Determinants of Dropout from a Virtual-Agent Based App for Insomnia Management: An Analysis in a Longitudinal, Self-Selected Sample of Users with Insomnia Symptoms (Preprint)

JMIR Mental Health, Journal Year: 2024, Volume and Issue: 12, P. e51022 - e51022

Published: Oct. 3, 2024

Fully automated digital interventions delivered via smartphone apps have proven efficacious for a wide variety of mental health outcomes. An important aspect is that they are accessible at low cost, thereby increasing their potential public impact and reducing disparities. However, major challenge to successful implementation the phenomenon users dropping out early. The purpose this study was pinpoint factors influencing early dropout in sample self-selected virtual agent (VA)-based behavioral intervention managing insomnia, named KANOPEE, which freely available France. From January 2021 December 2022, 9657 individuals, aged 18 years or older, who downloaded completed KANOPEE screening interview had either subclinical clinical insomnia symptoms, 4295 (44.5%) dropped (ie, did not return app continue filling subsequent assessments). primary outcome binary variable: having after completing assessment (early dropout) all treatment phases (n=551). Multivariable logistic regression analysis used identify predictors among set sociodemographic, clinical, sleep diary variables, users' perceptions program, collected during interview. mean age 47.95 (SD 15.21) years. Of those treatment, 65.1% (3153/4846) were women 34.9% (1693/4846) men. Younger (adjusted odds ratio [AOR] 0.98, 95% CI 0.97-0.99), lower education level (compared middle school; high school: AOR 0.56, 0.35-0.90; bachelor's degree: 0.35, 0.23-0.52; master's degree higher: 0.22-0.55), poorer nocturnal (sleep efficiency: 0.64, 0.42-0.96; number awakenings: 1.13, 1.04-1.23), more severe depression symptoms (AOR 1.12, 1.04-1.21) significant out. When measures included model, perceived benevolence credibility VA decreased 0.91, 0.85-0.97). As traditional face-to-face cognitive therapy presence plays an role dropout. This variable represents target address increase engagement with fully management programs. Furthermore, our results support contention can provide relevant user stimulation will eventually pay terms engagement.

Language: Английский

---

A New Research Model for Artificial Intelligence–Based Well-Being Chatbot Engagement: Survey Study (Preprint)

Published: April 27, 2024

Artificial intelligence (AI)–based chatbots have emerged as potential tools to assist individuals in reducing anxiety and supporting well-being. This study aimed identify the factors that impact individuals’ intention engage their engagement behavior with AI-based well-being by using a novel research model enhance service levels, thereby improving user experience mental health intervention effectiveness. We conducted web-based questionnaire survey of adult users China via social media. Our collected demographic data, well range measures assess relevant theoretical factors. Finally, 256 valid responses were obtained. The newly applied was validated through partial least squares structural equation modeling approach. explained 62.8% (<i>R</i>²) variance 74% behavior. Affect (β=.201; <i>P=</i>.002), (β=.184; <i>P=</i>.007), compatibility (β=.149; <i>P=</i>.03) statistically significant for engage. Habit (β=.154; <i>P=</i>.01), trust (β=.253; <i>P&lt;</i>.001), (β=.464; <i>P&lt;</i>.001) new extended provides basis studying users’ chatbot highlights practical points developers chatbots. It also importance create an emotional connection users.

Language: Английский

---

Acceptability and Feasibility of the English Version of Elevida, a Self-Guided Online Fatigue Intervention for People With Multiple Sclerosis

International Journal of MS Care, Journal Year: 2024, Volume and Issue: 26(Q4), P. 347 - 354

