Related Experiment Video
Updated: Jul 3, 2025
Assessment and Evaluation of the High Risk Neonate: The NICU Network Neurobehavioral Scale
Published on: August 25, 2014
Developing a practical neurodevelopmental prediction model for targeting high-risk very preterm infants during visit
Hao Wei Chung1,2,3,4, Ju-Chieh Chen2, Hsiu-Lin Chen1,5
1Division of Neonatology, Department of Pediatrics, Kaohsiung Medical University Chung-Ho Memorial Hospital, Kaohsiung Medical University, Kaohsiung, Taiwan.
Insights
Predicting neurodevelopmental outcomes for very preterm infants (VPI) is crucial. New machine learning models offer transparent and explainable predictions using fewer variables, improving early intervention strategies for VPI.
Area of Science:
- Neonatal follow-up
- Neurodevelopmental trajectories
- Machine learning in pediatrics
Background:
- Follow-up visits are critical for very preterm infants' (VPI) neurodevelopmental outcomes.
- Ensuring VPI attend follow-up before 12 months corrected age (CA) is challenging.
- Post-discharge data can enhance neurodevelopmental predictions due to brain plasticity.
Purpose of the Study:
- To develop and evaluate machine learning models for predicting cognitive and motor function in VPI.
- To compare the performance of an evolutionary-derived machine learning method (EL-NDI) against traditional models.
- To create transparent and explainable models for clinical use in VPI follow-up.
Main Methods:
- Developed four prediction models (cognitive/motor, 6/12 months CA) using VPI data (2010-2017).
- Defined cognitive/motor decline at 6 months CA and delay at 12 months CA using Bayley Scales of Infant Development 3rd edition (BSIDIII).
- Utilized an evolutionary-derived machine learning method (EL-NDI) and compared it with lasso regression, random forest, and support vector machine.
Main Results:
- EL-NDI models required fewer variables (4-10) compared to other methods (≥29) for similar performance.
- EL-NDI achieved areas under the receiver operating curve (AUC) of 0.76-0.83 for 6-month CA predictions.
- EL-NDI achieved AUCs of 0.73-0.82 for 12-month CA predictions.
Conclusions:
- The EL-NDI model demonstrates strong predictive performance with enhanced simplicity, transparency, and explainability.
- Clinical implementation of EL-NDI can aid in identifying high-risk VPI for timely early intervention.
- These models can foster better communication between clinicians and parents regarding VPI neurodevelopmental progress.
Background:
Follow-up visits for very preterm infants (VPI) after hospital discharge is crucial for their neurodevelopmental trajectories, but ensuring their attendance before 12 months corrected age (CA) remains a challenge. Current prediction models focus on future outcomes at discharge, but post-discharge data may enhance predictions of neurodevelopmental trajectories due to brain plasticity. Few studies in this field have utilized machine learning models to achieve this potential benefit with transparency, explainability, and transportability.
Methods:
We developed four prediction models for cognitive or motor function at 24 months CA separately at each follow-up visits, two for the 6-month and two for the 12-month CA visits, using hospitalized and follow-up data of VPI from the Taiwan Premature Infant Follow-up Network from 2010 to 2017. Regression models were employed at 6 months CA, defined as a decline in The Bayley Scales of Infant Development 3rd edition (BSIDIII) composite score > 1 SD between 6- and 24-month CA. The delay models were developed at 12 months CA, defined as a BSIDIII composite score < 85 at 24 months CA. We used an evolutionary-derived machine learning method (EL-NDI) to develop models and compared them to those built by lasso regression, random forest, and support vector machine.
Results:
One thousand two hundred forty-four VPI were in the developmental set and the two validation cohorts had 763 and 1347 VPI, respectively. EL-NDI used only 4-10 variables, while the others required 29 or more variables to achieve similar performance. For models at 6 months CA, the area under the receiver operating curve (AUC) of EL-NDI were 0.76-0.81(95% CI, 0.73-0.83) for cognitive regress with 4 variables and 0.79-0.83 (95% CI, 0.76-0.86) for motor regress with 4 variables. For models at 12 months CA, the AUC of EL-NDI were 0.75-0.78 (95% CI, 0.72-0.82) for cognitive delay with 10 variables and 0.73-0.82 (95% CI, 0.72-0.85) for motor delay with 4 variables.
Conclusions:
Our EL-NDI demonstrated good performance using simpler, transparent, explainable models for clinical purpose. Implementing these models for VPI during follow-up visits may facilitate more informed discussions between parents and physicians and identify high-risk infants more effectively for early intervention.
