The Addition of Decision Support into Computerized
Physician Order Entry Reduces Red Blood Cell
Transfusion Resource Utilization
in the Intensive Care Unit
Evans R. Ferna ´ndez Pe ´rez,1*Jeffrey L. Winters,2and Ognjen Gajic1
1Mayo Clinic College of Medicine, Division of Pulmonary and Critical Care Medicine, Mayo Clinic, Rochester, Minnesota
2Mayo Clinic College of Medicine, Division of Transfusion Medicine, Department of Laboratory Medicine and Pathology,
Mayo Clinic, Rochester, Minnesota
Computerized physician order entry (CPOE) has the potential for cost containment in crit-
ically ill patients through practice standardization and elimination of unnecessary interven-
tions. Previous study demonstrated the beneficial short-term effect of adding a decision
support for red blood cell (RBC) transfusion into the hospital CPOE. We evaluated the effect
of such intervention on RBC resource utilization during the two-year study period. From the
institutional APACHE III database we identified 2,200 patients with anemia, but no active
bleeding on admission: 1,100 during a year before and 1,100 during a year after the interven-
tion. The mean number of RBC transfusions per patient decreased from 1.5 ± 1.9 units to
1.3 ± 1.8 units after the intervention (P ¼ 0.045). RBC transfusion cost decreased from
$616,442 to $556,226 after the intervention. Hospital length of stay and adjusted hospital
mortality did not differ before and after protocol implementation. In conclusion, the imple-
mentation of an evidenced-based decision support system through a CPOE can decrease
RBC transfusion resource utilization in critically ill patients. Am. J. Hematol. 82:631–633,
C 2007 Wiley-Liss, Inc.
Key words: erythrocyte transfusion; medical order entry systems; cost savings
CPOE can decrease the costs of overuse and mis-
use of health services and improve compliance with
clinical guidelines . In a retrospective cohort study
 the addition of decision support for RBC trans-
fusion within the hospital CPOE effectively reduced
the number of ICU patients receiving transfusions
outside evidence-based indications. However, the
short follow-up period of the study (3 months) did
not allow adequate assessment of the durability and
the financial impact of such intervention. The current
study seeks to determine whether the implementation
of a CPOE decision support system for critically ill
patients could decrease the rate and cost for RBC
MATERIALS AND METHODS
The study was approved by the Institutional
Review Board. We studied consecutive critically ill
patients with anemia (Hct < 30%) from three multi-
disciplinary ICUs (medical, surgical and mixed) dur-
ing the periods before (May 2003 through April 2004)
and after (January 2005 through December 2005) the
introduction of decision support for RBC transfusion.
Patients with admitting diagnosis of bleeding, those
who underwent massive transfusion (>10 units) and
excluded. The decision support was incorporated into
the existing transfusion electronic orders on 15
*Correspondence to: Dr. Evans R. Ferna ´ ndez Pe ´ rez, Mayo
Clinic, 200 First Street SW, E18, Rochester, MN 55905.
Received for publication 4 August 2006; Accepted 20 November
Published online 18 January 2007 in Wiley InterScience
American Journal of Hematology 82:631–633 (2007)
C 2007 Wiley-Liss, Inc.
November 2004. The decision support algorithm justi-
fied RBC transfusion for patients whose Hgb were >7
g/dL in the presence of active bleeding, ischemia or
early septic shock.
APACHE III, transfusion and laboratory databases,
data was collected on: age, gender, ICU location,
Acute Physiology and Chronic Health Evaluation
(APACHE) III score and predicted mortality, number
of RBC transfusions and hematocrit. The primary
outcome variables were the number of RBC transfu-
sions per ICU patient and cost of RBC transfusions.
The ICU total cost per unit of leukoreduced-RBC
transfused was estimated at $168.53. To this was
added the costs incurred by the hospital for cross-
match testing ($52.43) and administering blood in the
ICU ($165.00). Since staffing was not altered to imple-
ment the CPOE decision support, it was not necessary
to attribute staffing costs to the intervention group.
Technical cost associated with integration of the deci-
sion support within the existing institutional elec-
tronic information infrastructure and validating and
testing the implementation of the CPOE decision sup-
port was estimated to be $600.00. This protocol-
driven cost was added to intervention period.
