---
title: CliniGuard Diagnosis ICD NER
emoji: 🏥
colorFrom: blue
colorTo: green
sdk: gradio
sdk_version: 6.15.1
app_file: app.py
short_description: PubMedBERT clinical diagnosis entity extraction NER
python_version: "3.12"
startup_duration_timeout: 30m
---

# CliniGuard Diagnosis ICD NER

Interactive demo for [genzeonplatform/cliniguard-diagnosis-icd-ner](https://huggingface.co/genzeonplatform/cliniguard-diagnosis-icd-ner),
a transformer-based clinical Named Entity Recognition model built on **PubMedBERT** (`BertForTokenClassification`)
for automated extraction of diagnoses, conditions, and coding-support entities from unstructured clinical text.

The model recognizes **9 diagnosis and coding entity types** (19 BIO labels):

| Entity | Description |
|--------|-------------|
| `PRIMARY_DIAGNOSIS` | Principal reason for encounter |
| `SECONDARY_DIAGNOSIS` | Additional diagnoses |
| `DIFFERENTIAL_DIAGNOSIS` | Diagnoses under consideration |
| `COMORBIDITY` | Co-existing conditions |
| `COMPLICATION` | Hospital-acquired or treatment complications |
| `CHRONIC_CONDITION` | Long-term conditions |
| `ACUTE_CONDITION` | Acute episodes or events |
| `DIAGNOSIS_STATUS` | Current clinical status (active, resolved, improving) |
| `DIAGNOSIS_DATE` | Date of diagnosis or onset |

Enter clinical text (discharge summaries, progress notes, clinical narratives) and the model will
highlight and structure the recognized diagnosis entities using the standard HuggingFace
`token-classification` pipeline with `aggregation_strategy="simple"`.