First Year Report Card: Inside the ICD National Registry

Medtronic’s newest implantable cardioverter defibrillator (ICD), Virtuoso, received FDA approval in May 2006. The stop watch-sized device is equipped with Medtronic’s Conexus wireless telemetry with SmartRadio technology, which uses the Medical Implant Communications Service. When the ICD detects notable changes in the patient’s condition or device status, a CareAlert is sent to the physician via the internet.

As the quest for more data in cardiac care grows with its link to reimbursement and quality of care measures, the number of national registries is sure to blossom. Last year, at the urging of the Centers for Medicare & Medicaid Services (CMS), the National ICD Registry was developed through a partnership of the American College of Cardiology Foundation (ACCF) and the Heart Rhythm Society (HRS).

The implantable cardioverter defibrillator (ICD) Registry joins the other registries in ACC’s National Cardiovascular Data Registry: ACTION Registry for acute coronary syndrome patients; Cath PCI Registry for diagnostic cardiac catheterizations and percutaneous coronary interventions; and CARE Registry for carotid revascularization and endarterectomy procedures.

The idea of the ICD registry was prompted from the National Coverage Decision that was published by CMS in 2005 and based on the Sudden Cardiac Death in Heart Failure Trial. HRS took the lead by chairing the National ICD Registry Working Group which then developed defining core characteristics and questions to be answered.

The National ICD Registry became the sole repository of ICD implantation data for Medicare beneficiaries as of April 1, 2006. From that date through July 2007, data were collected from 1,117 U.S. hospitals. Some 108,341 implantable cardioverter defribillators implantation procedures were included that were performed by 3,249 physicians during those first 15 months.

To provide a more accurate picture on the scope of the procedures, hospitals have been encouraged to enter data on all device recipients, not just Medicare beneficiaries. Nearly three-quarters (73 percent) of the participating hospitals elected to enter all of their ICD implementations into the registry. Because these hospitals are for the most part large institutions, their data accounts for 88 percent of the implementations.

In April 2007, the first benchmarking report was sent to participating hospitals. Reports are sent quarterly and include the entire outcome report for each hospital comparing their volume and data with results of similar-sized hospitals and a national aggregate. Additionally, as of July, quality improvement reports highlight key registration and utilization metrics comparing the individual hospital with a registry aggregate.

The first report showed an average patient age of 68. More than 74 percent were male and 83 percent were white. CMS beneficiaries accounted for 70 percent of the patients.

In July, the data and a review of the registry’s first year were published in the National ICD Registry Annual Report 2006, which was approved by the Hearth Rhythm Society board of trustees on July 13.

Now, as a well-functioning database, the ICD registry and its proponents face analyzing the information. According to their report, “Aspects of the current data collection require correction including missing and ambiguous data elements. Recognized limitations include the absence of device firing data and the need for longer follow-up to record complications that occur after discharge such as infection and lead dislodgement.” To address these matters, they are currently developing Version 2.0 to correct the limitations.

Some of its next steps include: affirming its purpose, goals and target audience; streamline data collection by removing data fields not used and clarifying data definitions; expanding the registry to follow pacemaker and ICD leads; meeting public policy, including use as a performance reporting tool; and adding longitudinal data collection that includes merging data with Medicare claims data, the National Death Index and home device monitoring systems.

As one answer to refining the registry, the working group developed the Longitudinal ICD Registry study to complement the baseline registry and obtain device firing data, which was approved by CMS in May. The study will include 350 randomly selected physicians and follow some 3,500 CMS beneficiaries receiving a primary prevention ICD. End points will include survival at three years and five years; death from any cardiovascular cause; total number and rate of device therapies; and ratio of inappropriate to total device therapies.

What is evident from this first year is that both short-term and long-term data on ICD implementation provides hospitals and payors with a vital tool for quality improvement and pay-for-performance programs.

One On One

Chair of the ICD National Registry steering committee and former president of the Heart Rhythm Society, Stephen C. Hammill, MD, FHRS, clinical professor of medicine and ECG laboratory director at Mayo Clinic, in Rochester, Minn., details Year One.

