ConnectED App – Model Card (Clinician-Facing Version)
🔍 Purpose
The ConnectED App is a digital screening tool developed to assist clinicians in recognising patterns consistent with three underdiagnosed conditions:
- Hypermobile Ehlers-Danlos Syndrome (hEDS)
- Postural Orthostatic Tachycardia Syndrome (POTS)
- Mast Cell Activation Syndrome (MCAS)
These conditions often present with multisystem symptom clusters and are frequently misdiagnosed or overlooked due to bias, system complexity, or symptom attribution to mental health.
:max_bytes(150000):strip_icc()/terms_a_artificial-intelligence-ai_asp-FINAL-ddba8ac599f3438d8064350d2ee1ae5a.jpg)
🤖 Model Overview
ConnectED uses a weighted scoring algorithm to evaluate user responses across five phenotype groupings:
- Cardiovascular
- Immune/MCAS
- Metabolic Support
- Neurodivergent (formerly Mixed)
- Cystic/Secretory
Each question is assigned a base value of 1 point, with additional weights (1–3) based on the strength of supporting evidence:
Weight 1 = Internationally accepted guidelines (e.g., Brighton Criteria, CHOP MCAS)
Weight 2 = Peer-reviewed research (e.g., PubMed-indexed)
Weight 3 = Evidence-informed but non-peer-reviewed (e.g., Mast Cell Action guidelines)
📈 Scoring Thresholds (V2)
For each condition:
- Likely: 50–100%
- Possible: 30–49%
- Unlikely: <30%
This adjustment improves sensitivity and identifies more potential candidates for further evaluation while minimising false negatives.
🧬 Diagnostic Enhancements
- Includes safe, at-home autonomic testing (e.g., NASA Lean Test, MALMO Protocol) to avoid adverse events common with tilt table testing
- Links clinical patterns to research-based symptom domains
- Maps likely phenotype expression (e.g., cystic = CFTR, KLK4 variants; ND = pyridoxine pathway)
🧠 Clinical Model Justification
- Developed by clinicians and researchers in collaboration with lived-experience experts
- Aims to shorten diagnostic delay, reduce gender bias, and flag multisystem clusters early
- Promotes trauma-informed and neurodivergent-affirming care
🧾 Output & Integration
- Patients receive a downloadable report containing Likely/Possible/Unlikely results, with evidence-based explanatory text for their GP/specialist
- Future versions will include phenotype-specific modular reports (e.g., Cystic Phenotype, Cardiovascular Report, etc.)
⚠️ Limitations
- ConnectED is not a diagnostic tool
- Requires clinician interpretation alongside clinical judgment
- User-reported data is not externally verified
📚 Evidence Base
ConnectED is grounded in:
- Brighton Diagnostic Guidelines for hEDS (Castori et al., 2017)
- CHOP and Molderings criteria for MCAS
- POTS consensus statement (Raj et al., 2020)
- Recent publications in neuroinflammation, autonomic dysfunction, and post-COVID phenotypes
- KLK gene research by Norris Lab (Gensemer et al., 2023 preprint)
📲 For more information or to access model validation materials
visit: www.connectedapp.health