To determine predictors of oral anticoagulation (OAC) for atrial fibrillation (AF) in long-term care (LTC).

Chart review.

Six LTC facilities in a metropolitan area.

One hundred seventeen residents with AF identified from 934 total residents.

Data was obtained from the medical chart, pharmacy record, and Minimum Data Set (MDS) regarding demographics, medical conditions, falls, fractures, gastrointestinal bleeding (GIB), peptic ulcer disease, dementia, anemia, and physical/cognitive function scales. The recursive partition algorithm was used to construct a model reflecting physician decision patterns that predict prescription of OAC.

Among those 117 residents (12.5% of 934) who had AF (age, 84.6 ± 8 years), OAC was prescribed for 54 (46%); aspirin or clopidogrel: 47 (40%); neither OAC nor any antithrombotic treatment (ATT): 25 (21%). Prior stroke was the primary determinant of OAC. Residents with prior stroke were less likely to be prescribed OAC if they had prior GIB, were non-Caucasian, or had no history of coronary artery disease (CAD). Those without a stroke were less likely to be prescribed OAC if they were younger, had dementia or lower functional status.

Prior stroke was the primary predictor of OAC use. Our model suggests that physicians may also incorporate concerns of age, bleeding, cognitive and physical function, and ethnicity into the decision-making process. Further study is needed to explore the reasons why 21% of the residents receive neither OAC nor ATT, and why OAC may be less likely to be prescribed to non-Caucasian LTC residents.