Mining Severe Drug Hypersensitivity Reaction Cases in Pediatric Electronic Health Records: Methodology Development and Applications
Bibliographic record
Abstract
BACKGROUND: Severe drug hypersensitivity reactions (DHRs) refer to allergic reactions caused by drugs and usually present with severe skin rashes and internal damage as the main symptoms. Reporting of severe DHRs in hospitals now solely occurs through spontaneous reporting systems (SRSs), which clinicians in charge operate. An automatic identification system scrutinizes clinical notes and reports potential severe DHR cases. OBJECTIVE: The goal of the research was to develop an automatic identification system for mining severe DHR cases and discover more DHR cases for further study. The proposed method was applied to 9 years of data in pediatrics electronic health records (EHRs) of Beijing Children's Hospital. METHODS: The phenotyping task was approached as a document classification problem. A DHR dataset containing tagged documents for training was prepared. Each document contains all the clinical notes generated during 1 inpatient visit in this data set. Document-level tags correspond to DHR types and a negative category. Strategies were evaluated for long document classification on the openly available National NLP Clinical Challenges 2016 smoking task. Four strategies were evaluated in this work: document truncation, hierarchy representation, efficient self-attention, and key sentence selection. In-domain and open-domain pretrained embeddings were evaluated on the DHR dataset. An automatic grid search was performed to tune statistical classifiers for the best performance over the transformed data. Inference efficiency and memory requirements of the best performing models were analyzed. The most efficient model for mining DHR cases from millions of documents in the EHR system was run. RESULTS: For long document classification, key sentence selection with guideline keywords achieved the best performance and was 9 times faster than hierarchy representation models for inference. The best model discovered 1155 DHR cases in Beijing Children's Hospital EHR system. After double-checking by clinician experts, 357 cases of severe DHRs were finally identified. For the smoking challenge, our model reached the record of state-of-the-art performance (94.1% vs 94.2%). CONCLUSIONS: The proposed method discovered 357 positive DHR cases from a large archive of EHR records, about 90% of which were missed by SRSs. SRSs reported only 36 cases during the same period. The case analysis also found more suspected drugs associated with severe DHRs in pediatrics.
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 | 0.005 | 0.010 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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".