P.082 Using machine-based learning to predict neonatal hypoxic-ischemic encephalopathy (HIE) severity based on patient-specific clinical factors
Bibliographic record
Abstract
Background: Severity of HIE is based on Sarnat classification; however, it is difficult to predict precise neurodevelopmental outcomes as this only provides a single snapshot in time. We aimed to use machine-based learning to better understand variables contributing towards HIE severity. Methods: Patients with HIE treated with hypothermia were studied between 2014 and 2020 at level 3 NICUs in Calgary, Alberta. Clinical information contained 23 features including specifics of clinical examination, blood work, MRI and EEG findings, and medications. Random forest models were trained to examine features most predictive of HIE severity. Results: Two hundred and six patients were eligible. By grouping patients based on the initial Sarnat score and post-cooling exam, features correctly predicted groups 43% and 73% of the time, respectively. Precision, accuracy, and recall was best for the mild group. Using MRI and day 1 seizures it was 54% and 67% predictive, respectively. Features contributing most included arterial pH, initial lactate, and overall EEG findings. There are ongoing analyses for further classification. Conclusions: Machine-based learning can improve predictive models for patient outcomes. There is benefit in using variables outside of the initial examination to improve classification. We aim to expand this model to include detailed neurodevelopmental outcomes to improve prognostication.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.011 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.007 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".