Hospital-Acquired Symptomatic Urinary Tract Infection in Patients Admitted to an Academic Stroke Center Affects Discharge Disposition

Objective: To test the role of hospital-acquired symptomatic urinary tract infection (SUTI) as an independent predictor of discharge disposition in the acute stroke patient. Study Design: A retrospective study of data collected from a stroke registry service. The registry is maintained by the Specialized Programs of Translational Research in Acute Stroke Data Core. The Specialized Programs of Translational Research in Acute Stroke is a nationalnetwork of 8 centers that perform early phase clinical projects, share data, and promote new approaches to therapy for acute stroke. Setting: A single university-based hospital. Participants: We performed a data query of the fields of interest from our universitybased stroke registry, a collection of 200 variables collected prospectively for each patient admitted to the stroke service between July 2004 and October 2009, with discharge disposition of home, inpatient rehabilitation, skilled nursing facility, or long-term acute care. Main Outcome Measures: Baseline demographics, including age, gender, ethnicity, and National Institutes of Health Stroke Scale (NIHSS) score, were collected. Cerebrovascular disease risk factors were used for independent risk assessment. Interaction terms were created between SUTI and known covariates, such as age, NIHSS, serum creatinine level, history of stroke, and urinary incontinence. Because patients who share discharge disposition tend to have similar length of hospitalization, we analyzed the effect of SUTI on the median length of stay for a correlation. Days in the intensive care unit and death were used to evaluate morbidity and mortality. By using multivariate logistic regression, the data were analyzed for differences in poststroke disposition among patients with SUTI. Results: Of 4971 patients admitted to the University of Texas at Houston Stroke Service, 2089 were discharged to home, 1029 to inpatient rehabilitation, 659 to a skilled nursing facility, and 226 to a long-term acute care facility. Patients with an SUTI were57%less likely to be discharged home compared with the other levels of care (P < .0001; odds ratio 0.430 [95% confidence interval 0.303-0.609]). When considering inpatient rehabilitation versus skilled nursing facility, patients with SUTI were 38% less likely to be discharged to inpatient rehabilitation (P<.0058; odds ratio 0.626 [95% confidence interval, 0.449-0.873]). We performed interaction analyses for SUTI and age, NIHSS, urinary incontinence, serum creatinine level, and history of stroke. We noted an interaction between SUTI and NIHSS for discharge disposition to a skilled nursing facility versus a long-term acute care facility. For patients with SUTI, a 1-unit increase in NIHSS results in a 10.6% increase in the likelihood of stroke rehabilitation in a long-term acute care facility compared with 5.6% increased likelihood for patients without SUTI (P = .0370). Conclusions: Acute stroke patients with hospital-acquired SUTI are less likely to be discharged home. In our analysis, if poststroke care is necessary, then patients with SUTI are more likely to receive inpatient stroke rehabilitation at the level of care suggestive of lower functional status. For every point increase in NIHSS, stroke patients with SUTI are 10.6% more likely to require continued rehabilitation care in a long-term acute care facility versus a skilled nursing facility compared with 5.6% for patients without SUTI. The combination of premorbid urinary incontinence and urinary tract infection has no additional impact on discharge disposition. This study is limited by its retrospective nature and the undetermined role of psychosocial factors related to discharge. Prospective studies are warranted on the efficacy of early catheter discontinuation, identification of new-onset urinary incontinence, use of genitourinary barriers, and catheter care every shift as variables that can decrease the risk of infection. The information obtained from prospective studies will have an impact on resource use that is of prime importance in the current health care climate.