Computerized Physician Order Entry: a focus on medication prescription
Nicolette de Keizer & Saeid EslamiDept Medical InformaticsUniversity of Amsterdam
Outline
Definition and context Advantages of CPOE Disadvantages of CPOE Outcome measures and examples Same system other outcome
What is Computerized Physician Order Entry (CPOE)?
Ordering of tests, medications, and treatments for patient care using computers
Involves electronic communication of the orders
Often use rules-based methods for checking appropriateness of care
CPOE, EHR and DSS
EHR
Documentation
Medication
Test reports (EKG, PFT)Radiology, lab results
CPOE
DSS
CPOE Advantages
Automate ordering process Reduces Order Errors
Standardized, legible complete orders
DSS/Alerts Data collected on variances in
practice
Case Example
Metformin is prescribed to a patient with an elevated creatinine level. A drug-lab interaction alert warns that
use of this medication could result in an increased risk of fatal lactic acidosis.
Case Example
A physician prescribes warfarin for a patient with chronic atrial fibrillation.
System: vitamin K rich food likely to interfere with
the efficacy of the drug. Specific patient information
Example DSS in CPOE – medication prescription
Allergy Age (check drug name and dose) Duplicate drugs on active orders,
not one-time Severe drug interactions
Drug-drug, drug-food Dose maximum Drugs with opposite actions
PubMed indexed papers on CPOE
0
20
40
60
80
100
120
140
160
<2000 2000 2001 2002 2003 2004 2005 2006
no. of papers
CPOE
Has a positive influence on patients’ outcome
Has a negative influence on patients’ outcome
On which outcome measure?
Outcome measures
Adherence to guideline Alerts - user response Time Safety
Medication errors ADEs (mortality)
Cost and Efficiency Medication costs Pharmacists interventions
Satisfaction, usage and usability
Example CPOE improves adherence to guideline
Teich JM et al. Arch Intern Med. 2000 Oct 9;160(18):2713-4.
Example CPOE reduce errors
Potts studied ADE rates in 13,828 medication orders before/after CPOE implementation at Vanderbilt Children’s PICU:
Potts AL, Barr FE, et al. Pediatrics. 2004 Jan;113(1 Pt 1):59-63.
• Brigham and Womens' Hospital, Boston introduced a CPOE
• After implementation, the rate of intercepted Adverse Drug Events (ADE) doubled!
• Reason: The system allowed to easily order much too large dosages of potassium chloride without clear indicating that it be given in divided doses.
• Bates et al The impact of computerized physician order entry on medication error prevention. JAMIA 1999, 6(4), 313-21.
Example CPOE introduces errors
pre period1 period2 period3
Potential ADEs/1000 pt-days
15.8 31.3 59.4 0.5
Example CPOE introduces errors
Association with increased PICU mortality: 2.8% 14 months before CPOE 6.4% 5 months after CPOE
Han YY, Carcillo JA, et al. Pediatrics. 2005 Dec;116(6):1506-12.
Example CPOE reduce costs
Cost: $3.7 million
implementation $ 600,000 to $1.1
million operational costs
Results: Decreased drug costs ADE cost is
approximately $4,700
Brigham and Women’s Experience: Cost-Effective
Kausal R et al. J Am Med Inform Assoc. 2006; 13(3): 365-7
CPOE and cost
Huge variation in actual costs based on hospital size and complexity of system Hardware and Software: $1-$5 million Staff training Ongoing maintenance Total costs for large, fully integrated
systems could be up to $60 million Costs will decrease when DSS is
geared to cost reduction
Usability Problems
Potential selection errors Similar medication names Similar patient names Overly trust default values
Influence workflow and communication Physician resistance
Frequent data entry required Must not require additional time Most decision-support steps must be turned off
to encourage use “Not-invented-here” syndrome
• Health information systems has to deal with the actors, the artefacts, and their interaction.
CPOE as a Sociotechnical intervention
CPOE =socio-technical systems
Same system other outcome
Upperman vs. Han same hospital: Pittsburgh Childrens
Hospital other outcome measure: ADE vs mortality
Han vs. Del Beccaro Same CPOE Different hospitals Same outcome measure (mortality)
Upperman et al. J Pediar Surg. 2005;40:57–59; Han et al. Pediatrics. 2005 Dec;116(6):1506-12; Del Beccaro. Pediatrics 2006;118;290-295.
Upperman et alPre-CPOE Post-CPOE
Verbal order regulatory compliance
80% 95%*
All ADEs 0.3 0.37
Harmfull ADEs 0.05 0.03*
* p<0.05
Del Becarro et al
Pre-CPOE Post-CPOE
Mortality 4.22% 3.46%
Same results for:
• transported patients
• congenital cardiovascular disease patients
Upperman vs Han
Surrogate outcome measures such as medication errors and ADEs are not sufficient
More studies on mortality are necessary
Han vs. Del Beccaro
Methodological differences: Han used unbalanced before after
periods Del Beccaro did not correct for case
mix differences in before vs after period Populations Han vs. Del Beccaro were
not comparable
Han vs. Del Beccaro
Pittsburgh implemented too fast without pilot and usability testing: Order entry only allowed when patient
physically entered hospital All medication in central pharmacy Pharmacy processed order only after
complete activation by nurse Go live without necessary order sets
Han vs. Del Beccaro
Seattle hospital learned from Pittsburgh Factors influencing successful
implementation: response time and user time meeting information needs (e.g.using order
sets) multidimensional integration with workflow Involvement leaders, support staff, clinicians training improvement through evaluation and learning
Successful CPOE implementation
Technical Infrastructure: EHR Drug information database DSS
Sociotechnical implementation strategy Usability pilots Teams of informaticians, physicians, nurses, clerks,
therapists Strong leader and good project management Training and support
R. Gardner: “implementation is 90% organizational and political and 10% technical”