Follow-up of patients with severe asthma using a conversational AI virtual assistant: feasibility and impact

Abstract

<bold>Introduction:</bold> The follow-up of patients with severe asthma (SA) represents a current challenge due to disease control variability requiring long-term monitoring. This study aimed to evaluate the feasibility and impact, in terms of reach and adherence, of a follow-up program using a conversational artificial intelligence (AI)-based virtual assistant in patients with SA. <bold>Methods:</bold> This single-center, prospective observational study followed patients at the Asthma Unit for 22 weeks. The virtual assistant Lola (Tucuvi®) conducted automated monthly calls to collect clinical data. Adults with SA were included, excluding those with language barriers or recent clinical trial participation. <bold>Results:</bold> A total of 168 patients were included, with a median age of 63 years, of whom 66% were female (demographic data in Fig <fig><object-id>erj;66/suppl_69/PA3638/F1</object-id><object-id>F1</object-id><object-id>F1</object-id><graphic></graphic></fig>1). Over the 22-week period, 818 automated calls were fully completed (4.86 calls per patient), resulting in a program reach of 88% and an adherence rate of 96%. Alerts generated by the AI led to the early detection of 14 exacerbations, 17 medication adjustments, 16 unscheduled evaluation visits, and 3 emergency department referrals. <bold>Conclusions:</bold> A follow-up program using a conversational AI virtual assistant achieved high reach and adherence, facilitating early detection of clinical deterioration and prompt interventions. However, its clinical efficacy in improving outcomes remains unproven.

Ähnliche Arbeiten