HSR&D Home » Research » IIR 11-330 – HSR&D Study
VA Post-Stroke Rehabilitation: Comparing Institutional Long-Term Care Settings
Huanguang Jia, PhD MPH BA
North Florida/South Georgia Veterans Health System, Gainesville, FL
Funding Period: July 2013 - December 2015
Stroke is prevalent among VHA Veteran enrollees. Effective post-acute stroke rehabilitation and/or restorative care can speed stroke survivors' recovery, minimize their functional disabilities, and improve their independence in daily activities of living. VA Community Living Centers (CLCs: formerly known as VA Nursing Homes) and VA-contracted Community Nursing Homes (CNHs) are two major sources of institutional long-term care for Veterans diagnosed with stroke. No study, however, has systematically examined rehabilitation and restorative care as well as health-related outcomes of these Veterans residing in CLCs and/or CNHs.
The primary objective of this study is to compare rehabilitation and restorative care utilization and the health-related outcomes between Veterans residing in CLCs vs. their counterparts residing in CNHs. The three primary aims are to compare national and regional differences between CLCs and CNHs in (1) facility and Veterans' characteristics; (2) Veterans' adjusted utilization of rehabilitation therapy, restorative nursing care, and mental health management; and (3) Veterans' changing status in adjusted physical function, cognitive performance, mental health, as well as mortality and stroke rehospitalization, 12-month post nursing home admission.
This retrospective national study included all VHA Veteran enrollees diagnosed with stroke and admitted to CLCs or VA-contracted CNHs from January 1, 2006 through December 31, 2009. Eligible Veterans were followed for a maximum of 12 months after their nursing home admission date. Two primary databases used in the study were the VA Minimum Data Set (MDS) 2.0 for CLC information and the Centers for Medicare and Medicaid Services (CMS) MDS 2.0 for CNH information. Veterans' service utilization and functional outcomes were collected as part of the scheduled MDS assessments at CNHs and CLCs, respectively. MDS data were linked with VA national inpatient, outpatient and vital status data, VA Fee Basis, CMS Medicare inpatient (MedPAR) and outpatient data files. These linked files provided Veterans' sociodemographic and clinical information, as well as mortality and rehospitalization outcome information. (1) Facility and Veteran characteristics were compared between CLCs and CNHs nationally and by the five MyVA regions (North Atlantic, Southeast, Midwest, Continental, and Pacific). (2) Utilization referred to average weekly days for the following services received by Veterans during their CLC or CNH stays: Rehabilitation therapy (i.e., physical, occupational, speech, and respiratory therapy); restorative nursing care (i.e., active range of motion, walking, passive range of motion, bed transfer, dressing and/or grooming, eating and/or swallowing, bed mobility, and communication); and depression management (i.e., psychological therapy, antidepressant, antipsychotic and anti-anxiety medication). A two-part model (logistic regression for Part 1 probability model and general linear regression for Part 2 utilization model) was applied for each utilization measures. (3) Veterans' Activities of Daily Living (ADL) function, cognitive performance and depression status referred to the sum score difference between the baseline and 3-, 6-, 9-, and 12-months. The ADL scale was comprised of seven MDS 2.0 ADL self-performance items (i.e., bed mobility, transfer, locomotion, dressing, eating, toilet use, personal hygiene); the cognitive scale included four MDS 2.0 items (i.e., short-memory problem, daily decision making, making self-understood, and ADL in eating); and the depression status was measured by a set of seven MDS 2.0 depression/mood items (i.e., negative statements, persistent anger or irritability, expression of unrealistic fears, repetitive health complaints, repetitive anxious complaints, sad or pained facial expressions, and crying or tearfulness). Linear mixed model was applied for these three functional outcomes. Veterans' 12-month mortality was assessed using logistic regression analysis. Finally, 12-month rehospitalization referred the Veterans' rehospitalization for acute stroke occurred inside and/or outside the VHA system, and Poisson regression model was applied. The above utilization and health outcome analyses were risk-adjusted, case-mixed, and compared between CLC Veterans vs. CNH Veterans nationally and across the five MyVA regions. Significance level was p<0.05.
Among the 18,272 eligible Veterans, 69.3% were from 133 CLCs and 30.7% from 2,346 CNHs. Regional distributions of Veterans ranged from 3,531 for Midwest CLCs to 1,833 for Continental CLCs; and from 1,514 for North Atlantic CNHs to 903 for Continental CNHs. Compared with CNHs, CLCs were significantly more likely to be in urban areas, hospital-based, with fewer facility beds, and lower resident/bed ratios. Compared with CNH Veterans, CLC veterans were significantly more likely to be younger, male, less educated, African American, separated or divorced, urban residents, high VA healthcare priority, and with more comorbid conditions.
