Prediction Model for Hospital-Acquired Pressure Ulcer Development: Retrospective Cohort Study.

Area of Science:

• Medical Informatics
• Clinical Nursing
• Public Health

Background:

• Hospital-acquired pressure ulcers (HAPUs) are costly injuries, with intensive care unit (ICU) patients at higher risk.
• Current risk assessment tools may not fully capture the complexities of HAPU development in ICUs.

Purpose of the Study:

• To develop and validate an ICU-specific prediction model for HAPUs using logistic regression.
• To compare the performance of this novel model against the established Braden scale in an ICU population.

Main Methods:

• A retrospective cohort study analyzed 12,654 patient encounters from an academic medical center's electronic health records.
• Multiple logistic regression identified significant predictors of HAPU, including age, diabetes, and ventilator use.
• Model performance was evaluated using sensitivity, specificity, AUC, and Youden index, compared to the Braden scale.

Main Results:

• The study identified 735 HAPUs (5.81% incidence) among ICU patients.
• The developed model achieved 91.7% overall accuracy and an AUC of 0.737.
• Factors like male gender, diabetes, and isolation status significantly increased HAPU risk.

Conclusions:

• An ICU-specific, data-driven HAPU prediction model demonstrates acceptable performance, comparable to the Braden scale.
• This model can augment clinical decision-making for HAPU risk assessment in intensive care settings.
• Implementation of this model may contribute to reducing HAPU incidence.