Shorter Length of Stay Is Associated With Worse Functional Outcomes for Medicare Beneficiaries With Stroke

• Published in: Physical therapy 2013-12, Vol.93 (12), p.1592-1602
• Main Authors: O'Brien, Suzanne R, Xue, Ying, Ingersoll, Gail, Kelly, Adam
• Format: Article
• Language: English
• Subjects: Activities of Daily Living; Aged; Aged, 80 and over; Beneficiaries; Care and treatment; Confidence intervals; Cross-Sectional Studies; Disability Evaluation; Economic aspects; Female; Health care policy; Hospitals; Humans; Length of Stay - statistics & numerical data; Logistic Models; Male; Medicare; Medicare - statistics & numerical data; Patient Discharge - statistics & numerical data; Patient outcomes; Prognosis; Prospective payment systems; Prospective payment systems (Medical care); Severity of Illness Index; Stroke; Stroke (Disease); Stroke Rehabilitation; Studies; Subacute care; United States
• ISSN: 0031-9023; 1538-6724
• DOI: 10.2522/ptj.20120484

Understanding of the potential impact that length of stay (LOS) may have on Medicare beneficiaries' poststroke discharge function and discharge destination since implementation of a prospective payment system is lacking. This study examined the trends and associations between LOS and discharge outcomes in Medicare beneficiaries with stroke treated in inpatient rehabilitation facilities (IRFs). A serial, cross-sectional analysis of the Inpatient Rehabilitation Facility Patient Assessment Instrument dataset was conducted. The sample consisted of 371,211 patients with stroke who were over 65 years of age in all IRFs in the United States between January 1, 2002, and June 30, 2007. Annual trends for means of LOS, admission and discharge Functional Independence Measure (FIM) scores, and percent community discharge were examined using generalized estimating equations (GEEs) with facility level control and post hoc testing. The association between discharge FIM scores and LOS was examined using a continuous, multivariate GEE model. The association between community discharge and LOS was examined using a logistic, multivariate GEE model. Time trends showed mean LOS decreased 1.8 days; admission and discharge FIM scores declined 4.4 points and 3.6 points, respectively; and mean community discharges declined 5.4%. Controlling for study year and covariates, each day was associated with an increase of 0.50 discharge FIM points (95% confidence interval=0.48, 0.52). Each day also was associated with a 0.3% decrease in odds of community discharge (95% confidence interval=0.994, 0.999). Reliability and validity of the Inpatient Rehabilitation Facility Patient Assessment Instrument (IRF-PAI) are lacking. Results may have been biased by a lack of control at the patient and facility levels. Medicare beneficiaries with stroke treated in IRFs experienced shorter LOS, had worsening admission and discharge function, and had fewer community discharges. Worsening admission function and shorter LOS may contribute to worsening discharge outcomes, which may indicate a lack of readiness for IRF treatment and that facility-level factors may be playing a role in shorter LOS.