CLC (vs. CNH) Veterans had significantly higher user rates for restorative care (33.5% vs. 30.6%) but lower user rates for depression management (68.3% vs. 70.6%); and CLC (vs. CNH) Veterans had fewer average weekly days for rehabilitation therapy (4.9 3.5 vs. 6.4 4.4) and depression management (6.9 4.0 vs. 7.6 4.1) but more average days for restorative care (9.4 9.3 vs. 5.9 6.2). Part 1 model results showed that the odds for CLC (vs. CNH) Veterans in receiving rehabilitation therapy, restorative nursing care and depression management were 1.16, 2.28, and 0.98, respectively. Part 2 model analyses demonstrated that CLC (vs. CNH) users averaged 1.5 fewer days in rehabilitation therapy, 5.5 more days in restorative nursing care, and 0.8 fewer days in depression management. Significant regional variations in rehabilitation and depression management days were observed with more pronounced variation in restorative care. Specifically, compared with their counterparts in the Pacific region, Veterans in other regions had an average of 0.8 (for North Atlantic) to 4.8 (for Southeast) more days in restorative care.
Lower ADL, cognitive and depression summary scores represent better or improved patient functionality.
First, CNH (vs. CLC) Veterans' ADL change scores averaged significantly lower at 9 months (-0.4 6.4 vs. 0.0 7.1) and 12 months (-0.2 6.6 vs. 0.3 7.1), but higher at 3 months (-1.6 5.3 vs. -1.0 4.8). Longitudinal analysis results showed that CLC (vs. CNH) Veterans had significantly more ADL improvement (Coefficient SD: -0.71 0.18) during their nursing home stays. And, major CLC vs. CNH differences occurred in Continental (-1.39), Southeast (-0.77), and Midwest (-0.77) regions.
Second, CLC (vs. CNH) Veterans' cognitive change scores averaged significantly lower at 3 months (-0.1 1.5 vs. 0.0 1.7) and 12 months (0.2 2.2 vs. 0.4 2.3). Longitudinal analysis results showed that CLC (vs. CNH) Veterans had significantly better cognitive performances (Coefficient SD: -0.13 0.06) during their nursing home stays. And major CLC vs. CNH differences occurred in Continental (-0.23), Midwest (-0.21), Southeast (-0.16), and Atlantic (-0.12) regions.
Third, CLC (vs. CNH) Veterans' depression change scores averaged significantly lower at 3 months (0.0 1.0 vs. 0.1 1.3) with no significant differences being observed for the rest of the follow-up time periods. Longitudinal analysis results showed that CLC (vs. CNH) Veterans had significant improvement in depression (Coefficient SD: -0.21 0.04) during their nursing home stays. And major CLC vs. CNH differences occurred in Midwest (-0.36), North Atlantic (-0.31), Continental (-0.30), and Southeast (-0.23) regions.
Fourth, CLC (vs. CNH) Veterans had a significantly higher 12-month crude mortality rate (25.1% vs. 23.4%) and shorter average survival length (308.7 days vs. 322.4 days). Our logistic regression results showed that CLC (vs. CNH) Veterans were significantly more likely to die (Odds Ratio=1.13) within 12 months post-admission after adjusting for propensity score and service utilization. Significant regional variation in mortality was found. Among CNH Veterans, the chances of dying in North Atlantic and Continental regions were 17% and 19% less than in the Pacific. And among the CLC Veterans, the chances of dying in the Continental region were 33% more than that in the Pacific.
Fifth, during the 12-month follow-up time period, 18.8% of the entire study sample was rehospitalized for acute stroke care; the stroke rehospitalization rate was higher for CNH (vs. CLC) Veterans (27.4% vs. 15.0%) and 90.3% of CNH Veterans' rehospitalizations occurred under the Medicare program as compare to 57.5% for CLC Veterans. Poisson regression results showed that CLC (vs. CNH) Veterans were less likely to be rehospitalized for stroke within the 12-month follow-up time period (Coefficient SD: -0.65 0.08). Significant regional variation in rehospitalization was found. Among CNH Veterans, North Atlantic, Midwest, and Continental regions had 0.51, 0.37 and 0.57 more rehospitalizations than the Pacific. And among the CLC Veterans, the North Atlantic and Southeast had -0.35, -0.26 and 0.44 less rehospitalizations than the Pacific region.
Finally, our factor analysis and Rasch analysis showed the following results: the seven ADL items measured one construct; Rasch analysis results were consistent with findings from the linear mixed model and an interval ADL scale could assist clinicians better understand patient functionality.
CLCs and CNHs are two major sources of institutional long-term care for Veterans diagnosed with stroke. However, systematic studies of these Veterans' health care use and the care quality they receive at CLC and VA-contracted CNHs have been lacking. This study filled in the gaps by providing the evidence on CLC-CNH differences nationally and regionally in facility and patient characteristics, rehabilitation service utilization, functional health status, as well as related health outcomes. These end-point data can inform VHA policy makers in planning for outsourcing institutional long-term care services needed by VHA enrollees. These findings can also inform both VA clinicians' and Veterans' decisions regarding appropriate placement in rehabilitation facilities for post-stroke care.
External Links for this Project
NIH ReporterGrant Number: I01HX000803-01A2
Dimensions for VADimensions for VA is a web-based tool available to VA staff that enables detailed searches of published research and research projects.
If you have VA-Intranet access, click here for more information vaww.hsrd.research.va.gov/dimensions/
VA staff not currently on the VA network can access Dimensions by registering for an account using their VA email address. Search Dimensions for this project
DRA: Aging, Older Veterans' Health and Care, Cardiovascular Disease
DRE: Treatment - Observational
MeSH Terms: none