A controlled trial of educational outreach to improve blood transfusion practice

To determine whether brief, face-to-face educational outreach visits can improve the appropriateness of blood product utilization.

Randomized, controlled multicenter trial with 6-month follow-up.

Surgical and medical services of two pairs of matched community and teaching hospitals in Massachusetts.

One hundred one transfusing staff surgeons and attending medical physicians.

A professionally based transfusion specialist presented one surgical- or medical-service-wide lecture emphasizing appropriate indications, risks, and benefits of red blood cell transfusions; brief, graphic, printed educational guidelines; and one 30-minute visit with each transfusing physician. No data feedback was provided. Educational messages emphasized the lack of utility of the traditional threshold for red blood cell transfusions (hematocrit, 30%) and transfusion risks (eg, viral hepatitis).

Proportion of red blood cell transfusions classified as compliant or noncompliant with blood transfusion guidelines, or indeterminate 6 months before and 6 months after an experimental educational intervention.

Based on analyses of 1449 medical record audits of red blood cell transfusions that occurred 6 months before and 6 months after the educational intervention, the average proportion of transfusions not in compliance with criteria declined from 0.40 to 0.24 among study surgeons (-40%) compared with an increase from 0.40 to 0.44 (+9%) among control surgeons (P = .006). These effects were consistent across procedure type and specialty. On average, study surgeons in the postintervention period performed transfusions when hematocrits were 2.0 percentage points lower than before the intervention (28.3% preintervention vs 26.3% postintervention), and lower than in the control group (28.3% preintervention and postintervention; P = .04). Likely savings in blood use for surgical services probably exceeded program costs, even without considering reduced risks of infection. No effects were observed among transfusions occurring in medical services, possibly because of substantially lower transfusion rates and lower pretransfusion hematocrits.

Brief, focused educational outreach visits by transfusion specialists can substantially improve the appropriateness and cost-effectiveness of blood product use in surgery. More data are needed regarding the durability of changes in practice patterns and the health and economic benefits of such interventions.