http://stroke.ahajournals.org/content/48/Suppl_1/ATMP40
Abstract
Background:
Interpersonal relationships are understudied components of the stroke
treatment paradigm, which become important when patients require
long-term care. In this study, we analyzed sociodemographic factors that
impact return to home after inpatient rehabilitation (IR).
Methods:
Stroke patients were identified by ICD9/10 code from a prospective
multicenter rehabilitation registry between Jan 2005 & July 2016
(n=6447). Patients were analyzed based upon "Home" vs "Not Home" or
"Married" vs "Not Married" groups. Descriptive statistics were provided
for all patients. Marital status was used as a proxy for caregiver
support. We hypothesized that increased discharge functional
independence measure (FIM), ambulation and no insurance predicted return
to home. A “return home model” was developed using multivariable
regression with a stepwise approach. Odds ratio & 95% CI were
calculated.
Results: 5378 patients
returned Home, 1069 did not return Home. Home patients tended to be
younger, married, ambulatory and minorities, with a discharge FIM>75
(p<0.0001). Aphasia, dysphagia and UTI were significantly higher in
the “Not Home” group (p<0.0001). Married patients had more stroke
risk factors and impairments, indicating increased caregiver needs
(Figure). In the model, being a minority and being a woman increased the
odds of returning home. Advancing age, being widowed, divorced,
separated or never married decreased the odds of returning home. We
confirmed that ambulation, increasing discharge FIM and no insurance
predicted return to home (Figure).
Conclusions:
Being married, a woman or a minority increases the odds of returning
home after inpatient rehabilitation. Caregiver training and social
support for unmarried and male patients are important areas of
improvement. Strategies to ensure the successful transition of stroke
rehabilitation patients to home are needed, including prospective
studies of non-spousal caregiver support.
No comments:
Post a Comment