On-admission pressure ulcer prediction using the nursing needs score

Area of Science:

• Nursing
• Clinical Prediction Models
• Patient Safety

Background:

• Pressure ulcers (PUs) pose a significant challenge in nursing care, necessitating effective preventive strategies.
• Current PU risk assessment methods often rely on data not available upon patient admission.
• The Nursing Needs Score (NNS), routinely collected in Japanese hospitals for staffing, has not been utilized for PU risk assessment.

Purpose of the Study:

• To develop a predictive model for PU development using only data accessible at the time of hospital admission.
• To evaluate the utility of the on-admission NNS score in predicting PU risk.

Main Methods:

• Logistic regression analysis was employed to construct the PU risk prediction model.
• A random undersampling technique was implemented to address data imbalance issues.

Main Results:

• A predictive model incorporating gender, age, surgical duration, and on-admission NNS score (NNS-B) demonstrated strong predictive performance.
• The optimal model achieved an average sensitivity of 69.2%, specificity of 82.8%, and an Area Under the Curve (AUC) of 84.0%.

Conclusions:

• A validated model for predicting PU development has been established using readily available admission data (gender, age, surgical duration, NNS-B).
• This model offers a practical tool to enhance nursing efficiency and reduce PU incidence by enabling early identification of high-risk patients.