Background: Chatbots could be a scalable solution that provides an interactive means of engaging users in behavioral health interventions driven by artificial intelligence. Although some chatbots have shown promising early efficacy results, there is limited information about how people use these chatbots. Understanding the usage patterns of chatbots for depression represents a crucial step toward improving chatbot design and providing information about the strengths and limitations of the chatbots. Objective: This study aims to understand how users engage and are redirected through a chatbot for depression (Tess) to provide design recommendations. Methods: Interactions of 354 users with the Tess depression modules were analyzed to understand chatbot usage across and within modules. Descriptive statistics were used to analyze participant flow through each depression module, including characters per message, completion rate, and time spent per module. Slide plots were also used to analyze the flow across and within modules. Results: Users sent a total of 6220 messages, with a total of 86,298 characters, and, on average, they engaged with Tess depression modules for 46 days. There was large heterogeneity in user engagement across different modules, which appeared to be affected by the length, complexity, content, and style of questions within the modules and the routing between modules. Conclusions: Overall, participants engaged with Tess; however, there was a heterogeneous usage pattern because of varying module designs. Major implications for future chatbot design and evaluation are discussed in the paper.

Subthreshold depression has been found to be associated with considerable impairment and an increased risk of developing major depression. Although several randomized trials have examined the effects of psychological interventions for subthreshold depression in children and adolescents, no meta-analysis has integrated the results of these trials. We searched 4 bibliographic databases and included randomized trials comparing psychological interventions with control conditions in children and adolescents scoring above a cut-off of a depression questionnaire but not meeting diagnostic criteria for major depression (or persistent depressive disorder) according to a diagnostic interview. Effect sizes and incidence rates of major depression were pooled with random effects meta-analyses. A total of 12 trials with 1,576 children and adolescents met inclusion criteria. The overall effect size indicating the difference between treatment and control at post-test was g = 0.38 (95% CI = 0.14-0.63), which corresponds to a number-needed-to-treat (NNT) of 8.4. Heterogeneity was moderate to high (I² = 61; 95% CI = 28-79), and there was significant risk of publication bias (p < .04). The 2 studies in children less than 12 years of age showed nonsignificant effects (g = 0.01; 95% CI = -1.16 to 1.18). We found no significant effect on the incidence of major depression at follow-up (relative risk = 0.52; 95% CI = 0.25-1.08), although this may be related to low statistical power. Interventions for subthreshold depression may have positive acute effects in adolescents. There is currently insufficient evidence, however, that these interventions are effective in children less than 12 years of age, or that they prevent the onset of major depression at follow-up.

There is now substantial evidence that major depressive episodes can be prevented and treated effectively with psychological interventions administered by trained lay and professional providers. There is also evidence that depression can be prevented and treated using self-help digital interventions such as websites and apps, and that their effectiveness increases if these interventions are supported by guides or coaches. In this commentary we describe digital approaches for depression developed and tested across the world by Palo Alto University’s Institute for International Internet Interventions for Health. We recommend the creation of digital apothecaries to increase access to in-person interventions by 1) training more providers, and 2) developing tools providers can use to enhance the effectiveness of these interventions, and to provide access to self-help interventions that can be used directly by anyone with 3) either guides or coaches to reduce dropout, or 4) as fully automated interventions. Existen evidencias importantes que demuestran que los episodios depresivos mayores pueden prevenirse y tratarse eficazmente con intervenciones psicológicas administradas tanto por profesionales como por no profesionales debidamente capacitados. También hay evidencia de que la depresión se puede prevenir y tratar mediante intervenciones digitales de autoayuda tales como páginas web y aplicaciones y que su eficacia aumenta si estas intervenciones son apoyadas por guías o ,coaches,. En este comentario describimos enfoques digitales para la depresión desarrollados y probados en todo el mundo por el Instituto de Intervenciones Internacionales en Salud por Internet de la Universidad de Palo Alto. Recomendamos la creación de apotecarias digitales para aumentar el acceso a las intervenciones: 1) capacitando a más proveedores de ayuda y 2) desarrollando herramientas que estos puedan utilizar para mejorar la eficacia de estas intervenciones y para proporcionar acceso a intervenciones de autoayuda que puedan ser utilizadas directamente por cualquier persona con 3) guías o ,coaches, con el objetivo de reducir el abandono o 4) como intervenciones totalmente automatizadas.

