Feature: DES use in stable patients doubles fiscal impact; not so w/ ICDs
 |
Peter W. Groeneveld, MD, an assistant professor of medicine at University of Pennsylvania and the Philadelphia VA Medical Center, reported the study at the American Heart Association (AHA) Quality of Care and Outcomes Research in Cardiovascular Disease and Stroke meeting last week. Shortly after, Cardiovascular Business News spoke with Groeneveld about his study.
CVB: What was the impetus for undertaking the study?
Groeneveld: While we know from the data that DES and ICDs are among the most common and most costly interventional therapies used in patients with coronary artery disease [CAD] and chronic heart failure [HF], respectively, we didn't know the actual fiscal impact of these technologies. It turns out that from 2003 to 2006, the overall Medicare costs associated with DES were double the projected economic estimates, while those associated with ICDs were not out of line with projections.
CVB: What is the projected cost estimate versus the real-world cost impact of DES?
Groeneveld: One way to figure this out is to guess and say it's the price difference between the new technology versus the old. We found the impact of DES on healthcare costs has been well above what you would expect just based on the price difference.
DES in this population were used in about 500,000 Medicare beneficiaries per year [ages 65-84]. Medicare pays $3,000 more for each procedure using a DES versus bare-metal stent. With a back of the envelope calculation, you'd say that should be about $1.5 billion. We actually find a much higher cost impact. Our current number is $3.3 billion. The reason is when DES were introduced, they changed the way medicine was practiced. Doctors ordered more tests to see if more patients would be candidates for DES and interventionalist started performing more PCIs. And there are various downstream costs because outcomes are different.
CVB: Wouldn't you expect an increased cost impact with new technology?
Groeneveld: You might say this always happens with new technology, but it didn't seem to happen with ICDs. With ICDs, the cost to Medicare is almost completely wrapped up with the cost of the device itself. If you do the same multiplication—what's been the volume of care for new ICDs, the price difference of treating someone with an ICD or not—it winds up being around $750 million. And we found a cost increase of about $774 million. Essentially, for ICDs, there hasn't been as big of a cost impact.
It turns out that there is an amplification effect for some new technologies where the presence of this new therapy means that people start practicing more expensive medicine. In the case of DES, we found that a large proportion of additional costs occurred in patients with relatively stable coronary disease. And that is concerning because there is no really good evidence that people with stable coronary disease should get PCI at all. In some sense, increasing the cost of a procedure, which really shouldn't be done—in patients with really stable coronary disease—is not where medicine should be headed if we're ever going to bend the cost curve. We should go in the opposite direction by trying to reduce costs by reducing unnecessary procedures, not by making unnecessary procedures even more expensive. That's bending the cost curve in the wrong way.
CVB: What is the ratio of DES being used in patients with AMI, non-AMI and stable coronary disease?
Groeneveld: We found that $1.3 billion was spent for patients with AMI, $700 million for patients with non-AMI but acute coronary syndromes and about $1.4 billion for patients with stable coronary disease. That last figure is 58 percent of the cost to Medicare.
CVB: How can your findings be used to possibly enact change in healthcare spending?
Groeneveld: There are a couple of things. Medicare has to take into account more than the price differential when making coverage decisions for new technologies. Also, more than half of the DES that Medicare covered were used off-label. Maybe the next generation of coverage decisions needs to be more circumscribed, much more in line with coverage with evidence development, where Medicare agrees to cover these but with the requirement for more real-world data that shows taxpayers are getting their money's worth.
CVB: Is there any way to know in advance which technologies will have an amplification effect?
Groeneveld: It's hard to know that in advance. In the world of interventional cardiology, the technologies that have behaved like ICDs in terms of being more in line with projected costs have been few and those that have behaved more like DES have been ample. One of the reasons why DES have proliferated widely is because it's really easy to use them. There was very little difference in terms of technique between delivering a DES versus a bare-metal stent.
On the one hand, it's not possible to predict perfectly what will happen with new technologies. On the other hand, it's easy to say there are not a lot of roadblocks to prevent DES from being used far afield. Back when the coverage decision was being debated in 2002, one could probably guess that off-label use was going to be an issue, but they probably did not anticipate it would be such a huge issue.
There are some things that could have put into place. In fact, the coverage decisions for DES that other countries, such as Canada and the U.K., have put into place mandate payment only to the extent that they demonstrated actual clinical benefit in the trials. It's evidence-based coverage.
CVB: Another AHA study you presented, a poster, showed ICDs have been underutilized among HF patients in the years following issuance of guidelines recommending use of primary prevention ICDs for selected systolic HF patients and that this has resulted in a significant excess in mortality. Wouldn't the increased use of these devices result in an overall increased cost impact to healthcare spending?
Groeneveld: Here is a technology that has shown clinical benefit and which seems to cost the amount we projected—and is probably not being used as frequently off-label. Unfortunately, it's on the opposite side of the technology coin, which is to say there are probably many patients whose lives would be saved if they had ICDs, but they are not getting them for one reason or another.
CVB: Do you have any theories on why that is?
Groeneveld: It turns out that many patients who get DES have no idea they have stents inside them. They signed a consent form, but that's just something a cardiologist does inside your heart. It's amazing how many patients who have had the procedure done and have no idea there is something inside their coronary arteries. They are surprised. We get people to consent, we tell them all about it and it just doesn't sink in.
Conversely, it's pretty hard to not know you have an ICD inside you. Patients know it will shock them. It turns out this type of thinking is the critical element. For whatever reason, patients, primary care physicians and cardiologists have hesitated to go the route of device care even in patients where it's entirely appropriate because of inhibitions to having a box sitting in your chest that will shock your heart. That would be my guess. There are probably other issues such as access to providers who do these types of procedures.
Our poster findings suggest that the areas in the country that have implanted these devices more frequently have lower rates of mortality in the HF population. No surprise, we think the device works. If the device is used more frequently in patients with systolic HF, lives are saved. The take home message from both studies is that if we're going to develop technology to save lives, we should use it, especially if the costs are coming in about what we expect and those costs are reasonable given the benefits we project are going to occur. Primary prevention ICDs are a reasonable deal. It's very ironic in a sense that our DES message is that there is a lot of money being spent on a therapy that has marginal, if any, benefit in a vast number of patients. The flip side is we're not doing enough to treat patients whose lives would be saved if they had an ICD.