especially true for interventionalists, who need to see enough cases to maintain their expertise, Dr. Farah said. Another difficulty is that physicians may not be working alongside other clinicians who share their specialty. Without collaboration, “you end up practicing based on limited resources and input from others around you,” he commented.
And while it may be possible to recruit a few experienced nurses and other staff members from larger hospitals, most will be coming from within the institution. “As much as you train them, they’re just not exposed to that sheer volume [necessary] to become an experienced nurse or technologist . . . . By creating that gap in experience, the support staff is not able to quickly recognize complications when they develop,” Dr. Farah explained. “We all know if you have a delay between the time a complication develops and the time that it’s treated, the outcome is worse.”
Moreover, hospital infrastructure may be lacking. “Financially, it’s just not feasible to have all of the support structure you have in a tertiary care center,” said Dr. Farah, citing the example of off-hours physician coverage and CT availability.
But like Drs. O’Neill and Ross, Dr. Farah was careful to point out that low-volume hospitals do not have uniformly bad outcomes. “In general, in larger hospitals you do have more structured care that’s guideline based, where you have standardized protocols for patients who come in with acute MI,” but such practices can be put in place at a hospital of any size, he advised.