Beyond Hospital-Level Aggregated Data: A Methodology to Adapt Clinical Data From the Electronic Health Record for...

Area of Science:

• Health Informatics
• Nursing Research
• Healthcare Management

Background:

• Existing nurse staffing studies often rely on aggregated hospital-level data, lacking the granularity for effective unit-specific decision-making.
• This limitation hinders the ability of nurse managers to make informed, frontline staffing adjustments based on real-time unit needs.

Purpose of the Study:

• To develop and describe a methodology for linking patient-level electronic health record (EHR) data to the specific nursing units where care was delivered.
• To create a national dataset that accurately reflects patient care at the unit level, facilitating context-specific research and operational analysis.

Main Methods:

• Utilized Veterans Health Administration (VHA) acute care hospitalizations from 2019.
• Linked patient-level EHR and barcode medication administration (BCMA) data to nursing units via a crosswalk.
• Segmented hospitalizations into ward stays based on time-stamped locations and calculated unit-level patient risk scores.

Main Results:

• Successfully linked 89.46% of ward stays and 93.10% of medication administrations to specific nursing units across 125 VHA facilities.
• Extracted data for 1117 nursing units, encompassing over 1.1 million ward stays and 147 million medication administrations.
• Demonstrated a higher average patient severity score at the unit level (4.71) compared to the facility level (4.53).

Conclusions:

• Accurate identification of nursing units is essential for leveraging EHR data to study unit-level nursing phenomena.
• This approach provides critical context-specific information vital for frontline managers making staffing and resource allocation decisions.
• The methodology enhances the utility of EHR data for improving nursing research and healthcare operational management.