Association of Display of Patient Photographs in the Electronic Health Record With Wrong-Patient Order Entry Errors

Abstract

Importance: Wrong-patient order entry (WPOE) errors have a high potential for harm; these errors are particularly frequent wherever workflows are complex and multitasking and interruptions are common, such as in the emergency department (ED). Previous research shows that interruptive solutions, such as electronic patient verification forms or alerts, can reduce these types of errors but may be time-consuming and cause alert fatigue.

Objective: To evaluate whether the use of noninterruptive display of patient photographs in the banner of the electronic health record (EHR) is associated with a decreased rate of WPOE errors.

Design, setting, and participants: In this cohort study, data collected as part of care for patients visiting the ED of a large tertiary academic urban hospital in Boston, Massachusetts, between July 1, 2017, and June 31, 2019, were analyzed.

Exposures: In a quality improvement initiative, the ED staff encouraged patients to have their photographs taken by informing them of the intended safety impact.

Main outcomes and measures: The rate of WPOE errors (measured using the retract-and-reorder method) for orders placed when the patient's photograph was displayed in the banner of the EHR vs the rate for patients without a photograph displayed. The primary analysis focused on orders placed in the ED; a secondary analysis included orders placed in any care setting.

Results: A total of 2 558 746 orders were placed for 71 851 unique patients (mean [SD] age, 49.2 [19.1] years; 42 677 (59.4%) female; 55 109 (76.7%) non-Hispanic). The risk of WPOE errors was significantly lower when the patient's photograph was displayed in the EHR (odds ratio, 0.72; 95% CI, 0.57-0.89). After this risk was adjusted for potential confounders using multivariable logistic regression, the effect size remained essentially the same (odds ratio, 0.57; 95% CI, 0.52-0.61). Risk of error was significantly lower in patients with higher acuity levels and among patients whose race was documented as White.

Conclusions and relevance: This cohort study suggests that displaying patient photographs in the EHR provides decision support functionality for enhancing patient identification and reducing WPOE errors while being noninterruptive with minimal risk of alert fatigue. Successful implementation of such a program in an ED setting involves a modest financial investment and requires appropriate engagement of patients and staff.

Figure 1.. Patient Photograph Feature in the Epic Electronic Health Record

Demographic data are not from a real patient. The image of the face is also not of a real human; it was generated using Style Generative Adversarial Network (owned by Nvidia). The screenshot is used with permission from Epic Systems Corporation.

Figure 2.. Segmented Regression Analysis of the Percentage of Orders Placed for Patients With Photographs

Dashed lines represent the linear fit for each segment.

Figure 3.. Segmented Regression Analysis of the Rate of Retract-and-Reorder (RAR) Events Before and After the Implementation of the Photographs Program

Dashed lines represent the linear fit for each segment.