1117 — Design and Evaluation of an Override Mechanism to Enhance Medication Safety for Veterans

Russ AL, Roudebush VAMC HSR&D Center for Health Information and Communication (CHIC), CIN 13-416, Indianapolis, IN; Chen S, Roudebush VAMC HSR&D CHIC, Indianapolis, IN; Melton BL, School of Pharmacy, University of Kansas, Lawrence, KS; Saleem JJ, Department of Industrial Engineering, University of Louisville, Louisville, KY; Spina JR, Department of Veterans Affairs, VA Greater Los Angeles Healthcare System, Los Angeles, CA; Weiner M, Roudebush VAMC HSR&D CHIC, CIN 13-416, Indianapolis, IN; Daggy JK, Roudebush VAMC HSR&D CHIC, Indianapolis, IN; Zillich AJ, College of Pharmacy, Purdue University, West Lafayette, IN;

Objectives:
The VA's computerized provider order entry system requires prescribers to enter a free-text override reason to bypass high-risk medication alerts. However, override reasons entered are often inappropriate and not useful to pharmacists who review prescriptions. Our objectives were to: 1) develop a menu of override choices, derived from literature evidence on what override reasons are useful to pharmacists; and 2) improve the override mechanism for prescribers. We hypothesized that the new override mechanism would improve usability for prescribers and increase the clinical appropriateness of override reasons entered.

Methods:
We conducted a counterbalanced, crossover study with VA prescribers to evaluate the original versus new override mechanism. Data were collected in the VA HSRD Human-Computer Interaction and Simulation Laboratory. All outpatient primary care prescribers were invited to participate. Prescribers were video recorded as they completed prescribing tasks for fictitious patients. We assessed usability (learnability, perceived efficiency) and used a-priori criteria to evaluate the clinical appropriateness of override reasons. Outcomes for perceived efficiency and clinical appropriateness of override reasons were analyzed with the Wilcoxon signed-rank test to compare the original to the new override mechanism.

Results:
Twenty prescribers participated. The number of clinically appropriate override reasons significantly increased for the new versus original mechanism (median change of 3.0 (interquartile range = 3.0); p < .0001). Five prescribers had difficulty learning how to use the override mechanism when first encountering the new design. Prescribers rated the new override mechanism as more efficient than the original (p = 0.032).

Implications:
A menu-based choice, compared to free-text entry, improved prescribers' perceived efficiency and significantly increased the clinical appropriateness of override reasons. Further design modifications are needed to aid learnability, but the new override mechanism may enhance physician-pharmacist coordination.

Impacts:
This study represents an important first step towards evidence-based design that fosters clinically appropriate overriding of medication alerts. Results may inform VA's future electronic health record and a more standardized override design across alert systems. These efforts may ultimately improve care coordination and enhance medication safety for patients.