Though clinical indications and insurance coverage of ICDs are expanding, researchers have only recently begun exploring the correlation between hospital and operator characteristics with procedural success.

Hospital volume

From 1990 to 2005, the annual rate of U.S. hospitalizations for the implantation of ICDs increased ten-fold (Am J Card 2008;101(12):1753-1755). Despite the increase, little was known about what influenced survival and complications after ICD implantation, when Al-Khatib et al undertook their study of Medicare patients who received an ICD between January 2002 and September 2005 (Circ: Arr&EP 2008;1:240-249). Examining 8,581 patients with ICDs, they found a correlation between an increase in the number of implants per physician and a significant decrease in the rate of complications and one-year mortality.

At ACC.10, James V. Freeman, MD, from Stanford University School of Medicine in California, and colleagues presented a similar study in a broader, more contemporary patient population. They examined the influence of hospital volume on outcomes, assessing 224,233 patients from 1,201 hospitals who received an initial ICD implantation between January 2006 and December 2008 in facilities participating in the ACC NCDR ICD Registry. They found that centers performing a higher volume of ICD implants have a lower rate of procedural complications.

They also assessed the different types of ICDs: single-chamber, dual-chamber and biventricular devices, finding that the trend of higher hospital implant volume and procedural success maintained, regardless of device type. 

Implanter type trumps volume

Questions still remain about how the individual operator—both personal volume and specialty—affects complication rates. “Data have shown that implanter training, rather than volume, has a greater influence on outcomes,” says Stephen C. Hammill, MD, from Mayo Clinic in Rochester, Minn.

Curtis et al found that implantations by non-electrophysiologists were associated with a higher risk of procedural complications and lower likelihood of receiving a cardiac resynchronization therapy-defibrillator device, compared with patients whose ICD was implanted by an electrophysiologist (JAMA 2009;301(16):1661-1670). ACC’s ICD Registry also revealed that non-electrophysiologists implanted 29 percent of ICDs during the 2006-2007 study period.

Likewise, Al-Khatib et al found the best results with electrophysiologists. They noted that patients who had their device implanted by a non-electrophysiology cardiologist were significantly more likely to have congestive heart failure than patients who had their ICD implanted by a thoracic surgeon; and patients who had their ICD implanted by a thoracic surgeon were more likely to have cerebrovascular disease.

In their accompanying editorial, Hammill and Curtis wrote that thoracic surgeons “will maintain that their patients are sicker in ways that administrative data simply cannot capture, and their concerns are not without merit” (Circ: Arr&EP 2008;1:235-237).

Assessing quality

CMS and other payors encourage the use of databases for reporting outcomes. For reimbursement, Medicare requires that providers implanting ICDs report data on each primary prevention ICD procedure. Currently, ACC’s ICD Registry provides the greatest reflection of U.S. performance outcomes with information on more than 380,000 implants from almost 1,500 hospitals, as of June 2009. 

While an argument can be made that only certain operators and providers should be implanting ICDs, Freeman stops short of this recommendation. “While we are now able to identify clear trends, we have yet to establish thresholds. Ultimately, the purpose of this data collection and research is to discover why adverse events occur, and how best to improve outcomes,” he notes.

However, physician consensus recommends that longer term outcomes are truly necessary to properly assess quality measures with ICD implantations. “The use of a registry that includes follow-up data will be the best approach for providers to assess outcomes,” says Hammill. He recommends that outcomes be assessed at 30 days for complications such as death, pneumothorax or pericardial effusion, and at 90 days to assess whether an infection occurred or a new defibrillator is needed.