Trigger-first large language model-assisted adverse drug reaction surveillance: A pharmacist-supervised retrospective cohort study
Source check: Source checked. Not independent scientific confirmation.
Finding: Among 7,950 exposure episodes, pharmacists adjudicated 28 ADR-positive; trigger selected 167 candidates and caught all 28. LoRA improved false-positive control significantly; sensitivity increase was nonsignificant, and 2 positives remained model-non-positive.
Technology: Qwen3-4B-Instruct assesses structured EHR case records for ADR evidence; a local LoRA model trained on 410 patient-disjoint cases involving other medicines is compared against the base model after prespecified electronic-trigger filtering.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Qwen3-4B-Instruct assesses structured EHR case records for ADR evidence; a local LoRA model trained on 410 patient-disjoint cases involving other medicines is compared against the base model after prespecified electronic-trigger filtering.
- Test and result
- Among 7,950 exposure episodes, pharmacists adjudicated 28 ADR-positive; trigger selected 167 candidates and caught all 28. LoRA improved false-positive control significantly; sensitivity increase was nonsignificant, and 2 positives remained model-non-positive. Two blinded pharmacists adjudicated the full cohort with third-reviewer resolution; the LLM comparison is paired on 167 trigger-positive cases with exact McNemar testing. Internal single-center evaluation for one preparation; no temporal/external validation, explicitly bounded in interpretation.
- Limitation / still unknown
- not significantly (P=.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC13718469/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.