The United States Latine/x population is now over 62 million and comprises nearly 19% of the population. Over 42 million speak Spanish and comprise the second-largest language group in the United States. As the largest non- English language group, Spanish speakers pose the greatest need for language-concordant mental health services. Providing language-concordant and culturally responsive services is critical to ensure meaningful access and improve treatment outcomes. The current authors use the pan-ethnic terms Latine or Latinx as encompassing the terms Hispanic (derived from Spain plus the entire Spanish colonized empire), Latino/a (gendered, for those with ties to Latin America specifically), and Chicano/a (a racist slur reclaimed as a symbol of pride for Mexican Americans specifically). Latine/x includes widely varied acculturation, language, and immigration statuses, all of which impact an individual’s ability and willingness to enter the United States mental health system and seek services. Indigenous individuals from Latin America likely do not align with this term, and may or may not speak Spanish, however, are likely placed in the above monolithic categories in the databases referenced herein.

Despite a general decline in smoking rates among the U.S. population, smoking among low-income populations remains disproportionately high, likely due to the social determinants of health. To inform tailored approaches and responsive public health policies, the aim of this study was to qualitatively explore the sociocultural contexts, attitudes, and behaviors regarding smoking and quitting in a sample of low-income adults who smoke. In-depth, semi-structured qualitative interviews were conducted with 20 adults. Participants were recruited through local safety-net primary care clinics and community sites. A thematic analytic approach was utilized to analyze transcribed interviews. In exploring smoking and quitting within the context of low-income individuals, the following six themes were identified: caught between health and tobacco use; the nuances of context; roadblocks to quitting; motivation without movement; a temporary escape; and one size does not fit all. Insights into sociocultural and environmental contexts that shape smoking and quitting among low-income individuals revealed a complex interplay of factors that perpetuate smoking behavior and make it difficult to achieve sustained cessation. The study findings point to the importance of patient-centered and collaborative approaches that tailor smoking cessation efforts to the unique needs and lived experiences of low-income people who smoke.

Background: Parent management training (PMT) is an evidence-based intervention for children with disruptive behavior problems. However, access to care is often limited by cost, availability of clinicians, scheduling, and transportation barriers. Integrating artificial intelligence (AI) into group PMT may improve accessibility, personalization, and adherence, while preserving therapeutic quality. Objective: This study explored caregivers’ experiences with a hybrid PMT program that combined live therapist-led group sessions with asynchronous support from Pat, an AI conversational agent designed to augment the therapists’ support to caregivers. Methods: A total of 88 caregivers of children aged 3-14 years (mean 7.98, SD 2.45 years) from Argentina and Paraguay participated in eight weekly online group sessions led by human therapists and supplemented by Pat. Caregivers were asked to provide the net promoter score (NPS) and their perceived contribution of their experience in remote group sessions and with Pat. Caregiver perspectives were analyzed using thematic analysis by multiple coders with consensus and interrater reliability assessment. Results: The average NPS was 76.92, indicating excellent satisfaction. Regarding participants’ perceptions of the overall program, the most frequent theme was useful strategies (73/202, 36.1%), reflecting the value placed on clear, structured, and practical tools to address everyday parenting challenges. Most comments about Pat were positive (156/164, 95.1%), particularly highlighting its 24/7 accessibility and constant availability (69/164, 42.1%). Recommendations for improvement mainly focused on enhancing the user experience and incorporating additional functionalities. Regarding perceived contribution to progress, caregivers attributed, on average, 61% to Pat and 46% to the group sessions. Conclusions: Combining therapist-led PMT group sessions with AI support appears feasible, acceptable, and valued by caregivers and may expand reach without sacrificing quality. Overall, caregivers valued the useful strategies and the professional and peer support and reported positive changes. Regarding Pat, the most valued aspect is the constant and immediate support. This model integrates human expertise with the accessibility and continuity provided by AI, reducing barriers, such as time, cost, and clinician availability.