Improving the usability of the computerized provider order entry (CPOE) interface and integrating it with workflow may reduce the patient-safety risk of inconsistent communication, according to a prospective study published in the May 25 edition of the Archives of Internal Medicine.
Although several types of CPOE-related errors may occur, errors related to inconsistent information within the same prescription (such as a mismatch between the structured template and the associated free-text field) have not been described, to the researchers' knowledge.
Hardeep Singh, MD, of the Houston Veterans Affairs Health Services Research and Development Center of Excellence, and colleagues sought to determine the nature and frequency of such errors and identify their potential predictive variables.
In the study, the researchers enrolled pharmacists to report prescriptions containing inconsistent communication over a four-month period at a tertiary-care facility. They also electronically retrieved all prescriptions written during the study period containing any comments in the free-text field and then randomly selected 500 for manual review to determine inconsistencies between free-text and structured fields. Of these, prescriptions without inconsistencies were categorized as controls.
They collected data on potentially predictive variables from reported and unreported errors and controls. For all inconsistencies, the investigators sought to determine their nature (such as drug dosage or administration schedule) and potential harm and used multivariate logistic regression models to identify factors associated with errors and harm.
Of 55,992 new prescriptions, 532 (0.95 percent) were reported to contain inconsistent communication, a rate comparable to that obtained from the unreported group, according to the authors.
They found that drug dosage was the most common inconsistent element among both groups. Certain medications were more likely associated with errors, as was the inpatient setting (odds ratio, 3.30) and surgical subspecialty (odds ratio, 2.45).
Based on their findings, Singh and colleagues noted that about 20 percent of errors could have resulted in moderate to severe harm to patients.