We established a successful method to reduce transfusions of all blood products using strict adherence to evidence-based criteria and continuous monitoring. Our model translates into improved patient safety by decreasing the number of unnecessary transfusions. This also led to a significant reduction in hospital expenses.

No abstract

e20544 Background: Despite their poor prognosis, few cancer patients have advanced care directives (ACD). Most often, ACD is discussed in an inpatient setting, when the patient is acutely ill and often nearing the end of life. Timely and thoughtful discussion would be better accomplished for both the patient and the family before such hospitalization. Methods: Interviews were done in 108 outpatients in an ambulatory chemotherapy center. A questionnaire noted whether ACD was executed or not, demographics (including education, religion and insurance status), performance status, characteristics of their cancer, and family status. All patients without ACD were counseled on the importance of ACD. The presence of an ACD was documented at each visit. Statistical comparison of patients with and without ACD at each visit was done. Results: The following variables were significantly associated with having an ACD at the first visit. Patients living alone could not be included in the logistic regression model, as none had advanced directives. However ACD execution differed significantly in patients living alone, compared with those living with family (0/33 vs 15/75; p=0.005). An additional 42 patients executed an ACD by the third visit. At the third visit, only living with family vs living alone remained significantly associated with having ACD. [OR=3.17 95% CI: 1.34, 7.50; p=0.009]. Conclusions: The following findings are noteworthy. Only 15/108 (14%) cancer outpatients had ACD. This disturbingly low rate increased to 55/108 (51%) after physician- initiated discussion. The following characteristics predicted not having ACD before intervention: younger age, absence of metastasis, better performance status, lower education status and living alone. After several physician interventions, only living alone remained a significant barrier to ACD execution by cancer outpatients. These results provide new insights and strategies to encourage improved advanced directive execution. [Table: see text] No significant financial relationships to disclose.

1126 Background: The complications of blood transfusions and the increased morbidity and mortality associated with liberal use of blood products have been convincingly demonstrated. The clinical problems they pose have prompted development of evidence-based transfusion guidelines. Nevertheless, most hospitals do not enforce these guidelines and effective strategies have not been developed. Continuing medical education, such as conferences, do not appear to be effective in changing physician behavior, as confirmed in the literature and by our hospital's experience. New York Methodist Hospital therefore undertook a more proactive approach to reduce unnecessary transfusions and in the process to change physicians' attitude toward transfusions. Methods: In October 2008, our hospital implemented strict criteria for transfusions of all blood products including red blood cells (RBC), platelets, fresh frozen plasma, and cryoprecipitate. These criteria were applied by selected experienced clinicians who each served as monitors for their own clinical department. The blood bank required the approval of the monitor to release blood products that did not meet the criteria. This helped ensure that patients who indeed require transfusion based on clinical situation would be treated appropriately. Physicians had easy access to the monitors 24 hours a day, 365 days a year. Transfusion requests from operating rooms and post-anesthesia care units were excluded from monitoring. Transfusion of patients with active bleeding was also exempted because it was included in the RBC justification criteria. In most patients with symptomatic anemia, a single unit of RBC was approved. We compared data from the 12 months preceding (Year 1) and the 12 months after the enforcement of the criteria (Year 2). The data included the number of hospital admissions, the amount of blood products transfused per month, the type and number of transfusion-related complications, and the number of type and screens performed by the blood bank. Results: Transfusion usage declined sharply from Year 1 to Year 2: RBC use fell by 27.9% (9,149 vs. 6,599 units), fresh frozen plasma use fell by 40.3 % (3,615 vs. 2,158), platelets use fell by 22.1 % (1,207 vs. 940), and cryoprecipitate use fell by 37.4% (1,421 vs. 890). These decreases occurred in the face of a 2.9% increase in hospital admissions over 12 months (from 35,348 to 36,421). The decreased transfusion use was accompanied by reduced complications, which declined by 28.6 % (35 transfusion reactions in Year 1 vs. 25 in Year 2). The total cost of blood products declined by $1,027,869 (from $3,694,069 in Year 1 to $2,666,200 in Year 2). The number of type and screen tests requested and performed in the hospital also decreased by 6.3% (from 28,488 in Year 1 to 26,681 in Year 2). Conclusion: We report our one-year experience with a new model for improving transfusion use. Monitoring and intervention in described settings with possibilities of appeal reduced transfusion use by 22–40%. It also appeared to modify physician attitude as exemplified by a 6.3% reduction in type and screen requests in spite of 2.9% increase in admissions. Although small, this decline supports the trend toward fewer transfusions and suggests a change in physician culture and behavior. We recommend other hospitals consider similar versions of this strategy appropriate for their circumstances since it benefits patients and reduces health care costs. Disclosures: No relevant conflicts of interest to declare.