Need for CDS to Prevent Too Rapid Injections and Infusions Due to CPOE?

Congress contribution

Affiliations
a Research Center for Medical Informatics, University Hospital Zurich, Switzerland

b Division of Internal Medicine, University Hospital Zurich, Switzerland

Publié le 15.10.2014

Introduction

Computerised physician order entry (CPOE) has been reported to improve patient safety. However, unintended effects can occur by introducing CPOE. Responsibilities may be shifted from nurses to physicians since CPOE may require physicians to provide information formerly defined by nurses, e.g. on drug dilutions and infusion rates. Consequences of too rapid intravenous administration may range from minor impairments of well-being to death. The purpose of this quality assessment was to determine potential hazards of the CPOE process for intravenous (IV) medication not assisted by clinical decision support (CDS). Orders for IV administered drugs were analysed retrospectively to identify those with too short durations in order to determine potential benefits of CDS-enhanced CPOE.

Methods

All IV drug prescriptions for inpatients at the University Hospital Zurich (tertiary care, 850 beds) were included over a 25 month period following the introduction of CPOE except for treatments in intensive care units, emergency and operating rooms. Infusion and injection times which were shorter than the minimal periods listed in the Swiss repository of approved drug information were defined as too rapid.

Results

A total of 459,099 IV orders resulted in 1,575,062 IV administrations to 55,682 patients. Analysis of the 150 most frequently ordered drugs showed that the duration of IV administrations was not defined in 113,397 of 248,915 orders (45.6%), prescribed with too short durations in 3,376 orders (1.4%) and documented as too rapid administrations in 14,206 of 779,304 injections or infusions (1.8%). Considerable differences in the proportion of orders with too short periods of IV administration were observed between drugs (table 1).
Table 1
Percentage of orders with too short duration of IV administration in a selection of drugs.
DrugMinimal durationaOrders:Prescribed administration durationb
Too rapid /Total (in %) Min. Mean Max. SD
Aciclovir60 min52 /811(6.4%) 6048024
Chlorpromazine50 min17 /486(3.5%) 9351,440577
Ciprofloxacin60 min460 /1,355(34.0%) 521,440106
Diclofenac15 min30 /2,199(1.4%) 2972046
Iloprost360 min666 /2,642(25.2%)153272,880194
Paracetamol15 min560 /24,985(2.2%) 17110,080413
Potassium chloride15 min7 /18,218(0.0%) 99110,080619
Vancomycin60 min108 /3,425(3.2%) 892,16083
a Minimal recommended duration according to the Swiss repository of approved drug information 

b Minimum, mean, maximum, and standard deviation in minutes

Discussion

The numbers of orders with unspecified or too short durations of IV infusions were high and patient safety might be a concern with some orders. However, nurses review orders before drug administration and thereby reduce the risk of incidents. Collaboration of physicians, nurses and pharmacists differ between countries and even within clinics. Introduction of CPOE may modify the collaboration of physicians and nurses and should therefore be monitored and managed as an iterative process of business improvements. Management policies have to define how detailed physicians have to order IV prescriptions. Nevertheless technical interventions could further minimise safety hazards by providing CDS including drug-specific default values for minimum infusion periods and specifying constraints for the input of IV administration durations.
Correspondence:

Emmanuel Eschmann

University Hospital Zurich

Directorate of Research and Education

Research Center for Medical Informatics

Rämistrasse 10

CH-8091 Zurich

emmanuel.eschmann[at]usz.ch

Avec la fonction commentaires, nous proposons un espace pour un échange professionnel ouvert et critique. Celui-ci est ouvert à tous les abonné-e-s SHW Beta. Nous publions les commentaires tant qu’ils respectent nos lignes directrices.