Private MedHELM binary task: given a pediatric ADHD visit note, say whether the clinician recommended parent training in behavior management.
unassessed
| Category | domain |
|---|---|
| Subcategory | pediatric ADHD clinical-note classification: parent training in behavior management |
| Page status | active |
| Metric | exact_match |
| Direction | higher_is_better |
| Unit | % |
| Dataset size | 423 |
| Publisher | Stanford University School of Medicine (Packard Children's Health Alliance / Stanford Medicine Children's Health); packaged in MedHELM by Stanford CRFM |
ADHD-Behavior (HELM scenario shc_ptbm_med) tests whether a model can read an English pediatric primary-care note and decide if the clinician recommended parent training in behavior management (PTBM). PTBM is the first-line evidence-based treatment for young children with ADHD in AAP-style guidelines. The HELM prompt asks for A (yes) or B (no) only. The notes come from Packard Children's Health Alliance visits of children aged 4-6 years with an ADHD diagnosis, 2015-2019. The same classification problem was first studied as an NLP quality-of-care measure, not as a public leaderboard task.
Binary A/B classification over a clinical note plus a question; English text; HELM multiple-choice joint adaptation.
No model card in ModelSpec reports this benchmark yet.