We studied 2,220 patients over a 24-months period
(1,110 in the control group and 1,110 in the inter-
vention group). The mean number of RBC transfu-
sions (1.5 ± 1.9 vs 1.3 ± 1.8, P ¼ 0.045) and the
percentage of transfused patients decreased (53% vs
48%, P ¼ 0.013) decreased after protocol implemen-
tation. There were no significant differences in age,
gender, hematocrit, ICU, and hospital length of stay
(LOS) between the two groups (Table I). Apache III
scores, predicted hospital mortality and actual hos-
pital mortality were higher during the second period
(Table I). RBC transfusion cost decreased over the
study period from $616442 to $556226. When ac-
counted for the cost of the addition of decision sup-
port, the calculated cost saving was $59,616 during
the first year of implementation in the three ICUs.
We found no differences in the mean rate of RBC
transfusion between the first and last two months
of protocol implementation (1.2 ± 1.8 vs. 1.3 ± 1.8,
P ¼ 0.217).
In this study, the implementation of an evidenced-
based decision support system through a CPOE
in critical care units lead to a small but significant
decrease in RBC transfusion resource utilization.
The effect remained consistent throughout the post-
intervention period. Although transfusion cost in the
ICU tends to rise considerably in the presence of
patients with higher probability of death, [4,5] in
this study we observed that a CPOE decision sup-
port prevented the increase of RBC transfusion
resource utilization despite the increase in severity
of illness during the intervention period. It is impor-
tant to emphasize that the cost associated with RBC
transfusion represent only a small part among over-
all cost of care for critically ill patients. In this
study, costs common to both control and interven-
tion periods such as ICU and hospital LOS were
not significantly different. With any quality improve-
ment intervention, one must consider the cost of the
intervention itself and the learning curve for pro-
viders [6,7]. The incremental cost of adding the deci-
sion support for RBC transfusion into the existing
institutional CPOE was indeed trivial ($600). In
conclusion, the use of blood transfusions for treat-
ment of anemia in critically ill patients has signifi-
cant financial implications. The addition of a deci-
sion support to the CPOE can enhance the restric-
TABLE I. Patient Characteristics and Outcome of ICU Patients With Anemia One Year Before and
After Decision Support Implementation
Patient characteristics Before decision support (N ¼ 1,100)After decision support (N ¼ 1,100)
Age in years
APACHE III score
APACHE III ICU predicted mortality
ICU length of stay (days)
Hospital length of stay (days)
For continuous variables, median values and interquartile ranges are reported in parenthesis while for categorical variables counts with percentages in
parenthesis are reported.
CPOE, computerized physician order entry; APACHE, acute physiology and chronic health evaluation.
aPre-transfusion hematocrit or the lowest hematocrit value during the ICU stay in patients who did not get blood transfusion.
632Brief Report: Ferna ´ndez Pe ´rez et al.
American Journal of Hematology DOI 10.1002/ajh
tive transfusion practice in the ICU, generating cost Download full-text
savings from reduced RBC transfusion.
1. Kuperman GJ, Gibson RF. Computer physician order entry:
Benefits, costs, and issues. Ann Intern Med 2003;139:31–39.
2. Rana R, Afessa B, Keegan MT, et al. Evidence-based red cell
transfusion in the critically ill: Quality improvement using
computerized physician order entry. Crit Care Med 2006;34:
3. Rivers EP, Nguyen HB, Huang DT, Donnino M. Early goal-directed
therapy. Crit Care Med 2004;32:314–315. Author reply 315.
4. Levy MM, Abraham E, Zilberberg M, MacIntyre NR. A descrip-
tive evaluation of transfusion practices in patients receiving me-
chanical ventilation. Chest 2005;127:928–935.
5. Hebert PC, Wells G, Martin C, et al. Variation in red cell trans-
fusion practice in the intensive care unit: A multicentre cohort
study. Crit Care (Lond) 1999;3:57–63.
6. Henderson RA, Pocock SJ, Sharp SJ, et al. Long-term results of
RITA-1 trial: Clinical and cost comparisons of coronary angio-
plasty and coronary-artery bypass grafting. Randomised Inter-
vention Treatment of Angina. Lancet 1998;352:1419–1425.
7. Drummond MF, Sculpher MJ, Torrance GW, O’Brien BJ, Stod-
dart GL. Cost analysis. In: Methods for the Economic Evalua-
tion of Health Care Programmes, 3rd ed. Oxford: Oxford Uni-
versity Press, 2005. pp 55–101.
633Brief Report: RBC Transfusion Resource Utilization
American Journal of Hematology DOI 10.1002/ajh