Viet Cuong Nguyen et al. · 2026 · arXiv (Cornell University) · Open access
Therapeutic dialogue is not a sequence of isolated responses: client goals, motivation, resistance, and therapeutic alliance evolve over time. Yet current LLM-based mental health dialogue systems often lack explicit mechanisms for tracking these dynamics across extended interactions, which can lead to poorly timed interventions or premature goal resolution. We introduce CALM-IT, a framework for generating and evaluating long-form Motivational Interviewing dialogues through explicit modeling of evolving client and counselor states, guiding both counseling strategy selection and utterance generation. We evaluate CALM-IT on a large-scale corpus of 8,232 synthetic dialogues spanning multiple dialogue lengths and frameworks. Compared with all baselines, CALM-IT achieves the best performance on most MITI 4.2 global ratings, including Empathy, Partnership, and Softening Sustain Talk, as well as on other key performance metrics while exhibiting minimal performance degradation as dialogue length increases. Notably, although CALM-IT initiates fewer change-directed prompts, it produces the highest client acceptance rate (64.3%) on average across different length conditions. We release a reproducible generation framework, a MITI-grounded process-level evaluation protocol, and a large-scale synthetic corpus for studying therapeutic LLMs under realistic long-form interaction conditions.
Research and education tool only. Not for diagnosis, emergency care, legal advice, or treatment recommendations. Verify citations against original sources.
Donna Pettey · 2015 · uO Research (University of Ottawa) · Open access
While tobacco use remains a leading preventable risk factor for mortality and morbidity in Canada (Patra, Rehm, Baliunas & Popova, 2007), the overall smoking prevalence rate of Canadians has decreased substantially from close to 50% of the population in 1965 to 16.1% of the population in 2012 (Canadian Tobacco Use Monitoring Survey (CTUMS) 2012; Reid, Hammond, Rynard & Burkhalter, 2014). However, up to 85% of individuals with a mental illness continue to use tobacco products (Harris, Parle & Gagne, 2007), contributing to an inequitable distribution of negative health outcomes for this population. Individuals with severe mental illness die an estimated twenty-five years earlier than the general population, with sixty per cent of these deaths due to cardiovascular, pulmonary and infectious disease (Parks, Svendsen, Singer, & Foti, 2006). A recent study that examined specific tobacco-attributable deaths in these populations found that tobacco accounted for 53% of deaths in individuals with schizophrenia, 50% of all deaths for those diagnosed with a depressive disorder, and 48% of all deaths for those with a diagnosis of bipolar disorder (Callaghan et al., 2014). This research project is intended to increase our understanding of what constitutes an effective intervention for smoking cessation and smoking reduction in a population of individuals with severe mental illness who are homeless or vulnerably housed, living in a large urban setting. Two areas of inquiry were proposed. The first inquiry examined data collected as part of a needs assessment to determine the overall prevalence rate of smoking and related behaviours for a population of individuals with severe mental illness receiving services from a community mental health agency. We found that the tobacco use prevalence was 72%, and 62% of smokers had high or very high levels of nicotine dependence; however almost half of respondents (47%) were interested in quitting or reducing tobacco within the next 6 months. Smokers were found to be over 9 times more likely to have a co-occurring substance use disorder (OR=9.44, 95%CI[6.33,14.08]). The second inquiry was a pilot study conducting a randomized controlled trial design to evaluate smoking cessation and smoking reduction outcomes for two groups of individuals (n=61) with severe mental illness receiving different smoking cessation interventions. Clients randomly assigned to the routine Smoking Cessation group (SC-R) received up to 24 weeks of no-cost Nicotine Replacement Therapy (NRT) and clients assigned to the Smoking Cessation Plus group (SC+) received up to 24 weeks of no-cost Nicotine Replacement Therapy (NRT) plus two initial individual sessions of motivational interviewing followed by weekly psychosocial group interventions for up to 24 weeks. Primary outcomes were levels of tobacco use at the 3-month and 6-month follow-up. The 7-day point prevalence abstinence rate measured at 3 months was 21.9% (n=7) for the SC+ group and 13.8% (n=4) for the SC-R group (OR=1.75,95%CI[.46,6.74]). At 6 months, the 7-day point prevalence abstinence rate was 12.5% (n=4) for the SC+ group and 6.9% (n=2) for the SC-R group (OR=1.93, 95%CI[.33,11.41]). Secondary outcomes included change in reported quality of life, physical health and mental health status functioning over the course of the study. We found that there were no statistically significant differences in the smoking quit or smoking reduction rates between the two treatment groups. At the 3-month time point the overall quit rate for both groups combined was 18% (n=11) and at the 6-month time point the quit rate was 10% (n=6). Reduction in the number of daily cigarettes smoked was statistically significant over time (F [1.68, 98.90] = 55.13, p < .001, η p 2 = 0.48) for both groups, as was the overall reduction of the FTND score (F [2, 94] = 17.98, p < .001, η p 2 = 0.28). This research demonstrates that collecting vital tobacco prevalence and dependency information is a straightforward and important task for community mental health agencies. Individuals with mental illness have both the interest and ability to quit or reduce their use of tobacco. Practitioners need to be aware of alternative smoking practices that may contribute to understanding tobacco use patterns and dependence in this population. Other factors such as co-morbid substance use disorder and level of community functioning may influence smoking status and, consequently, how treatment is provided. The findings of the pilot trial demonstrate the feasibility of conducting smoking cessation research with the population. Findings also suggest that a larger definitive trial is warranted to examine the effectiveness of the SC+ intervention. This research adds to the limited but growing knowledge base of how to address tobacco use and provide treatment to this vulnerable group, and will contribute to advances in population health by informing effective interventions with the attendant implications for program and policy development.
