Search

BACKGROUND AND OBJECTIVES: Failed extubations are associated with pulmonary morbidity in hospitalized premature newborns. The objective of this study was to use quality improvement methodology to reduce failed extubations through practice standardization and integrating a real-time extubation success calculator into the electronic medical record (EMR). METHODS: A specific, measurable, achievable, relevant, and time-bound aim was developed to reduce failed extubations (defined as reintubation <5 days from primary extubation) by 50% among infants <32 weeks' gestational age (GA) or <1500 g birth weight by December 31, 2022. Plan-do-study-act cycles were developed to standardize postextubation respiratory support and integrate the EMR-based calculator. Outcome measures included extubation failure rates. Balancing measures included days on mechanical ventilation and number of patients intubated <3 days. Process measures were followed for guideline compliance. Statistical process control charts were used to track time-ordered data and detect special cause variation. RESULTS: We observed a reduction in failed extubations from 10.3% to 2.3%, with special cause variation noted after both plan-do-study-act cycle #1 and #2. Special cause variation was detected in both GA subgroups: <28 weeks' GA (22.0%-8.6%) and ≥28 weeks' GA (4.6%-0.3%). Additionally, the average number of infants intubated <3 days increased (60.2%-73.6%), whereas average ventilator days decreased (10.8-7.0). Finally, the time from infants' extubation score reaching threshold (≥60%) to extubation decreased (14.1-6.4 days) after launching the EMR-integrated calculator. CONCLUSIONS: Practice standardization and implementation of an EMR-based real-time clinical decision support tool improved extubation success, promoted earlier extubation, and reduced ventilator days in premature newborns. [ABSTRACT FROM AUTHOR]

Diagnosis of adverse neonatal outcomes is crucial for preterm survival since it enables doctors to provide timely treatment. Machine learning (ML) algorithms have been demonstrated to be effective in predicting adverse neonatal outcomes. However, most previous ML-based methods have only focused on predicting a single outcome, ignoring the potential correlations between different outcomes, and potentially leading to suboptimal results and overfitting issues. In this work, we first analyze the correlations between three adverse neonatal outcomes and then formulate the diagnosis of multiple neonatal outcomes as a multi-task learning (MTL) problem. We then propose an MTL framework to jointly predict multiple adverse neonatal outcomes. In particular, the MTL framework contains shared hidden layers and multiple task-specific branches. Extensive experiments have been conducted using Electronic Health Records (EHRs) from 121 preterm neonates. Empirical results demonstrate the effectiveness of the MTL framework. Furthermore, the feature importance is analyzed for each neonatal outcome, providing insights into model interpretability.