Facebook Twitter YouTube

16053

Neonatal Handoffs

Friday, October 19, 2012
Room R02-R05 (Morial Convention Center)
Chris DeRienzo, MD, MPP, Monica Horvath, Rob Lenfestey and Jeff Ferranti, Division of Neonatal-Perinatal Medicine, Duke University Health System, Durham, NC

Purpose: To propose a core dataset for neonatal intensive care unit (NICU) handoffs and analyze errors in written handoffs performed by Duke NICU residents.

Methods: We identified and defined core data elements for a NICU handoff (Table 1) and performed a retrospective analysis of quality improvement data collected over 1 month in 2011. Throughout the QI project, one neonatal fellow was present within the NICU for a total of 5 nights, during which on-service residents printed additional copies of written handoff lists at signout. The investigator compared these lists to real-time auto-populated data within the tablet-based Clinical Care View (CCV) environment or actual bedside data. Resulting data were then analyzed for errors of omission and commission.

Results: A total of 103 handoffs (33 unique patients) were analyzed. Among data elements controlled by providers (e.g., not auto-populated) there were 223 errors of commission and 92 errors of omission for a total of 315 errors (3.06 errors/handoff). Most patients were subjected to more errors of commission than of omission; one patient had notably more omission errors (13 vs. 3), while 5 had equal numbers of both error types. Though patients who were handed off more than once experienced more errors, 5 patients had ≥5 errors / handoff. Figure 1 then shows our analysis of errors by data element. Three elements – Top 5 Problems, Access, and Feeding Details – accounted for >50% of all errors. Data elements that were auto-populated into the signout lists by definition had no errors in comparison to the gold-standard (CCV), as both auto-populate simultaneously through the same EMR system.

Conclusion: After defining core data elements and analyzing errors for NICU handoffs, we found that residents commit more errors of commission than omission in written handoffs. The former occurred most often within data elements that required continual manual updating but which CCV demonstrates could be auto-populated within our existing IT environment. Auto-population may thus present opportunities for reducing errors in written handoffs.

Table 1 – Core Content for NICU Handoff

Data Element

Description

Demographics

Name, Team, Location, Medical Record Number, Adjusted Gestational Age

Weight

Today's Weight and Dosing Weight

Top 5 Problems

First 5 problems in daily note problem list

Medications

Auto-populated from CPOE system

Access

Vascular access/dialysis catheters

Respiratory Support (Basic)

Basic description of respiratory support (“Room Air,” “CPAP,” “SIMV”)

Respiratory Support (Details)

Extended description of respiratory support – e.g., CPAP 5 and 21% FiO2, SIMV with PIP 17, PEEP 5, rate 25 and 30% FiO2

Total Fluids

Volume/kg of total fluids order

Feeds

Volume of feeds

Feeding Details

Formula/breastmilk, additives, route, schedule

Pending Studies

Pending lab / radiographic / other studies to follow-up

Figure 1 – Errors of Omission and Commission in NICU Handoffs by Data Element