Ariana Prawda · 2010 · Rutgers University Community Repository (Rutgers University) · Open access
Relatively little research has been conducted on improving adherence to treatment for Latino patients. Given the changing demands in the field of Latino mental health and the difficulties of treatment engagement and treatment retention with Latinos, finding effective mental health interventions is of utmost importance. The following study analyzes Motivational Interviewing (MI) techniques that have been used with depressed Latinos in an effort to increase their adherence to antidepressant medication and improve symptoms of depression. The analysis consists of a series of pragmatic case studies, which are intended to capture the contextual factors that contribute to either the success or failure of the MI intervention in facilitating medication compliance and that contribute to the manner in which psychological interventions can be adapted to special populations. The MI intervention involved two sessions delivered within a two-week time period and one booster session delivered approximately two months later. Throughout their participation in the study, participants were monitored with regard to their level of depression, motivation to adhere to their antidepressive medication treatment, and actual compliance in taking their antidepressant medication. In total, three subjects from the research study’s database were selected for case study analysis: one who was found to have a positive outcome, one who was found to have a negative outcome, and one who was found to have mixed results with the MI intervention. The results suggest that while Motivational Interviewing has the potential to work as a treatment-enhancing intervention, its success in producing behavioral change largely depends on how well the techniques are adjusted to the individual’s stage of change.
The growing demand for scalable mental health support has increased interest in AI-based counseling systems grounded in Motivational Interviewing (MI).However, existing MI datasets do not explicitly model the structured progression of MI phases, which is essential for effective and goal-oriented counseling.To address this gap, we introduce PHASEMI, a phasestructured MI dataset, together with a data generation framework that employs therapist, client, and supervisor LLMs to explicitly control phase transitions.Compared to the best alternative baseline, PHASEMI achieves improved coverage of MI phases, with gains of 12.3% in Exploring, 37.6% in Guiding, and 61.1% in Choosing, and experimental evaluations demonstrate that it yields higher overall counseling quality than baseline datasets.
Todd Sage · 2019 · UND Scholarly Commons (University of North Dakota) · Open access
This research explored the use of a cognition primer to increase the perception of applicability of Motivational Interviewing for child welfare workers. Andragogy informed the need for cognition priming as a way to increase participants’ receptiveness to training by making it more applicable to their direct practice. The theory of implementation science was used to inform how organizational supports impede or enhance the likelihood of child welfare workers using Motivational Interviewing in their practice. A cross-sectional quasi-experimental mixed modal nonequivalent group design was used with a convenience sample of 41 front line child welfare workers from one Midwest urban county social services agency. A modified version of the Application Potential of Professional Learning Inventory (APPLI 31) (Curry, 2011) was used to measure applicability of training in a control group that received Motivational Interviewing training-as-usual compared to the intervention group that received the training along with a cognition primer. This study explored participant’s perception of the applicability of Motivational Interviewing, willingness to use Motivational Interviewing, and the personal and organizational factors that contribute to the adoption of Motivational Interviewing. Pre and post-intervention surveys were administered, and results were analyzed utilizing independent samples t-tests, multiple linear regression, and thematic analysis of the qualitative responses. The results of this study demonstrated that organizational supports and participants’ prior experiences with training increased the likelihood of adopting Motivational Interviewing. No differences were found between the control group that received training-as-usual and the intervention group that received the training with cognition priming. Quantitative and qualitative analysis revealed that child welfare workers see Motivational Interviewing applicability for their practice, but they do not feel equipped due to time constraints and a lack of system support to use this approach. Analysis from this research adds to the literature that organizational and supervisor supports are a key factor in the adoption of practice behaviors in child welfare agencies. Additionally, this research found that worker’s views related their perceived lack of time to use and implement Motivational Interviewing must be addressed as part of priming to overcome child welfare workers’ reluctance to implement Motivational Interviewing in their practice.
Rod Ali Mazloomi · 2023 · TSpace (University of Toronto) · Open access
The advent of large language models, such as BERT, GPT-3, and ChatGPT has galvanized research in chatbots, particularly in the context of addiction therapy. One such chatbot is MIBOT, developed through active collaboration between researchers at the University of Toronto and MI clinicians at the Centre for Addiction and Mental Health (CAMH). The goal of the MIBOT project is to develop a chatbot to emulate the principles of Motivational Interviewing in the context of smoking cessation therapy. While chatbots provide benefits in reaching underserved populations, there is a challenge in reviewing conversation transcripts due to the large number of conversations they may have. To address this challenge, topic modeling methods are analyzed in their ability to review transcripts. We find that topic modeling approaches help researchers (1) explore and summarize the MIBOT transcripts; (2) plan appropriate responses in an M-Iinspired conversation; and (3) stimulate further research and question generation.
