Background: Although the use of implantable cardioverter-defibrillators (ICDs) for the primary prevention of sudden cardiac death varies by sex, race, and hospital, geographic variation in ICD use remains unexplored. Our objective was to quantify regional variations in the utilization of primary prevention ICDs in the United States, and to evaluate if an association exists between utilization and physician supply or the proportion of patients meeting the trial inclusion criteria.
Methods and results: This is a cross-sectional analysis among the Medicare, fee-for-service population from the National Cardiovascular Data Registry. Using hospital referral regions, we calculated the age-, sex-, and race-adjusted rates of ICD placement for each region and assessed the correlation between these rates and (1) physician supply and (2) the proportion of patients meeting trial inclusion criteria. Substantial variation was found across quintiles of rate ratios of ICD implantation, ranging from 0.39 to 1.77 (compared with a national mean rate of 1.0). This ratio was not correlated with the supply of cardiologists (R(2)=0.01), electrophysiologists (R(2)=0.01), or with the proportion of patients meeting trial inclusion criteria (R(2)<0.01). Over all, 13% of all patients receiving ICDs did not meet trial criteria.
Conclusions: Marked geographic variation in the use of primary prevention ICDs exists across the United States that is not correlated with physician supply. Although >1 in 10 patients received ICDs outside of trial criteria, this potential overuse did not explain the variation. Future studies should consider underuse or misuse of primary prevention ICDs as causes of geographic variation.