The Use of Electronic Health Record Metadata to Identify Nurse-Patient Assignments in the Intensive Care Unit:...

Area of Science:

• Health Informatics
• Nursing Research
• Critical Care Medicine

Background:

• Individual nurse characteristics significantly impact patient outcomes in intensive care units (ICUs).
• Previous research aggregated nursing data, limiting granular analysis of individual nurse contributions.
• Linking individual nurses to patients retrospectively has been a significant research barrier.

Purpose of the Study:

• To assess the feasibility of using electronic health record (EHR) nurse metadata for retrospective nurse-patient assignment identification in ICUs.
• To develop and validate an algorithm for determining individual nurse assignments from EHR data.

Main Methods:

• Utilized EHR data from 38 ICUs across 18 hospitals (2018-2020).
• Developed a deterministic algorithm based on nurse charting frequency and timing (clinical assessments, medication administration).
• Validated algorithm accuracy and precision through manual chart review on a subset of patient shifts.

Main Results:

• The algorithm successfully identified a single nurse for 97.3% of patient shifts.
• Achieved 93% accuracy and 94.4% precision in identifying primary nurses.
• Common misclassifications occurred during patient care transitions (transfers, discharges, admissions).

Conclusions:

• EHR metadata accurately identifies individual nurse-patient assignments in the ICU.
• This capability enables novel research into nurse staffing at the individual level.
• Findings support leveraging individual nurse-patient assignment data to improve patient outcomes.