Dados do Trabalho

Título

Real world evidence of a multicomponent CBTi program with chatbot and AI on reducing anxiety and depressive symptoms

Introdução

The co-occurrence of insomnia with anxiety and depression is prevalent and well observed in the general population. The gold-standard treatment for insomnia, Cognitive-Behavioral Therapy for Insomnia (CBTi), has been shown to ameliorate such symptoms, applied by a therapist or in digital formats. However, few digital solutions have been described with real-world data, and particularly testing the extent at which they might improve anxiety or depression.

Objetivo

Our goal was to describe therapeutic response and changes in anxiety and depressive symptomatology after a multicomponent digital program developed by Vigilantes do Sono (Sleep Watchers) using chatbot and Artificial Intelligence (AI).

Métodos

A virtual assistant interacts with users daily, for 5-10 minutes during ~7 weeks, asking them to complete diaries and delivering CBTi knowledge pills in ~45 sessions, distributed in seven modules. The Insomnia Severity Index (ISI), the General Anxiety Disorder (GAD-2) and the Patient Health Questionnaire (PHQ-2) are used before and after the modules of the program. Anxiety and depressive symptomatology were defined by scores ≥3 in GAD-2 and PHQ-2, respectively. Participants (18+ years) were recruited through social media, organic search, or referred by health-care professionals, from February/2021 to August/2022. All of them electronically signed an informed consent.

Resultados

A total of 1.447 participants (69% women, mean age 41.1±12.7) had insomnia (ISI>7) and fulfilled the GAD-2 and the PHQ-2 at least twice, of whom 76% (n=1.100) and 63% (n=916) scored positively (≥3) for anxiety and depressive symptomatology, respectively. The correlations between baseline values of the ISI with both GAD-2 (⍴=0.44; p<0.001) and PHQ-2 (⍴=0.44; p<0.001) were significant, as well as the correlation between them (⍴=0.56; p<0.001). Modules 2, 3, 4, 5, 6, and 7 were completed by 66%, 53%, 45%, 38%, 32%, and 26% of participants, respectively, of whom 52%, 55%, 59%, 61%, 63%, and 66% reached either therapeutic response (ISI reduction >7) or insomnia remission (ISI<7). Among those who scored positively on GAD-2 and PHQ-2, 38% and 44% had a score reduction of at least 3 points or 50%, respectively. These rates increased to 57% and 60% for individuals who completed all modules.

Conclusões

The findings support the efficacy of the program in reducing anxiety and depressive symptomatology, and reinforce the use of AI in optimizing this new therapeutic modality.

Palavras -chave

Depression, mental health, digital therapeutics, Cognitive-Behavioral Therapy for Insomnia (CBTi)