Architecture

From patient data to
explainable clinical insight

ClinAssist processes structured EHR data, clinical notes, and imaging in real time — then tells the clinician exactly why it reached its conclusion.

How It Works

A three-stage pipeline

Built to operate within the triage window — fast, multimodal, and fully transparent.

01
📋

Ingest Patient Data

Structured EHR data — vitals, labs, demographics, comorbidities — are pulled automatically at point of registration. Clinical notes are captured via NLP. Imaging is ingested where available.

02
🧠

Multi-Modal Prediction

A transformer-based NLP model processes free-text notes, a gradient-boosted model handles structured EHR data, and a CNN/ViT pipeline analyses diagnostic imaging where available. All three feed into a unified risk score.

03
🔍

Multi-Method Explanation Layer

Every prediction is explained in real time using SHAP feature attribution, LIME local approximations, and attention visualisation for NLP outputs. The clinician sees exactly which features drove the score — and by how much.

Medical records
EHR Ingestion
Medical AI
Model Inference
Data explanation
SHAP Explanation
ClinAssist — Patient Risk Dashboard · ED Bay 3 · Live
Risk Score
87/100
⇧ High acuity — immediate review
Predicted Triage
ATS 2
Emergency — seen within 10 min
Model Confidence
94%
High — 847 similar cases
Why this score? — SHAP Explanations
Feature contribution to risk prediction
O₂ Saturation (91%)
+0.42
Systolic BP (88 mmHg)
+0.31
Age (72)
+0.22
Arrival Mode (ambulance)
+0.18
No prior ED visits
−0.09
Heart Rate (last 5 min)
112 bpm · tachycardic
Core Features

Built for the realities of frontline clinical care

Explainability

SHAP, LIME & Attention Transparency

Every prediction comes with layered explanations — SHAP feature attribution for tabular data, LIME local approximations, and attention visualisation for NLP. Clinicians can interrogate, challenge, and override — supported by clear reasoning, not blind outputs.

NLP

Clinical Notes Understanding

Transformer-based NLP reads triage and clinical notes in real time, extracting structured insight from unstructured language including medical shorthand and abbreviations.

Risk Stratification

Real-Time Acuity Scoring

Risk scores update dynamically as new data arrives — vitals changes, new labs, updated notes — keeping the clinician's picture current throughout the encounter.

Integration

Workflow-First Design

Designed around the 3–5 minute triage window. ClinAssist surfaces the right information at the right moment — without adding cognitive burden to high-pressure environments.

Imaging

Diagnostic Imaging Pipeline

CNN and Vision Transformer (ViT) architectures analyse imaging inputs — such as chest X-rays — as part of the multimodal pipeline, extending ClinAssist beyond structured data into visual diagnostics.

Trust

Human-Factors Validated

Clinician trust and decision accuracy are primary outcome measures — not just model AUC. We measure whether ClinAssist actually improves decisions, not just predictions.