Nearly two-thirds of Medicare beneficiaries who had an ischemic stroke either died or required rehospitalization within 1 year in an analysis of a registry of more than 90,000 patients.

In-hospital mortality was 14% within 30 days of hospital admission, and 1-year mortality exceeded 26%, Dr. Gregg C. Fonarow of Ahmanson-UCLA Cardiomyopathy Center, Los Angeles, and his associates reported Dec. 16 in the journal Stroke.

Although there was wide variation in the rates of death or rehospitalization between the bottom 10th percentile of hospitals and the top 10th percentile, hospital characteristics accounted for only a small percentage of the variance in clinical outcomes.

Until now, there have been few studies assessing the burden of rehospitalization and death following ischemic stroke. Most previous studies were conducted outside the United States and were limited by "restricted cohorts, typically from a single or small number of institutions; [a] limited geographic area; select clinical trial patient populations; [and] short follow-up," the investigators said. Most of these studies were also focused on a limited number of outcomes, they noted.

Dr. Fonarow and his colleagues used data from the Get With the Guidelines–Stroke registry to quantify rates of readmission and mortality. They examined the records of 91,134 Medicare patients treated at 625 participating hospitals, representing all geographic regions in the United States. Patients were admitted for acute ischemic stroke during 2003-2006 and followed through the end of 2007.

The registry included a variety of types of treatment centers, from small community hospitals to academic institutions and facilities officially designated as primary stroke centers.

The mean patient age was 79 years, and most patients had comorbidities such as hypertension (77%), diabetes (28%), coronary artery disease (33%), and atrial fibrillation or flutter (24%).

The combined end point of rehospitalization or death was quite frequent after ischemic stroke, with rates of 21% at 30 days and 62% at 1 year.

There was marked variation among hospitals in rates of death and readmission, which was obvious within 30 days and persisted throughout follow-up. "This suggests there may be significant variation in care processes and systems of care in-hospital, during transition of care, and during the post-discharge management," the investigators wrote.

There was an approximately twofold difference in outcomes between the highest-performing and lowest-performing hospitals. After adjustment for risk factors, the 30-day mortality was 18% at hospitals in the bottom 10th percentile, compared with 10% at hospitals in the top 10th percentile. The corresponding 30-day rates of death or readmission were 26% and 16%.

However, hospital characteristics accounted for only 2%-5% of the overall variation in clinical outcomes. Neither the volume of stroke patients nor teaching status affected hospital performance. Outcomes were not significantly different between hospitals that were designated as primary stroke centers and hospitals that were not. Geographic location exerted a small effect, with hospitals in the West and Northeast performing slightly better than those in the South or Midwest.

One "disappointing" finding was that outcomes did not improve over time. The mortality and readmission rates in 2006 were nearly identical to the rates in 2003, Dr. Fonarow and his associates wrote.

The primary limitation of this study was that stroke severity was measured and documented in only 37% of patients, despite its being "an important component of basic patient assessment and determination of prognosis." Use of the National Institutes of Health Stroke Scale or other severity scales is not routine in clinical practice, they noted.

Without data on stroke severity, it was not possible to identify the reasons for poorer outcomes in some hospitals and better outcomes in others, or to identify which aspects of patient care might be the best targets for quality improvement interventions.

Moreover, "as evident by patients’ comorbid conditions, ambulatory status, and disposition at discharge, many of these patients may have been quite impaired, and a portion of postdischarge events may not have been preventable, even with improved transitional and outpatient care," Dr. Fonarow and his colleagues wrote.

But one finding pointed to an area where improvement is needed: Many patients arrived at the hospital by private transportation. This highlights the need for further patient and public education on the importance of recognizing stroke symptoms early and immediately calling for emergency medical assistance, they noted.

Several investigators cited financial ties to manufacturers of drugs for ischemic stroke.

Poor Outcomes Reported for Many Medicare Stroke Patients