Effect of an Inpatient Anticoagulation Service on Improving the Safe Use of Warfarin Sodium

Background:

Anticoagulation with warfarin in the hospital has many challenges, including a narrow therapeutic range, variable metabolism, and drug‐related interactions. These factors, combined with illness and the complexities of hospitalization, have the potential for significant morbidity and mortality because of sub‐ and supratherapeutic effects related to warfarin dosing. The Joint Commission has made the reduction in the likelihood of patient harm associated with anticoagulation therapy a priority for its 2008 National Patient Safety Goals in Hospital Programs. There is evidence for improved therapeutic use of warfarin in outpatient anticoagulation clinics, but the effect on hospitalized patients has not been studied.

Methods:

With IRB approval, we enrolled patients admitted to our Medicine service from April through September 2007 who were either currently taking or being started on warfarin during their hospitalization. We randomly identified 2 medical teams as an intervention group and the other 2 teams as a control group in the study. We collected baseline patient demographics and indications for anticoagulation in both groups. We followed warfarin doses, international normalized ratios (INRs), adverse events associated with subtherapeutic (INR < 1.5) or supratherapeutic (INR > 5) events, and length of stay (LOS) daily in each group. In the intervention group, an anticoagulation nurse in coordination with pharmacist and physician support provided evidence‐based risk assessment and ongoing dosing guideline recommendations for warfarin. The control group received usual care, dosing, and monitoring of warfarin therapy. Group comparisons were made using chi‐square tests, Wilcoxon tests, and Cox's regression model.

Results:

Our study enrolled 87 patients (39 intervention and 48 control). There were no significant differences in sex, age, indication for anticoagulation, or adverse events between the 2 groups. Intervention patients were less likely to have subtherapeutic INRs compared with controls (37.4% vs. 53.4%, P = .0002). There were no differences in supratherapeutic INRs between groups (17.1% of control and 20% of intervention). The INRs of the intervention patients were significantly more likely to be in therapeutic range versus controls (42.6% vs. 29.3%, P = .0018). There was a borderline significantly shorter median LOS between intervention and controls (6 vs. 8 days, P = .07).

Conclusions:

In this study, the coordinated monitoring and dosing of warfarin maintained more patients in a safe and therapeutic range. Although we did not demonstrate a reduction in adverse events, there is evidence that closely links such events to sub‐ and supratherapeutic INRs. Based on our study, warfarin can be administered in a manner that avoids such situations. Furthermore, such efforts showed a trend toward reducing LOS. These important factors of safer therapeutics and the potential for reducing the LOS support further development of inpatient protocols and services to ensure the safe use of warfarin.

Author Disclosure:

E. Merrens, none; C. DiPaola, none; T. MacKenzie, none.