Use of Warfarin Therapy Among Residents Who Developed Venous Thromboembolism in the Nursing Home

Abstract

BackgroundTreatment of venous thromboembolism (VTE) in long-term care (LTC) settings has received little empirical study. ObjectiveAmong residents with VTE in nursing homes, this analysis evaluated frequency of anticoagulant use, the proportion of residents newly started on warfarin who persisted on therapy (≥3 months), and the association of key resident characteristics, including bleeding risk, with warfarin use and persistence. MethodsUsing the AnalytiCare LTC database (US), eligible residents had deep vein thrombosis or pulmonary embolism coded in the Minimum Data Set (MDS) 2.0 during the uptake period April 1, 2007 through December 31, 2008 (earliest VTE was index date) and had 1 or more MDS assessment(s) over the 90-day preindex period, each negative for VTE. Logistic regression evaluated the association of resident characteristics with warfarin use. Cox regression evaluated persistence with warfarin therapy. ResultsThe median age of residents with VTE included in the analysis (N = 489) was 80 years; 73% received anticoagulant therapy and 66% were prescribed warfarin ±45 days of the index date. Multivariate logistic regression identified several factors significantly associated with warfarin use: location in South Central region (odds ratio [OR] = 1.94, P = 0.019) and the Western region (OR = 2.53, P = 0.005) [both vs reference South Atlantic]; body mass index categories normal (OR = 2.73, P = 0.045), overweight (OR = 4.21, P = 0.005), and obese (OR = 3.82, P = 0.010) (both vs reference underweight); Alzheimer's/dementia (OR = 0.52, P = 0.024); cancer (OR = 0.39, P = 0.008); and moderate-dependent versus independent physical functioning (OR = 2.59, P = 0.003). Of residents newly started on warfarin therapy with no history of cancer (n = 149), 28% discontinued warfarin within 90 days of initiation. Peripheral vascular disease (PVD) (OR = 4.07, P < 0.001), Alzheimer's disease/dementia (OR = 2.55, P = 0.046), and antipsychotic use (OR = 4.60, P < 0.001) were all significantly associated with discontinuation. ConclusionsPatients in specific geographic regions who were underweight, had Alzheimer's disease/dementia or cancer, or had independent physical functioning were less likely to receive warfarin. Nonpersistence of warfarin therapy was strongly related to antipsychotic use, presence of dementia, or PVD.