Background: Guidelines recommend measuring glycated hemoglobin (HgbA1C) levels on all inpatients with diabetes mellitus, if untested in the prior 3 months. In response, our health system applies default HgbA1C orders in the electronic insulin orders for qualifying patients. However, recent red blood cell transfusions may cause falsely low HgbA1c values, leading to inappropriate management. Clinicians often forget to deselect default HgbA1c orders after transfusion, leading to erroneous results. We evaluated a non-interruptive clinical decision support (CDS) intervention to discourage automatic HgbA1C testing in patients with recent transfusions. Method: Retrospective observational analysis of the number and percentage of HgbA1c tests performed on inpatients within 7 days of a red blood cell transfusion over a 40-month period Results: Pre-intervention, clinicians ordered an average of 827 HgbA1c tests/month on inpatients. Of these, 11.7% (97 tests/month) were on patients who received a red blood cell transfusion within the preceding 7 days. Post-intervention, clinicians ordered an average of 832 HgbA1c tests per month on inpatients, of which 5.8% (48 tests/month) were performed within 7 days of red blood cell transfusions. Discussion: A non-interruptive CDS intervention can significantly decrease the number of HgbA1c tests performed in hospitalized patients who received a red blood cell transfusion, a common cause of erroneous HgbA1c values. This approach reduced waste without restricting clinician autonomy or requiring interruptive alerts that generate alert fatigue.
White et al. (Mon,) studied this question.