Improving Accuracy of Handoff by Implementing an Electronic Health Record-generated Tool: An Improvement Project in an Academic Neonatal Intensive Care Unit

Abstract

Background: Written patient handoffs are susceptible to errors or incompleteness. The accuracy is dependent on the person inputting the information. Thus, handoff printouts generated by electronic health records (EHR) with automation reduces the risk of transcription errors and improves consistency in format. This single-center quality improvement project aims to increase the accuracy of handoff printouts with an EHR-generated handoff tool.

Methods: This project used a plan-do-study-act methodology. Participants included registered nurses, neonatal nurse practitioners, neonatal hospitalists, pediatric residents, neonatal fellows, and neonatologists. The goals were to (1) increase accuracy of information to 80%, (2) reduce verbal handoff time by 20%, (3) reduce the frequency of incorrectly listed medications below 20%, and (4) improve user satisfaction by 1 point (on a 5-point Likert scale) over 6 months. Baseline assessment included a survey and a review of handoff reports 4 months before transitioning to the new handoff tool. We created a new handoff tool using EHR autogenerated phrases (Epic SmartPhrases) and autopopulated fields for pertinent Neonatal Intensive Care Unit patient data.

Results: After the unit-wide implementation of the new tool, the accuracy of 16 patient data points increased from 51% to 97%, while the frequency of patients with incorrectly listed medications decreased from 51% to 0%. Handoff time remained unchanged, while a 5-question user satisfaction survey showed an increase on the Likert scale.

Conclusions: We demonstrated that handoff printouts generated by EHR have fewer inaccuracies than manually scripted versions and do not add to the time required to give verbal handoff.

Fig. 1.

Examples of the handoff tools before and after implementing changes. A, The old, existing handoff tool was a textbox with no structure, prompts, or autogenerated data. B, The new handoff tool features well-demarcated sections (demographics, patient problem list, medications list, respiratory support, daytime to-do list, and nighttime contingency plans). Note that much of the critical data is autopopulated by the EHR.

Fig. 2.

Quality improvement project work flow. A, This modified Ishikawa diagram depicts the contributing factors that make the prior handoff tool prone to error and incompleteness. B, Timeline of the quality-improvement project.

Fig. 3.

Free text responses in the baseline survey from 35 respondents are analyzed for frequent key terms. The Pareto chart depicts the most common responses. The critical contributors to handoff dissatisfaction at baseline include the old handoff tool’s lack of structure, outdated information, lack of consistency, inability to access the patient chart while editing the handoff, the lack of automation, and the wordiness of free-texted manually typed handoffs.

Fig. 4.

Statistical process control chart (U charts) demonstrates improvement in patient data accuracy in multiple categories after implementation of the new handoff tool. A, Patient composite data (demographics + medical information) accuracy improved from 51% to 97%. B, Statistical process control chart (U charts) demonstrating a reduction in the frequency of patients with incorrect or incomplete medication lists from an average of 51% to 0% after implementation of the new handoff tool. C, The control chart of time (in min) spent per patient during evening verbal handoff demonstrates no signals.

Fig. 5.

This user-satisfaction survey assesses the baseline satisfaction with the old handoff tool and twice after the new tool was launched. The questions are ranked on a 5-point Likert scale.