Preventing Wrong-Drug and Wrong-Patient Errors With Indication Alerts in CPOE Systems

Despite the widespread use of computerized prescriber order entry (CPOE) systems, wrong-drug and wrong-patient errors persist. An accurate problem list helps to prevent these errors by alerting prescribers when a medication order does not have a corresponding problem on a patient’s list. Problem lists are not well maintained and are often inaccurate and incomplete, jeopardizing patient safety. The use of indication alerts that prompt prescribers to add problems to a problem list when it is missing a corresponding problem provides situational awareness to prescribers and promotes self-interception of wrong-drug and wrong-patient errors.

Two types of interception events can be measured electronically: 1) wrong-patient abandon and reorder - when a prescriber begins a medication order, abandons it, and reorders for the correct medication or patient – and 2) wrong-drug retract and reorder - when a prescriber cancels an order soon after signing it, and then reorders for the correct medication or patient. This research examined how the use of indication alerts impacted the placement of problems on the problem list and impacted interception events.

The specific aims of the research were as follows:

• Implement a set of 30-50 indication alerts for medications that are vulnerable to look-alike and sound-alike errors at one hospital in Chicago and one in New York City, using two commercial electronic medical record systems.
• Conduct an interrupted time series study design to quantify the effect of indication alerts on 1) the combined rate of self-intercepted wrong-drug and wrong-patient CPOE errors and 2) on the rate of each type of error viewed separately.
• Assess the impact of indication alerts on the probability of adding new diagnoses to the problem list during encounters that include CPOE.

A before-and-after interrupted time series study was conducted to quantify the impact of indication alerts on five outcomes: wrong patient retract and reorder (WP-RAR), wrong drug retract and reorder (WD-RAR), wrong patient abandon and reorder (WP-AAR), wrong drug abandon and reorder (WD-AAR), and problems placed on the problem list. Two healthcare organizations participated: New York Presbyterian (NYP) with 86 study medications and Northwestern Medicine (NM) with 206 study medications. Indication alerts occurred when a study drug was prescribed that did not have a corresponding problem on the patient’s problem list. At NYP the alerts were non-disruptive and appeared only during inpatient orders that were not part of order sets. NM’s were interruptive alerts that targeted only attending physicians at some facilities.

Indication alerts resulted in a significant increase in placing problems on the problem list. The intervention had no effect on WP-RAR events. There was no immediate effect on WD-RAR events, but they did decline slowly over time after the intervention. WP-AAR and WD-AAR had a 77 percent and 33 percent relative increase, respectively. These changes were attributed to an increase in situational awareness by the prescriber and represent evidence of a patient safety improvement.

The researchers advocate that health systems consider implementing indication alerts while weighing the costs – implementation and an increase in alert fatigue – against the benefits – more complete problem lists and an increase in self-intercepted medication errors.