Robert A. Lyon · 2014 · UWE Research Repository (UWE Bristol) · Open access
The treatment and recovery of forensic psychiatric inpatients can take many months and often years, within which time health issues such as obesity and diabetes can develop. Encouraging beneficial health behaviours with this client group is notoriously difficult for a variety of reasons and anti-psychotic medication paradoxically often serves to exacerbate physical health concerns. There is a dearth of research concerning effective interventions for improved physical health for this client group, perhaps because security and staffing issues present significant challenges to conducting research in this environment. \n \nThis exploratory piece of action research employed multi-staged Motivational Interviewing (MI) sessions in nine case studies, the aim of which was to explore the utility of the approach in increasing physical activity (PA) with forensic psychiatric inpatients. Changes in PA and intrinsic motivation were assessed over three months and clients were surveyed on their views of the approach. A question exists over the propriety of client-centred care in a setting that is often necessarily restrictive and controlled. Challenges to a broader implementation of MI for health promotion amongst the multi-disciplinary care team (MDT) are considered, and were explored through the use of a staff survey. Outcomes for the research suggest that MI is a useful and valued approach to facilitating changes in physical activity levels with forensic psychiatric clients. There may be others from this environment for whom the approach is inappropriate, and there is a need for further research with clients whose health concerns are significant but who do not readily present for activity sessions. Outcomes from this research further suggest some of the MDT may be philosophically aligned with some aspects of the client-centred ethos of MI, yet still maintain a belief in the propriety of directive and authoritarian methods for promoting health. This may present a challenge to the training of MI if a broader implementation of MI for physical health issues is considered, and limited survey data may not have revealed the full extent of this challenge.
Sophia Joseph Parrilla · 2016 · ScholarWorks (Walden University) · Open access
Adolescents with acting-out behavior have an increased risk of dropping out of high school, incarceration, and early mortality. Researchers have indicated the need for studies using motivational interviewing (MI) to promote behavioral change. There is a gap in the literature about the efficacy of MI with adolescents and clinicians' knowledge of, and attitude toward, MI. Using self-determination theory as the grounding theory, this study examined the knowledge and attitudes of 73 clinicians on the island of St. Croix, Virgin Islands, about knowledge and attitudes of MI and whether it predicted the use of MI as a therapeutic intervention for adolescents who exhibit acting-out behavior. A cross-sectional survey was used to collect data using an adapted version of Leffingwell's Motivational Interviewing Knowledge and Attitudes Test (MIKAT). Two phases were required: a pilot study and a full study. A pilot study of 10 clinicians was used to establish the reliability of the revised MIKAT. In total a purposive sample of 73 clinicians participated in the full study, which includes the 10 from the pilot study. Results of the multiple linear regression test indicated that knowledge about MI and attitudes towards MI were not significant predictors of likelihood to use MI (p = .875). The results of this study may contribute to positive social change by supporting the development of effective training for clinicians who work with adolescents on St. Croix, where adolescent behavior is of great concern.
Rashid W. Flewellen · 2016 · eSpace (Curtin University) · Open access
Alcohol-related injury is a significant public health challenge particularly in Emergency Departments (ED). Staff responses are invariably influenced by a number of factors, that include perceptions of people affected by alcohol, confidence in the ability to respond and demands on limited resources. This thesis aimed to identify attitudes and beliefs associated with responses to alcohol related injury (ARI). The research objective was to answer the following question: “What factors influence an evidence-based response for ED clinicians managing such injuries?” Diverse factors such as low fidelity within intervention implementation, and misconceptions about the role of public health in emergency medicine as well as findings from this study, will contribute to the ability to understand the response.
Peter Sachs Collopy · 2015 · ScholarlyCommons (University of Pennsylvania) · Open access
In the late 1960s, video recorders became portable, leaving the television studio for the art gallery, the psychiatric hospital, and the streets. The technology of recording moving images on magnetic tape, previously of use only to broadcasters, became a tool for artistic expression, psychological experimentation, and political revolution. Video became portable not only materially but also culturally; it could be carried by an individual, but it could also be carried into institutions from the RAND Corporation to the Black Panther Party, from psychiatrists’ offices to art galleries, and from prisons to state-funded media access centers. Between 1967 and 1973, American videographers across many of these institutional contexts participated in a common discourse, sharing not only practical knowledge about the uses and maintenance of video equipment, but visions of its social significance, psychological effects, and utopian future. For many, video was a technology which would bring about a new kind of awareness, the communal consiousness that—influenced by the evolutionary philosophy of Henri Bergson—Pierre Teilhard de Chardin referred to as the noosphere and Marshall McLuhan as the global village. Experimental videographers across several fields were also influenced by the psychedelic research of the 1950s and early 1960s, by the development of cybernetics as a science of both social systems and interactions between humans and machines, by anthropology and humanistic psychology, and by revolutionary political movements in the United States and around the world.