Abstract 253: Out of Range INR Results Lead to Increased Health Care Utilization in Four Large Anticoagulation Clinics

Abstract

Background: The impact on healthcare utilization of a single out of range (OOR) INR value not associated with any bleeding or thromboembolic complication among chronic warfarin-treated patients is not well described. Methods: At four large phone-based anticoagulation clinics in Michigan, warfarin-treated patients with atrial fibrillation (AF) or venous thromboembolism (VTE) were identified and data collected via medical chart abstraction. Propensity score matching was used to identify two groups closely matched on patient characteristics: the OOR INR group (INR value < 1.8 or > 3.2 if target range was 2-3) and the control group (INR value between 1.8 and 3.2). Data for each anticoagulation clinic interaction and INR lab test were abstracted until the patient had 2 subsequent and sequential in-range INR values. Methods and frequency of interactions between the anticoagulation clinic and patient were recorded, described as median and interquartile ranges (IQR) and compared using Poisson regression with adjusted means. Results: Demographics were similar for the 116 OOR INR patients and the 58 control patients studied (mean±SD age 72.1±13.0 and 74.3±11.3 years, respectively). Indications for warfarin were more commonly venous thromboembolism in the OOR versus in-range patients (42.2% vs. 27.6%, p=0.06). OOR and in-range INR patients experienced a median of 3 (IQR 3-5) and 3 (IQR 3-3) with adjusted means of 4.2 and 3.2 (p<0.001) INR lab draws until two sequential tests were in-range. OOR INR patients required a median 5 (IQR 3-6) with adjusted means of 5.3 and 3.7 interactions with the anticoagulation clinic versus 3 (IQR 3-4) for in-range INR patients (p<0.001). OOR INR patients more often required phone calls (adjusted means 2.9 versus 0.9, p<0.001) but fewer mailed letters (adjusted means 1.3 versus 2.3, p<0.001) than in-range INR patients. OOR INR patients more often required multiple types of contact than in-range INR patients (83.6% versus 55.2%, p<0.001). Contact was most frequently performed by registered nurses (adjusted means 3.4 versus 2.9, p=0.059) and administrative assistants (adjusted means 1.3 versus 0.5, p<0.001) for both OOR and in-range INR patients. Blood count and renal function lab test were similarly rare for both groups (mean 0.4, median 0, p=0.74 for each). Conclusions: Warfarin-treated patients who experience OOR INR values without any bleeding or thromboembolic complication require more frequent interactions with the anticoagulation clinic, including more telephone calls and multiple types of contact.