As the Number of US TAVR Centers Grows, Patient Outcomes Vary
Substantial variation exists across hospitals in mortality and readmission rates following TAVR, with patients treated at certain centers more than twice as likely to die within 30 days as those treated at the best-performing hospitals. The disparity, likely related to as-yet unexplored hospital-level factors, highlights the need to measure performance going forward and raises questions about the minimum number of procedures needed to maintain certification for a TAVR program, experts say.
“This study serves as an important benchmark for quality measurement and future performance improvement efforts for TAVR,” Harlan Krumholz, MD, SM, of Yale University (New Haven, CT), and colleagues report in the December 15 issue of the Journal of the American College of Cardiology.
“Moving forward, as more centers and operators begin performing TAVR and existing centers and operators become more proficient, it will be important to continue to monitor the extent of hospital variation to ensure the delivery of optimal outcomes for patients,” they say.
The findings do not, however, indicate that the rollout of TAVR should be pulled back in any way, according to co-author Karthik Murugiah, MD, also of Yale. He noted that the degree of scrutiny TAVR has undergone during its introduction is unprecedented.
“We don’t want our study to be misinterpreted as saying that we need to step back, but it does raise a key issue: that we need to be mindful and continue to measure performance as this procedure is performed in higher and higher numbers over time,” he said in an interview with TCTMD.
TAVR has been rapidly adopted since its approval in November 2011, and although reports from the early commercial experience have shown good results, little is known about hospital-level variation in outcomes.
To explore the issue, the investigators looked at Medicare fee-for-service data on patients 65 and older who underwent TAVR between 2011 and 2013. The analysis included 14,722 procedures performed at 417 hospitals; the median number of procedures per hospital was 17 (range 2 to 46).
Median risk-standardized outcome rates showed substantial variation across centers:
- 30-day mortality: 6.0% (range 3.8% to 10.2%)
- 1-year mortality: 17.5% (range 11.8% to 25.6%)
- 30-day readmission: 20.9% (range 17.1% to 24.4%)
After adjustment for a wide range of patient characteristics, the likelihood of 30-day mortality was about twice as high at centers 1 SD above vs 1 SD below the national average (OR 2.07; 95% CI 1.91-2.25). Associations were attenuated but still significant for 1-year mortality (OR 1.76; 95% CI 1.69-1.82) and 30-day readmission (OR 1.41; 95% CI 1.37-1.44).
Hospital-Level Factors Driving Variation
Murugiah said some of the observed between-center variation could be explained by patient-level factors that were not captured by their models, which were based on administrative data lacking certain clinical details. The rest of the variation, he said, is likely related to hospital-level variables that need to be evaluated in future studies, like institutional TAVR volume, operator training and volume, how cardiac surgeons and interventional cardiologists interact when treating patients, and procedure flow in the operating room.
Isaac George, MD, of Columbia University Medical Center (New York, NY), told TCTMD in an email that the variation in outcomes across hospitals is not surprising, “given the complexity of the procedure, the significant investment a hospital system must make to establish a valve center (OR, anesthesia, echo, ICU, hybrid room), and the experience factor associated with TAVR.”
The findings partly reflect the fact that centers that have been performing TAVR longer, do more cases, and have greater experience with handling complex anatomy and complications will have better outcomes, he said.
“An important aspect of mortality is the concept of failure to rescue, with the understanding that larger centers have a better capacity to recognize complications early and prevent them from spiraling into a mortality,” he said. That is especially true for TAVR, in which decisions made during the postoperative period, when many patients can die from comorbid conditions, will be critical in determining outcome, he added.
What Center Volume is Appropriate?
The analysis was limited, George said, by not reporting actual center volume and including hospitals performing as little as 1 TAVR procedure during the study period. “Clearly a center that has only done 1 procedure will skew these results tremendously,” he said. “Maybe a more meaningful number would have been 25, as it would include low-volume centers but not centers just starting.”
“The important question to ask moving forward is how many TAVRs should be designated as a minimum to retain TAVR certification for a program,” he continued. “That will depend on outcomes such as those reported here but also on more in-depth analyses that can be provided by the TVT Registry. I suspect that TAVR will continue to expand to more centers but reach a saturation point in another year, while surgical centers will contract and become more regionalized.”
Murugiah K, Wang Y, Desai NR, et al. Hospital variation in outcomes for transcatheter aortic valve replacement among Medicare beneficiaries, 2011 to 2013. J Am Coll Cardiol. 2015;66:2678-2685.
- The study was supported by the National Heart, Lung, and Blood Institute.
- Krumholz reports being the recipient of research agreements from Johnson & Johnson (Janssen) and Medtronic through Yale University; having a contract with the FDA to develop methods and facilitate best practices for medical device surveillance; and serving as chair of a cardiac scientific advisory board for UnitedHealth.
- George and Murugiah report no relevant conflicts of interest.