Overall 30-Day Stroke Readmissions Decline, but Not All Trends Rosy

Patrice Wendling

September 05, 2018

A new nationwide study shows that overall 30-day hospital readmissions after stroke are declining. However, readmissions due to a second stroke or complication closely related to the primary event are not going down and, unfortunately in some cases, are on the rise.

"To me that's a point of grave concern and highlights that there is a continued need to focus on secondary stroke prevention," said senior investigator, Farhaan Vahidy, PhD, MBBS, MPH, associate professor in neurology, University of Texas Health Science Center, Houston.

Although prior studies have focused on 30-day readmissions for Centers for Medicare & Medicaid Services (CMS) beneficiaries with acute ischemic stroke (AIS), little is known about nationwide readmissions after other stroke subtypes and among younger patients, he explained to theheart.org | Medscape Cardiology.

To examine these trends, investigators identified 2,200,688 stroke hospital discharges between January 2010 and September 2015 in the Nationwide Readmissions Database. Of these, 87.6% were AIS, 8.7% were intracerebral hemorrhage (ICH), and 3.6% were subarachnoid hemorrhage (SAH).

Among the 2,078,854 index events, the average patient age was 70 years and 51.9% of patients were female. An average of 1397 hospitals per year were included in the analyses, and the median stroke volume was 99 cases per year.

The overall 30-day stroke-related readmission rate was highest in 2010 and gradually declined by an annual 3.3% between 2010 and 2014, the last year full data were available.

After controlling for case mix across the years, there was a significant annual decline in the likelihood of 30-day readmission by 4.0% during the study period (odds ratio, 0.96; 95% confidence interval, 0.95 - 0.97).

Multivariate analyses also indicated that the probability of a nationwide 30-day stroke-related readmission was 12.1% less likely after implementation of the CMS's Readmissions Reduction Program (HRRP) penalties that began October 1, 2012, than before implementation.

Unplanned and Costly

More than 90% of all stroke-related readmissions were unplanned, and up to 13.6% were deemed potentially preventable. The top two reasons for readmission were acute cerebrovascular disease and septicemia.

Thirty-day readmission rates were highest for patients with ICH (13.7%), followed by patients with AIS (12.4%) and those with SAH (11.48%).

"Hemorrhagic patients tend to be younger, so it's kind of a 'double whammy' for them because they experience really high levels of early mortality as well," — averaging about 25% to 30% during hospitalization vs about 5% to 10% for a patient with an ischemic stroke, said Vahidy.

"Because hemorrhagic strokes are more severe, more debilitating, and produce profound impairments in activities of daily living, we did expect they might have a greater burden for readmissions but that's the unique finding of our study," he said. "And because they are younger, a majority, up to 50% of them, are not covered by Medicare at the time of their initial event. So those findings are new and novel."

Importantly, 30-day readmission attributed to the same primary diagnosis as the index admission increased from 16.20% to 19.25% between 2010 and 2015 for patients with AIS, while holding relatively steady over the same period for those with ICH (from 8.36% to 8.81%) and SAH (from 7.22% to 8.15%).

For patients with AIS, in-hospital mortality was higher on readmission than during the index hospitalization (6.54% vs 5.13%), the average length of stay was longer on readmission (6.5 days vs 4.92 days), and the average inflation-adjusted cost per stay was higher on readmission ($12,303 vs $10,881).

"We need to consider that these are population-level data, and we are talking about more than 2 million patients, of whom more than 12% are readmitted," Vahidy said. "So if the average length of stay is longer by any magnitude than what their initial stay was and their cost to the healthcare system is higher than what their initial cost was, then I think it becomes quite significant on an absolute level in terms of healthcare cost savings."

Local readmission reduction programs that use  telemedicine-based follow-up, nurse- or healthcare worker-based follow-up including secondary prevention medication use, and early visits to primary care or specialty clinics have curtailed readmissions over time, but there is no national program for such efforts, he said.

"From a healthcare perspective, sometimes we're skeptical about the cost-benefit of these programs, so I think there's a need to prove that at the national level these programs work and do help make a difference."

Hospital Volume

Because of concerns that quality programs, including the HRRP, may disproportionately penalize larger safety-net hospitals that tend to treat more complicated, sicker patients, the investigators also analyzed stroke readmissions by hospital type and annual stroke discharge volume.

In adjusted analyses,  the likelihood of 30-day readmission did not differ between academic institutions and nonteaching hospitals with a low stroke volume (11 to 50 cases per year). However, the likelihood of readmission was significantly higher for patients discharged from nonteaching hospitals with an annual stroke volume of 300 patients or more per year, compared with those discharged from teaching hospitals with a similar stroke volume.

"So the extrapolation of that is that maybe the quality of care that is being provided at teaching hospitals or the capacity of teaching hospitals is better at handling large volumes of stroke patients," Vahidy said. "I think that's also a unique and novel analysis that at least for stroke patients had not been done before."

Limitations of the study, the authors note, include the inability to control for stroke-specific severity measures; the inability to track individual patients across multiple years (and thus the analyses represent "cross-sectional estimates" for each year); and variations in ICD-9 coding practices for stroke across various settings or over time, which may introduce bias.

"Although nationally derived, our estimates are best suited for understanding and improving care at a local level, rather than establishing a national performance standard," the authors concluded in the study, published in JAMA Network Open on August 17.

Vahidy and first author Arvind Bambhroliya report having no relevant financial relationships.

JAMA Network Open. 2018;1:e181190. Full text

Follow Patrice Wendling on Twitter: @pwendl. For more from theheart.org | Medscape Cardiology, follow us on Twitter and Facebook.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.