Crisis prediction among tele-mental health patients: A large language model and expert clinician comparison.

BACKGROUND: Due to recent advances in artificial intelligence (AI), large language models (LLMs) have emerged as a powerful tool for a variety of language related tasks, including sentiment analysis, and summarization of provider-patient interactions. However, there is limited research on these models in the area of crisis prediction. OBJECTIVE: This study aimed to evaluate the performance of LLMs, specifically OpenAI's GPT-4, in predicting current and future mental health crisis episodes using patient provided information at intake among users of a national telemental health platform. METHODS: De-identified patient provided data was pulled from specific intake questions of the Brightside telehealth platform, including the chief complaint, for 140 patients who indicated suicidal ideation (SI), and another 120 patients who later indicated SI with a plan during the course of treatment. Similar data was pulled for 200 randomly selected patients treated during the same time period who never endorsed SI. Six senior Brightside clinicians (three psychologists and three psychiatrists) were shown patients' self-reported chief complaint and self-reported suicide attempt history but were blinded to the future course of treatment and other reported symptoms including SI. They were asked a simple yes/no question regarding their prediction of endorsement of SI with plan along with their confidence level about the prediction. GPT-4 was provided similar information and asked to answer the same questions, enabling us to directly compare the performance of AI and clinicians. RESULTS: Overall, clinicians' average precision (0.698) was higher than GPT-4 (0.596) in identifying SI with plan at intake (n=140) vs. no SI (n=200) when using chief complaint alone, while sensitivity was higher for GPT-4 (0.621) than clinicians' average (0.529). The addition of suicide attempt history increased clinicians' average sensitivity (0.590) and precision (0.765), while increasing GPT-4 sensitivity (0.590) but decreasing GPT-4 precision (0.544). Performance decreased comparatively when predicting future SI with plan (n=120) vs no SI (n=200) with chief complaint only for clinicians (average sensitivity=0.399; average precision=0.594) and GPT-4 (sensitivity=0.458; precision=0.482). The addition of suicide attempt history increased performance comparatively for clinicians (average sensitivity=0.457; average precision=0.687) and GPT-4 (sensitivity=0.742; precision=0.476). CONCLUSIONS: GPT-4 with a simple prompt design produced results on some metrics that approached that of a trained clinician. Additional work must be done before such a model could be piloted in a clinical setting. The model should undergo safety checks for bias given evidence that LLMs can perpetuate the biases of the underlying data they are trained upon. We believe that LLMs hold promise to augment identification of higher risk patients at intake and potentially deliver more timely care to patients.

The utility of completing adjuvant video lessons based on the unified protocol during psychotherapy: A retrospective study using a telehealth platform in routine clinical care.

Digital supplements to tele-psychotherapy are increasingly needed. The purpose of this retrospective study was to investigate the association between outcomes and the use of supplemental video lessons based on the Unified Protocol (UP), an empirically supported transdiagnostic treatment.Participants included 7,326 adults in psychotherapy for depression and/or anxiety. Partial correlations were calculated between number of UP video lessons completed and change in outcomes after 10 weeks, controlling for number of therapy sessions and baseline scores. Then, participants were divided into those who did not complete any UP video lessons (n = 2355) and those who completed at least 7/10 video lessons (n = 549), and propensity-matched on 14 covariates. Repeated measures analysis of variance compared these groups (n = 401 in each group) on outcomes.Among the entire sample, symptom severity decreased as the number of UP video lessons completed increased, with the exception of lessons on avoidance and exposure. Those watching at least 7 lessons showed significantly greater reduction in both depression and anxiety symptoms than those who did not watch any.Viewing supplemental UP video lessons in addition to tele-psychotherapy had a positive and significant association with symptom improvement and may provide an additional tool for clinicians to implement UP components virtually.

Novel Telehealth Adaptations for Evidence-Based Outpatient Suicide Treatment: Feasibility and Effectiveness of the Crisis Care Program.

Background: Suicide rates in the United States have escalated dramatically over the past 20 years and remain a leading cause of death. Access to evidenced-based care is limited, and telehealth is well-positioned to offer novel care solutions. The Crisis Care program is a suicide-specific treatment program delivered within a national outpatient telehealth setting using a digitally adapted version of the Collaborative Assessment and Management of Suicidality (CAMS) as the framework of care. This study investigates the feasibility and preliminary effectiveness of Crisis Care as scalable suicide-specific treatment model. Methods: Patient engagement, symptom reduction, and care outcomes were examined among a cohort of patients (n = 130) over 16 weeks. The feasibility of implementation was assessed through patient engagement. Clinical outcomes were measured with PHQ-9, GAD-7, and the CAMS SSF-4 rating scales. Results: Over 85% of enrolled patients were approved for Crisis Care at intake, and 83% went on to complete at least four sessions (the minimum required to graduate). All patient subgroups experienced declines in depressive symptoms, anxiety symptoms, suicidal ideation frequency, and suicide-specific risk factors. Conclusions: Results support the feasibility and preliminary effectiveness of Crisis Care as a suicide-specific care solution that can be delivered within a stepped-care model in an outpatient telehealth setting.

