Examination of anticoagulation prescription among elderly patients with atrial fibrillation after in-hospital fall

Mechanical fall is common among elders and has been associated with a lack of anticoagulant therapy among patients with atrial fibrillation (AF). However, anticoagulant therapy is recommended despite frequent fall due to an increased risk of a thromboembolic event. Using data from a large health system, we investigated the predictors of anticoagulation prescription on discharge in AF elderly patients after an in-hospital fall. In this retrospective analysis, we examined patients aged 60 years and older discharged from 2013 to 2018 with a diagnosis of AF and a secondary diagnosis of in-hospital fall. The primary outcome was the prescription of anticoagulation at discharge. We obtained patients’ demographical (race, sex, and health insurance status) and clinical (management by a resident team, receipt of a head CT or a cardiology consultation, ambulation status and discharge location) data. We further categorized the type of anticoagulation prescribed as warfarin or novel oral anticoagulants (NOACs). We ran chi-square and Fischer’s exact tests on all data and multivariable logistic regressions on those of patients with pre-existing AF to identify the predictors of anticoagulation prescription on discharge. In total, 67% of 235 patients were discharged on anticoagulation. Of patients admitted on anticoagulation, 91% were prescribed anticoagulation on discharge (p < 0.001), while only 40% of patients with new-onset AF were discharged on anticoagulation (p < 0.001). Patients over the age of 90, compared to those aged 60–89, with existing AF had lower odds (OR = 0.34 [95% CI 0.12–0.98]) of being prescribed anticoagulation on discharge. Among patients with preexisting AF, being admitted on anticoagulation increased the odds (OR = 39.8 [15.2–104.0]) of anticoagulation prescription on discharge. Asian patients with prior AF were less likely (OR = 0.12 [0.026–0.060]) to receive anticoagulation on discharge. Of patients with new AF, 81% were prescribed a NOAC as opposed to warfarin (p < 0.05). These results suggest that provider’s decisions on anticoagulation initiation seem to be guided more by their concerns over bleeding complications than by the patient’s risk for stroke. However, anchoring bias strongly influences anticoagulation prescription. It may benefit AF patients already on anticoagulation, but it may prevent anticoagulation prescription in patients with new AF and Asian patients.