Evaluation and validation of a prediction model for extubation success in very preterm infants

Abstract

ObjectiveTo evaluate the performance of a publicly available model predicting extubation success in very preterm infants. Study DesignRetrospective study of infants born < 1250 g at a single center. Model performance evaluated using the area under the receiver operating curve (AUROC) and comparing observed and expected probabilities of extubation success, de ned as survival ≥ 5 d without an endotracheal tube. ResultsOf 177 infants, 120 (68%) were extubated successfully. The median (IQR) gestational age was 27 weeks (25-28) and weight at extubation was 915 g (755-1050). The model had acceptable discrimination (AUROC 0.72 [95% CI 0.65-0.80]) and adequate calibration (calibration slope 0.96, intercept − 0.06, mean observed-to-expected difference in probability of extubation success − 0.08 [95% CI -0.01, -0.15]). ConclusionsThe extubation success prediction model has acceptable performance in an external cohort, supporting its potential utility in clinical decision-making. Additional studies are needed to determine if its use can improve outcomes.74.

A clinical informatics approach to bronchopulmonary dysplasia: current barriers and future possibilities

“…While several prediction models using logistic regression or ML methods have been developed to aid clinicians in determining the appropriate timing for extubation in extremely preterm infants, the models do not include many patients in the high-risk category of 22–24 weeks GA and none have been routinely adopted into clinical practice ( 64 – 68 ). Moreover, it is worth noting that the extubation prediction models available have been developed for patients who are at risk of developing BPD and not for patients who have already been diagnosed with the condition ( 69 ). Therefore, the development of extubation prediction models for neonates with BPD could prove to be even more valuable.…”

EBNEO commentary: Prediction of extubation failure among low birthweight neonates using machine learning

“…To achieve an 80% probability of extubation success, their model had a sensitivity/specificity of 54%/81% 4 . Their model was validated at an external site in another study with similar results 5 …”

“…3 Predictive models using demographic and clinical data have recently shown limited ability to predict extubation success. [3][4][5][6] To be effective, these models need to have high specificity and sensitivity to predict extubation failure to avoid keeping infants intubated too long.…”

“…4 Their model was validated at an external site in another study with similar results. 5 In the current study, the XGBoost model combined the following variables: patient demographics + edications + vitals and ventilator readings, and predicted extubation success with 93.1% sensitivity, 46.2% specificity, 83.1% positive predictive value and 70.1% negative predictive value. Although the model performed well in detecting patients who will successfully extubate, the specificity in predicting extubation failure is less than 50%.…”