CVB: What is most noteworthy about the information gathered in the ICD registry during the first year? Dr. Hammill: The most important aspect of the registry is that it’s collecting information on outcomes at the time of ICD implantation and then that information is returned to the hospitals quarterly. An individual hospital is compared to the national average and to hospitals of their own size to see how they’re performing. They can see if their rates for various complications are within the range that we’re seeing with other hospitals. This is also a global overview of the type of data that are being collected by the registry which gives us a research and publication process where people can look at the data specifically and develop research proposals and publications. We have four of those research studies being presented at the American Heart Association meeting in Orlando in November. CVB: Were there any surprises in the annual report? Dr. Hammill: One thing that was a surprise—the complication rate is low; that’s a pleasing surprise. And the in-hospital death at the time of ICD implantation was 0.02 percent. In previous publications, the complication rate was about 0.5 percent—much higher than we’re seeing now. So that’s a pleasing find and to me suggests that over the years people have become more experienced with the procedure. The other pleasing find was that 85 percent of the physicians who are implanting ICDs have gone through some type of formal training to gain experience. Now we would like the other 15 percent to go through similar training. But a lot of that is dependent upon hospital credentialing committees and how strict they are when they are approving someone to implant ICDs. And the third thing that just needs to be looked at further and compared with other data and registries is the issue of race and gender. Nearly 75 percent of the people who received ICDs were male, and it’s a question of whether the device is being underutilized in women or if that’s an appropriate number. The only way we’ll be able to answer that question is to get an idea of how many women in general meet criteria for ICDs. There is an American Heart Association registry and the Medicare claims data. We’ll have to compare this registry data with those registries to get an idea if the device is being used appropriately in women; the same holds true for the race distribution. [Eighty-three percent of the patients were white.] CVB: How important is it to collect ICD data? Dr. Hammill: I think there are two places where it’s important. One is hospitals now are all working on ways of improving their own quality and outcomes. The registry gives them benchmarking data to compare their ICD program with other hospitals. Secondly, insurance companies are looking at the registry in terms of pay-for-performance or their quality programs and starting to look at these registries as something that needs to be done by their providers to insure quality. For example, United Healthcare has made it mandatory for participants in their program to participate in the registry. If a hospital wants to be labeled as a United Health Premium Cardiac Specialty Center, they have to be entering patients into the registry. Basically, United is doing that as a means to track performance. CVB: Are there any weaknesses within the IDC reporting system? Dr. Hammill: The weakness is getting good quality follow-up data. We have the data on ICD implantation, but we don’t have follow-up data. That’s something we need and we plan to do that in two ways. One: we will enhance some of the follow-up aspects of the registry data that are being collected. Two: we will combine this data with Medicare claims data and the National Death Index. If someone is readmitted for an infection, problem with a defibrillator, or other complication of the procedure, it gets entered into these two large databases. Then we’re able to match up the two databases to track longer-term results. CVB: Beyond the need for payments and the ability to collect data, what is driving the use of national registries? Dr. Hammill: Hospitals and physicians are worried about the report-carding approach that has been done in the past by insurance companies and state medical organizations or state departments of medicine. The two most visible examples are from New York and Pennsylvania, where in the past they did report-carding of cardiac surgeons with regard to their mortality rate. Because they didn’t have the data on how sick the patients were, some of the bigger hospitals and more experienced surgeons had a higher mortality rate. In the end, it was shown that it was because they were being referred sicker patients. That’s what prompted the Society of Thoracic Surgery (STS) to put together the STS registry in the 1990s, and that same kind of report is being looked at for other areas of medicine. I think the advantage of that kind of registry is that it collects data beyond the claims or administrative data collected by the insurance company. It gives a more accurate picture of the types of patient who are being treated. CVB: Do you expect more registries to be developed? Dr. Hammill: Yes, because it’s a good way of looking at real life experience with devices and therapies. We are looking to develop a complex catheter ablation registry for catheter ablation and ventricular tachycardia. Then I suspect there will be registries developed in other areas of medicine for Medicare and the Agency for Healthcare Research and Quality, who is interested in registries to see how practice is being performed outside of controlled trials. CVB: How do national registries affect the practice of cardiology? Dr. Hammill: National registries improve the programs locally, so if the program is noting a higher complication rate or longer length of stay, hospitals will try to figure out why and improve it. We think having this type of data out there will help to insure that physicians who do these procedures have adequate training because they will begin to see that this type of information is being tracked.

AHA Scientific Sessions 2007 Nov. 4-7, Orlando, Fla. ICD Presentations Tuesday, Nov. 6 → 9:15 a.m. ICD for Primary Prevention Session: Patients Who Receive ICD for MADIT-II Criteria in Clinical Practice Are Different from Patients Enrolled in MADIT-II Tuesday, Nov. 6 → 11 a.m. Disparities in Cardiac Care Session: Racial Differences in Nationwide Utilization of Cardiac Resynchronization Therapy Tuesday, Nov. 6 → 11:15 a.m. Disparities in Cardiac Care Session: Differences in Implantation-related Complications Between Men and Women Receiving ICD Therapy for Primary Prevention Tuesday, Nov. 6 → 11:45 a.m. Disparities in Cardiac Care Session: Sex Differences in the Characteristics of Patients Receiving ICD Therapy for Primary Prevention Wednesday, Nov. 7 → 2:57 p.m. Moderated Poster Session: Racial and Ethnic Differences in Cardiac Resynchronization Therapy Utilization Outside of Published Guidelines