Can Variables From the Electronic Health Record Identify Delirium at Bedside?

Area of Science:

• Gerontology
• Clinical Medicine
• Health Informatics

Background:

• Delirium is a prevalent and serious condition in elderly hospitalized individuals, frequently overlooked.
• Existing delirium prediction models seldom utilize electronic health record (EHR) data.
• Early identification of delirium is crucial for patient outcomes.

Purpose of the Study:

• To investigate the utility of EHR variables for bedside identification of delirium in older inpatients.
• To assess the association between specific EHR data points and the presence of delirium.

Main Methods:

• Prospective cohort study involving 408 inpatients aged 65 and older.
• Daily delirium screening using the 3-minute diagnostic Confusion Assessment Method (3D-CAM).
• Extraction of demographic and clinical data from the EHR.

Main Results:

• The overall delirium rate was 16.7%.
• Patients with delirium were older and had higher rates of infection, prior dementia, and higher comorbidity scores (Charlson, Morse Fall Scale) but lower Braden Scale scores.
• Multivariable analysis confirmed prior dementia (OR: 5.0) and Braden score <18 (OR: 2.8) as significant predictors of delirium.

Conclusions:

• EHR data, particularly prior dementia diagnosis and low Braden Scale scores, can aid in identifying hospitalized older adults with delirium.
• Further development of automated delirium prediction models using EHR data is warranted.