HRS: Mind the gap: ICD variation remains unexplained
BOSTON—While innovations within the healthcare industry remain electric, there is still unexplained variations in care patterns in the U.S., particularly for device implants. Cardiologist density and patient insurance type could be some of the culprits, Paul Heidenreich, MD, of Stanford University, Palo Alto, Calif., said, adding that more work needs to be done to figure out the reasons for these differentiations.

Heidenreich discussed the impact of insurance and geographic variation in the U.S. during a presentation May 9 at the 33rd annual scientific sessions of the Heart Rhythm Society (HRS).

Heidenreich cited a 2011 study by Kapoor et al that evaluated quality measures and implantable cardioverter-defibrillator (ICD) use to look at insurance status and how it was related to ICD implantation. “Those without insurance were significantly less likely to have an ICD or have planned for one implanted compared with those with insurance or Medicare,” he said. In fact, there was a 40 percent decrease in the odds ratio reported in patients who did not hold insurance.

He also referenced a similar study that evaluated off-label cardiac resynchronization therapy (CRT) in 45,000 patients enrolled in the American College of Cardiology's National Cardiovascular Data Registry (NCDR). It was estimated that 24 percent of patients had off-label CRT use. Those who had Medicaid saw significantly less CRT-D use. Off-label CRT-D use was increased in patients who held commercial insurance.

“The summary here is fairly straightforward: patients with no insurance were definitely less likely to receive an ICD,” Heidenreich added. “We should take insurance status into account.”

The Dartmouth Atlas project has outlined the regional variations of care for many procedures, including coronary angiography and coronary artery stenting, among others. “In states, there is a lot of variation between county to county. It is very hard to pinpoint the cause. You can’t just say, ‘Oh, it’s the east coast or the northeast’ that does or does not do these procedures,” he said.

Regional variation is wide, especially in terms of ICD use. In fact, Heidenreich said that 2006-2007 Medicare data from the NCDR ICD Registry showed that there was a four- to five-fold difference between the lowest quartile and highest quartile hospital referral region. “If you break this down by region it ranges from 22 to 736 [per million population]."

Heidenreich questioned whether the supply of cardiologists pushes the demand for ICD implantation. However, he said that only 1 percent of this variation can be explained by the density of cardiologists.

He added that population density was associated with very dense areas with more ICD use. Additionally, he noted that care at academic hospitals did not show a direct linear relationship to ICD use but noted that those in the upper quartiles were more likely to be academic centers.

“Just to have the capability to place an ICD at your hospitals was not associated with higher ICD use,” he said.

Referral regions that increased ICD use noted a 0.12 percent reduction in mortality. It was concluded that if every region of the country could reach the top quintile of ICD use it would result in 1,345 fewer deaths. “More patients would be alive at two years ... due to this increased ICD use," he said.

“We don’t really know why there is so much variation or why there is so much variation in off-label use,” he added. “There is a slightly larger effect of academic hospitals and some related to income but there is a lot that still needs to be explained.”

Heidenreich estimated that nearly 20 percent of the racial disparities seen in ICD implantation could be explained by regional variation.

“Overall, the conclusion is that most regional variation is unexplained,” Heidenreich summed. “If we are going to optimize ICD use we are going to need a better understanding of these factors.”