The 30-day hospital readmission rate for stroke patients declined by an annual average of 3.3% from 2010 to 2014, national data on more than 2 million stroke hospital discharges indicated.
About one in eight adult stroke patients were readmitted within 30 days of hospital discharge in that period, but stroke patients discharged from high-volume non-teaching hospitals were more likely to return than patients treated at high-volume academic centers, reported Farhaan Vahidy, MD, MPH, PhD, of McGovern Medical School at UTHealth in Houston, and colleagues in JAMA Network Open.
“This work provides the most comprehensive evidence to date that 30-day readmission rates have declined during 2010 to 2014 in the U.S. for patients with different types of stroke — ischemic stroke, primary intracerebral hemorrhage, and subarachnoid hemorrhage,” Vahidy told MedPage Today.
Reducing ischemic stroke readmissions is a quality metric of the Centers for Medicare & Medicaid Services (CMS), but less is known about readmission rates of stroke patients who are covered by other insurers or uninsured, or people who have hemorrhagic strokes.
In the current study, Vahidy and colleagues analyzed 2,078,854 index stroke events in the Nationwide Readmissions Database of the Healthcare Cost and Utilization Project between January 1, 2010, and September 30, 2015, representing 50% of all U.S. hospitalizations in 22 states. Of these events, 87.6% were acute ischemic strokes, 8.7% were intracerebral hemorrhages, and 3.6% were subarachnoid hemorrhages. The researchers defined readmission as any admission within 30 days of index hospitalization discharge.
Mean patient age was about 70 and 51.9% were female. More than 35% of patients in the study were younger than 65, and more than 25% were not insured by Medicare or Medicaid.
Overall, 13.7% of intracerebral hemorrhage patients, 12.4% of acute ischemic stroke patients, and 11.5% of subarachnoid hemorrhage patients were readmitted within 30 days.
About one in five readmissions had the same primary discharge diagnosis as the index hospitalization; the second most common readmission reason was septicemia. Based on CMS-defined algorithms, more than 90% of readmissions were unplanned; depending on the stroke subtype, up to 13.6% were considered potentially preventable.
Among patients with acute ischemic stroke, 30-day readmission rates were significantly lower for patients younger than age 65 compared with older patients (10.8% versus 13.3%). There also were significant differences in readmission rates between payer types (OR 0.70 for private insurance versus Medicare), even after controlling for age, suggesting CMS readmission estimates may not be nationally representative.
While teaching hospitals had steady readmission rates on average regardless of how many stroke cases they treated, non-academic centers had higher readmission rates with increasing numbers of stroke patients.
“Intuitively, you would think that the higher volume of stroke patients a hospital treats, the better the quality of care is and the better their outcomes,” observed Adam Webb, MD, of Emory University Hospital in Atlanta, who was not involved in the study. “In fact, this is just the opposite for readmission in non-teaching hospitals.”
“The authors correctly point out that high-volume teaching hospitals are more likely to participate in stroke quality data registries such as Get With The Guidelines and may be more likely to be Joint Commission–accredited stroke centers,” Webb told MedPage Today.“This is yet another argument why stroke systems of care should be set up, whenever possible, to route patients to these types of hospitals.”
This research “provides a nationally representative benchmark to examine readmission rates for patients hospitalized with stroke,” added Elan Guterman, MD, of the University of California San Francisco, who also was not involved with the study.
“It also demonstrates clear variation in readmission rate across different institutional settings and over time,” Guterman said. “This is an informative place to start and poses many important questions about what is driving this variability.”
The analysis has several limitations, the researchers noted; for example, the database did not allow them to control for stroke-specific severity measures or track individual patients across multiple years.
The findings are best suited for improving local care rather than establishing a national performance standard, they added: “Further work is warranted to establish the validity of national readmission measures, particularly in the context of emerging evidence of compromised survival associated with readmission reduction for other conditions.”
The authors declared no competing interests.
