Background:
For many years, there has been awareness that excessive transfusions can cause harm to patients. Despite this, practice patterns have been slow to adapt to evidence suggesting restrictive transfusion policies result in better patient outcomes. Our organization attempted to reduce transfusions across the institution using a passive decision support intervention with no measurable impact. We hypothesized that a focused, active intervention would be more successful.
Purpose:
Our goal was to increase adherence to a restrictive transfusion guideline for patients admitted to general medicine floors. Our guideline suggested a transfusion threshold of Hgb<7 in upper GI bleed patients except in cases of hemodynamic instability. In all other indications for transfusion our threshold was Hgb<8.
Description:
These guidelines were developed from existing evidence and consultation with departmental content experts. In March 2013 we launched our intervention during a monthly Patient Safety Conference attended by housestaff and hospitalists. This conference was attended by departmental and residency leadership. The guidelines were delivered and the evidence behind them was explained with the only educational focus being to reduce patient harm; reduction of hospital cost was not discussed. This was followed with monthly email alerts to the house officers for the following three months. Chief residents reviewed the guidelines with residents on rotation switch days. Hospital medicine faculty reviewed the guidelines at section meetings and via email alerts.
Conclusions:
Data was collected over representative time periods before and after the intervention. We restricted our measurements to transfusions given on our three general medical units. We measured the absolute number of units of PRBCs delivered as well as the pre‐transfusion Hgb associated with each unit transfused. We also measured the number of admissions to these units over the sample periods. When comparing the three month samples pre and post intervention, units transfused dropped from 481 to 273. When corrected for volume of admissions, the rate of units transfused per admission fell from 0.319 to 0.167 (47.6%). The median pre‐transfusion Hgb fell from 7.6 to 7.3 suggesting that of the units transfused, more were within guidelines. In the pre‐intervention sample, there were 95 units of blood delivered with a pre‐transfusion Hgb>8. In the post‐intervention group, this fell to 18 units. Overall, the effects of this focused intervention yielded significant reduction in the use of blood products. These findings suggest that a local, intensive patient safety intervention will be more effective at achieving change than a “top down” hospital wide intervention.