Authors:

Mansi Abhijit Dhamne

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Hypermobile Ehlers-Danlos Syndrome (hEDS) is a heritable connective tissue disorder characterized by joint hypermobility, pain, fatigue, and variable dysautonomia. Its rarity and variable symptom profile make diagnosis an arduous journey, further complicated by the absence of known genetic markers. Even post-diagnosis, traditional episodic care models often fail to capture the subjective and fluctuating nature of hEDS symptoms, leading to fragmented clinical insight and suboptimal treatment. To aid diagnosis and to facilitate a personalized treatment plan, we present Health Persona, an AI-powered mobile health system for continuous, patient-centric monitoring of autonomic physiology, symptoms, and contextual information (e.g., environment, time of day, activity). The system features a privacy-preserving architecture that captures and annotates multimodal data, including speech, text, and physiological signals from wearable devices and in-app interactions. A conversational AI interface supports patient engagement through symptom journaling and check-ins. Data are analysed using adaptive methods to identify patterns across symptoms, physiology, and context, enabling actionable insights such as flare predictors, symptom clusters, and consultation summaries. Initial development focuses on usability, flexible symptom entry, and chatbot accuracy based on user feedback. To assess the reliability of the speech-to-text component used in patient- clinician summaries, we compared ASR outputs to manually transcribed consultations (N=10). The system achieved an average Word Error Rate (WER) of 8.96%, indicating high transcription fidelity for downstream tasks. Health Persona offers a promising path toward earlier diagnosis and individualized, data-driven care for the holistic treatment of hEDS patients. Future work will evaluate its clinical relevance and impact on patient outcomes. Keywords: Machine Learning in Healthcare; Digital Health; Hypermobile Ehlers Danlos Syndrome; Conversational AI; Wearable Devices † Presenting Undergrad Author; ‡ Contributing Undergrad Author; * Undergrad Acknowledgment

Source:

Purdue University / 2025

Topics:

No topics listed

Co-authors:

Mansi Abhijit Dhamne