UMHS LOGOUniversity of Michigan
Department of Pediatrics

Evidence-Based Pediatrics Web Site

Computerized Order Entry on Inpatient Services Reduces Adverse Drug Events


  • Significant expense in hardware, education, and time is involved in implementing computerized order entry systems.  Does the improvement in patient care and cost savings justify this expense?

Clinical Bottom Lines

  1. Adverse Drug Events (ADEs) occurred at a varying rate of 10.7/1000 patient days.
  2. ADEs were greatest at the order entry and administration phase.
  3. Most ADEs were dosing errors and involved analgesics or antibiotics.
  4. The rate of ADEs can be reduced by approximately 55% (RRR) with simple physician order entry systems (POES).  The reduction in non-intercepted serious medication errors was 5.84 per 1000 patient days (ARR).  Thus, you would need to use a POES for 171 patient days to prevent one error of this type.
  5. The initial cost (dollars and time) to implement such a system is significant.  However, the potential for such a system to save time and dollars and prevent adverse outcomes is enormous.

Summary of Key Evidence

  1. Six groups consisting of medical, surgical general, and intensive care units pre-intervention, the same groups post intervention and two additional units to increase study size were used.3
  2. The study was done in Boston at a hospital comparable to U of M, a 726 bed tertiary care hospital.  At Brigham and Women's Hospital, Boston MA, approximately 16,000 orders are written daily, 40% of them for medications.
  3. Adverse drug events were monitored for 6 months prior and 9 months after the POE intervention.
  4. End points were number of non-intercepted serious medication errors (preventable) and non-intercepted potential ADEs.
  5. POE intervention:
    • menu-based list of medications on formulary,
    • default doses and a range of potential doses for each medication, requirements to enter dosage, route and frequency for all orders,
    • relevant laboratory results were displayed on the screen at the time of ordering (e.g., potassium levels when furosemide was ordered),
    • consequent orders, which are orders that should follow from other orders (e.g., suggestions to perform animoglycoside levels when aminoglycosides were ordered),
    • limited drug allergy checking, checking for the most frequent drug allergies,
    • drug-drug interaction checking, 80 of the most serious drug-drug interactions,
    • drug laboratory checking (e.g., potassium levels in patients receiving potassium),
  6. Non-intercepted serious medication errors decreased 55% (RRR) from 10.7 events to 4.86 events per 1000 patient days (p=0.01).  Preventable ADEs declined 17% from 4.59 to 3.88 (p=.37) while non-intercepted potential ADEs declined 84% from 5.99 to 0.98 per 1000 pt days (p=.002).

Additional Comments

  • The study relied upon time-series comparisons. Because of the scope of the implementation of the order entry system, the implementation could not be done in a simultaneous randomized fashion.
  • Time between interventions may have resulted in differences in physicians, patient population, practices, medications used. Most notably study had to correct for error rate and ADEs associated with greater use of multiple sedatives in intervention group. Multiple sedative use accounted for 42% of preventable ADEs in the intervention group.
  • Costs of ADEs can be extraordinary.3
  • Serious errors cannot be prevented if important information is not entered into the computer. Seven serious allergy errors were not caught because new allergies discovered during the hospitalization were not entered.
  • Effect of computer system likely is lower limit of what can be accomplished because a great many potential functions of the POES were not implemented.
  • Adverse drug events account for 19% of injuries in hospitalized patients.1
  • The rate of adverse drug events occurs in a range 0.7 to 6.5 per hundred admissions, depending on the definition of the degree of adverse event. The most frequent errors are at the ordering stage.3
  • A computerized decision support tool has been used for selection of antibiotic and anti-infective substances in critically ill patients linked to the computer-based patient records and microbiological results systems. The program presented detailed epidemiological recommendations and warnings regarding drug use and suggested proper antibiotic selection to the ordering physician. There were a number of valuable outcomes including a reduction in antibiotic use, number of doses and cultures ordered, a reduction in the length of stay in the intensive care unit and in the total hospital stay from 4.9 to 2.7 days and 12.9 to 19, respectively. Total costs were reduced from $38,283 to $26,315 per hospital stay. Mortality was decreased by 4%.2
  • The University of Michigan Medical Center uses a critical care information system that predicts and analyzes patient outcomes. In its 20-bed medical ICU unit, savings in direct variable costs were more than $3 million per year, coming largely from reduced nursing staffing and pharmacy costs. Average length of stay in the ICU declined 40% during the ensuing two and a half years, from 6.9 days to 4.1 days. They reduced its percentage of low-risk ICU admissions from 17% to fewer than 10%.6


  1. Leape LL, Brennan TA, Laird NM.  The nature of adverse events in hospitalized patients.  N Engl J Med 1991; 324:377-384.
  2. Evans RS, Pestotnik SL, Classen DC, et al.  A computer-assisted management program for antibiotic and other anti-infective agents.  N Eng J Med 1998; 232-238.
  3. Bates DW, et al.  Effect of computer prescribing on preventing medication errors.  JAMA, 1998; 280(15):1311-1316.
  4. Shiffman RN, Brandt CA, Freeman BG.  Transition to a computer-based record using scannable, structured encounter forms.  Arch Ped Adolesc Med, 1997; 151(12):1247-1253.
  5. Bates DW, Cullen DJ, et al.  Incidence of adverse drug events and potential adverse drug events, implications for prevention.  JAMA, 1995; 29-34.
  6. Dr. Charles Watts, University of Michigan, Personal Communication.
  7. Tierney WM, Miller ME, Overhage JM, McDonald CJ.  Physician impatient order writing on microcomputer workstations.  JAMA, 1993; 269:379-383.

CAT Author: J. Michael Kramer, MD

CAT Appraisers: Allan Olson, MD

Date appraised: February 8, 1999

Last updated March 23, 2003
Department of Pediatrics and Communicable Diseases
© 1998-2002 University of Michigan Health System