|
BACKGROUND/RATIONALE:
Little is currently known about the prevalence, characteristics, or quality of care for VA nursing home residents with serious mental illness (SMI). Researchers have questioned whether nursing home staff have adequate mental health resources, training, and experience to meet the care needs of residents with SMI. There is also little information regarding the distribution of individuals with SMI across nursing homes, or how this relates to treatment capacities, processes, or resident outcomes.
OBJECTIVE(S):
This study will test whether and how nursing home structural elements (including SMI specialization) and facility treatment processes influence risks of adverse outcomes, for patients with and without SMI.
Study objectives are to: 1) Describe the prevalence, characteristics, and services use of VA nursing home residents with SMI; 2) Determine whether risks of adverse health outcomes differ by resident SMI status; and 3) Determine whether individuals with SMI are less likely to experience adverse health outcomes if their nursing home has a high prevalence of SMI.
METHODS:
Retrospective cohort design, analyzing national VA administrative data from FY01-FY06. We will also survey VA nursing home Unit Nurse Managers.
The study will identify residents in VA nursing homes at the end of each year, FY01-FY06. We will assess resident characteristics and services utilization, by SMI status. We will evaluate trends in the prevalence of SMI. We will examine characteristics at admission and will evaluate associations between facility characteristics (e.g., size, location, prevalence of SMI) and resident characteristics.
We will identify all VA nursing home residents as of the start of FY01 and admissions before the end of FY06. We will conduct separate multivariable survival analyses for each outcome measure, except for mortality, which will be examined using generalized estimating equations. Analyses will evaluate whether resident SMI status is associated with risks of adverse outcomes. In models using continuous time variables (e.g., time until discharge), we will use Cox proportional hazards models. For models with discrete time variables (e.g., time until assessed functional status decline, where assessment are typically conducted at 3-month intervals), we will use discrete time hazards models.
We will examine the influence of structure and process indicators on risks of adverse outcomes. These will assessed at of the start of FY01 and then at 6-month intervals. Structure measures will include facility SMI specialization, assessed using a measure of sustained high prevalence and admission volume. We will use multivariable hazard rate analysis to assess the influence of structure and process measures on risks of negative health outcomes and GEE to model risks of mortality. In sensitivity analyses, we will include survey data, to validate key measures and inform interpretation.
FINDINGS/RESULTS:
In the period from FY98 through FY07, there were 321,412 veterans who received care in VA nursing homes. Among these individuals, 93,972 were included in the VA's National Psychosis Registry (29.2%); 144,143 individuals were included in the VA's National Registry for Depression (44.9%); and 143,010 (44.5%) had received a dementia diagnosis from VA providers at some point between FY97 and FY07. 56.3% of these veterans were included in either the National Psychosis Registry or the National Registry for Depression.
Analyses of VA nursing home resident cohorts are ongoing.
IMPACT:
Study investigators have contributed study information in the context of the VISN 11 Geriatrics and Extended Care Service Line Advisory Committee discussions and the VISN 11 Nursing Home Mental Health Task Force.
PUBLICATIONS:
None at this time.
DRA:
Aging and Age-Related Changes, Mental Illness
DRE:
Quality of Care
Keywords:
Behavior (patient), Behavior (provider), Severe mental illness
MeSH Terms:
none
|