Published: Dec. 9, 2024

Fatigue is common in multiple sclerosis (MS); it significantly impairs quality of life, and treatment options are limited. A randomized controlled trial Elevida, a self-guided, online German fatigue intervention, showed significant benefit. We tested an English version Elevida with people MS Australia. Participants were volunteers who self-reported at least mild (≥ 43 on the Scale for Motor Cognitive Functions scale), some mobility (Expanded Disability Status < 8), no or cognitive difficulties (≤ 32 Multiple Sclerosis Neuropsychological Questionnaire). completed 9-week program, commenting rating its acceptability. The Chalder was baseline, end-of-program, 2 months later. undertook qualitative (thematic analysis) quantitative (before/after differences, using paired t test) analyses. Thirty-eight expressed interest study; 26 eligible; 20 began study. Fifteen participants (75%) program (mean [SD]: 58.9 [10.5] years age, 67% women, 9 relapsing MS, 6 progressive MS). Over 90% completing rated acceptability as good very good, approximately 70% found helpful. Three themes identified: Positive negative comments features, incorrect assumptions content, personal experiences reflections. Significant improvement (P .01) scores from baseline to completion maintained after completion. acceptable effective MS-related fatigue. Identified will guide further development satisfy users' sense autonomy, competence, relatedness.

Language: Английский

---

Digital Psychiatry: Opportunities, Challenges, and Future Directions

Journal of Psychiatric Practice, Journal Year: 2024, Volume and Issue: 30(6), P. 400 - 410

Published: Nov. 1, 2024

Recently, the field of psychiatry has experienced a transformative shift with integration digital tools into traditional therapeutic approaches. Digital encompasses wide spectrum applications, ranging from phenotyping, smartphone wearable devices, virtual/augmented reality, and artificial intelligence (AI). This convergence innovations potential to revolutionize mental health care, enhancing both accessibility patient outcomes. However, despite significant progress in psychiatry, its implementation presents plethora challenges ethical considerations. Critical problems that require careful investigation are raised by issues such as data privacy, divide, legal frameworks, dependability instruments. Furthermore, there risks several hazards associated psychiatric practice. A better understanding growing is needed promote development effective interventions improve accuracy diagnosis. The overarching goal this review paper provide an overview some current opportunities highlighting benefits inherent challenges. also aims at providing guidelines for future research proper clinical

Language: Английский

---

Combining AI and Human Support in Mental Health: a Digital Intervention with Comparable Effectiveness to Human-delivered Care (Preprint)

Journal of Medical Internet Research, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 28, 2024

Escalating mental health demand exceeds existing clinical capacity, requiring scalable digital solutions. However, engagement remains challenging. Conversational agents enhance by making programs more interactive and personalized but have not been widely used. This study evaluated a program for anxiety against external comparators. The used an AI-driven conversational agent to deliver clinician-written content via machine learning, with clinician oversight user support. aimed evaluate the engagement, effectiveness, safety of this structured, evidence-based human support mild, moderate severe generalized anxiety. Statistical analyses determine whether reduced than propensity-matched waiting control was statistically non-inferior real-world face-to-face typed cognitive behavioral therapy (CBT). Prospective participants (N=299) were recruited from NHS or social media in UK given use up 9 weeks (study conducted October 2023 May 2024). Endpoints collected before, during after program, at one-month follow-up. External comparator groups generated through propensity-matching sample Talking Therapies (NHS TT) data ieso Digital Health (typed-CBT) Dorset Healthcare University Foundation Trust (DHC) (face-to-face CBT). Superiority non-inferiority compare symptom reduction (change on GAD-7 scale) group groups. included time spent per participant calculated. Participants median 6 hours over 53 days, 78% (n=232) engaged (i.e. completed 2 14 days). There large clinically meaningful symptoms (per-protocol (PP; n=169): change = -7.4, d 1.6; intention-to-treat (ITT; n=299): -5.4, d=1.1). PP effect superior (d 1.3), CBT (p <.001) typed-CBT <.001). Similarly, ITT sample, showed superiority (d=0.8) (p=.002) approaching significance (p=.06). Effects sustained Clinicians overseeing mean 1.6 (31 - 200 minutes) sessions participant. By combining AI support, achieved outcomes comparable human-delivered care while significantly reducing required 8 times relative global estimates. These findings highlight potential technology scale healthcare, address unmet need, ultimately impact quality life economic burden globally. ISRCTN id: 52546704.

Language: Английский

---