APPLICATION OF NEXT GENERATION TECHNOLOGIES: EP.320 AI-based processing of patient voice in rare neuromuscular disorders: Understanding patient experience and early disease detection

Abstract

Patients and caregivers share insights on experience in diagnostics and treatment while looking for the second opinion or striving to support each other, esp. now in pandemic era. The research objective was to analyze the potential of Patient Voice (PVO) as the source of data and tool for the improvement of patient journey in rare neuromuscular diseases. PVO means firsthand unsolicited testimony by patients/caregivers, gathered with the AI-based multilingual Natural language understanding technology from open social networks or disease-related forums, anonymized according to GDPR. Several disorders were included, with various prevalence: SMA5q, Pompe, and several ultra-rare diseases. To provide cross-country comparison, EU, China and Russia were selected for the analysis. There were processed stories of diagnosed patients (up to 1,320 patients per an indication/country) and “potential ones”: cases where a diagnosis is not confirmed or unknown, but a case history is a perfect match to a disease of interest (up to 41,239 patients per an indication/country). PVO proved to be valuable for deeper understanding of QoL and burden issues (up to 57 types of problems discovered per an indication/country). For ultra-rare diseases, key value proposition lays in early detection of patients and building the registries for further navigation to diagnosis. PVO provides evidence on comorbidities (incl. rare cases, e.g. the one of SMA5q and phenylketonuria in 1 case), previous/wrong diagnoses before the final correct one (up to 24 diagnoses per a disease) and symptom journey (up to 49 groups of symptoms per a disease, with thousands of linguistic expressions considered). The latter form the basis for early detection (not just individual symptoms but various disease patterns and patient profiles). Other aspects covered include stakeholder mapping (KOLs, centers of excellence in patient eyes, online communities, HCPs visited – up to 23 specialties), time before diagnosis, diagnostic procedures carried out, treatment used, drivers/barriers for diagnostics/treatment.