2012 HSR&D/QUERI National Conference Abstract
3049 — Quality of Care in VA Community Living Centers: 1997 to 2007
Kim HM, Mach J, Szymanski B, and McCarthy JF, Ann Arbor COE;
There are longstanding concerns regarding quality of care in nursing homes, particularly for individuals with psychiatric morbidity. This study assesses process-related quality for Veterans Affairs Community Living Centers (CLCs, formerly known as nursing homes) and evaluates differences with regard to the prevalence of serious mental illness (SMI) among residents.
Using VA administrative and Minimum Data Set records, we identified residents in 145 VA CLCs, for fiscal years 1997 (FY97) through FY07 and assessed trends in and predictors of four established process-related quality indicators. Quality indicators included: (1) treatment with only first-generation antipsychotics among residents with schizophrenia receiving oral antipsychotics; (2) depression without antidepressant therapy in first 6 months of stay; (3) occasional bladder/bowel incontinence without a toileting plan; and (4) physical restraint use among residents with dementia. We also assessed factors associated with variability in these indicators, including indicators based on the prevalence of residents with serious mental illness (SMI) and SMI admission volume.
All indicators showed significant improvement from FY97 to FY02 and continued with moderate improvement even beyond FY02. Specifically, from FY02 to FY07 there were declines in untreated depression (from 6.3% to 5.5%), restraint use (4.2% to 1.1%), and incontinence without a toileting plan (37.3% to 22.1%). However, among residents with schizophrenia the percentage treated only with first generation antipsychotics rose from 6.4% in FY97 to 10.8% in FY02, remaining stable thereafter. After FY05, by which time most indicators showed high quality and were stable, higher CLC prevalence of SMI was associated with lower percentage treated only with first generation antipsychotics (p = 0.04), and also with greater restraint use (p <0.001) and more incontinence without a toileting plan (p = 0.002). Facility staffing was neither associated with change in process indicators during early period (FY97 to FY02) nor with variability in these indicators during the stable period.
Between FY97 and FY07, VA CLCs had significant improvement on important quality indicators. SMI prevalence was inconsistently associated with CLC quality indicators.
This study documents important quality advances in VA CLCs. Further attention is needed regarding facility-level differences in treatment patterns and quality of care.