Skip to content Skip to footer

Predictive Model for Hospital-Acquired Infections in Pediatric Patients with Spinal Cord Injuries: A Comprehensive Study Review

Hospital-acquired infections (HAIs) pose a significant risk to pediatric patients with spinal cord injuries (SCIs). These infections can cause hospital stays to be extended and increase the patient’s morbidity. A study conducted by multiple hospitals in China aimed to identify patients at greater risk of contracting these infections. This study was retrospective and spanned from January 2005 to December 2023. It covered 220 pediatric patients with SCIs, with the patients divided into two sets: a training set consisting of 154 patients and a validation set with 66 patients. A multivariate logistic regression analysis was utilized to determine risk factors associated with HAIs, from which a predictive nomogram was established.

A receiver operating characteristic (ROC) curves analysis, area under the ROC curve, and calibration plots were used to assess the performance of the model. The model’s capacity for discrimination was evaluated through the decision curve analysis, which determined its net benefit over clinical decision thresholds in both training and validation sets. The predictive nomogram used metrics such as age, time from the injury to hospital admission, history of urinary and pulmonary infections, urobilinogen positivity, damaged segments, and admission American Spinal Injury Association (ASIA) scores.

The model proved to have excellent discrimination in the training set, with the area under the curve (AUC) equalling 0.957. On the other hand, the model demonstrated good discrimination in the validation set with an AUC of 0.919. Calibration plots revealed a suitable fit between predicted probabilities and observed outcomes, thereby validating the model. This predictive model for HAIs in pediatric SCI patients is a promising tool for clinical application. It offers healthcare providers a mechanism for identifying patients at high risk for HAIs, augmenting the potential for early interventions and better patient care strategies.


Sign Up to Our Newsletter

Be the first to know the latest updates

[yikes-mailchimp form="1"]