Consumer Expectations and Attitudes About Psychotherapy: Survey Study.

BACKGROUND: Although mental illness is common among adults in the United States, access to, and public perception of, mental health care continue to present as key barriers to care. OBJECTIVE: Given the importance of attitudes toward, and perceptions of, mental health treatment in the successful access to mental health care and treatment of mental health issues, the primary goal of this survey study was to further investigate consumer perspectives of psychotherapy among adults in the United States; specifically, adding to the literature by investigating perceptions of both the general public and patients receiving telehealth. More specifically, the aims were to better understand openness to, and satisfaction with, therapy; perceptions, preferences, and expectations around therapy; and perceptions of psychotropic medication. METHODS: An electronic survey was administered to current and former patients (those receiving psychotherapy) of Brightside, a nationwide telehealth company, as well as to the general public; both were convenience samples. Using the same survey questions, Brightside surveyed its own members (using Qualtrics; Qualtrics International Inc) and the general population (using SurveyMonkey's Audience tool; Momentive). This survey included questions about basic participant demographics, as well as questions about current mental health treatment, perceptions about therapy, and therapists' qualities. RESULTS: A total of 714 people completed the survey. The data were fairly evenly split between those collected from Brightside patients (368/714, 51.5%) and those collected from the general public (346/714, 48.5%). Combining both samples, overall participation was 67.1% (479/714) women; 73.1% (522/714) White, 7.3% (52/714) Asian, 6.7% (48/714) African American, and 7.4% (53/714) Hispanic or Latinx; largely aged 25 to 34 years (255/714, 35.7%) or 35 to 44 years (187/714, 26.2%); from either the Mid-Atlantic (131/714, 18.3%) or South Atlantic (129/714, 18.1%) regions of the country; and most (402/714, 56.3%) earning annual salaries of US $30,000 to US $100,000. There were generally favorable perceptions of both psychotherapy and psychiatric medication. Selecting a therapist as well as cost and insurance are the common factors in therapy that are important to patients. The most commonly held perception of psychotherapy duration was indefinite (250/714, 35%). Very few (58/714, 8.1%) thought that therapy typically lasts 1 to 3 months. Most of the participants (414/714, 58%) thought that evidence-based practice was important. CONCLUSIONS: Public education is needed to increase awareness of the typical duration and cost of psychotherapy. There seem to be generally favorable perceptions of both psychotherapy and psychotropic medication. Selecting a therapist as well as cost and insurance are the common factors in therapy that are important to patients. Practitioners and those marketing their services might consider using their marketing campaigns to counter some of the more common falsely held beliefs.

Symptom clustering of major depression in a national telehealth sample.

BACKGROUND: Major depressive disorder (MDD) is a heterogeneous disorder whose possible symptom combinations have not been well delineated. The aim of this study was to explore the heterogeneity of symptoms experienced by those with MDD to characterize phenotypic presentations. METHODS: Cross-sectional data (N = 10,158) from a large telemental health platform were used to identify subtypes of MDD. Symptom data, gathered from both clinically-validated surveys and intake questions, were analyzed via polychoric correlations, principal component analysis, and cluster analysis. RESULTS: Principal components analysis (PCA) of baseline symptom data revealed 5 components, including anxious distress, core emotional, agitation/irritability, insomnia, and anergic/apathy components. PCA-based cluster analysis resulted in four MDD phenotypes, the largest of which was characterized by a prominent elevation on the anergic/apathy component, but also core emotional. The four clusters differed on demographic and clinical characteristics. LIMITATIONS: The primary limitation of this study is that the phenotypes uncovered are limited by the questions asked. These phenotypes will need to be cross validated with other samples, potentially expanded to include biological/genetic variables, and followed longitudinally. CONCLUSIONS: The heterogeneity in MDD, as illustrated by the phenotypes in this sample, may explain the heterogeneity of treatment response in large-scale treatment trials. These phenotypes can be used to study varying rates of recovery following treatment and to develop clinical decision support tools and artificial intelligence algorithms. Strengths of this study include its size, breadth of included symptoms, and novel use of a telehealth platform.

Does Bupropion Increase Anxiety?: A Naturalistic Study Over 12 Weeks.

PURPOSE/BACKGROUND: There has long been a clinical belief that bupropion exacerbates anxiety. The purpose of the current retrospective study is to compare anxiety severity over time in those prescribed selective serotonin reuptake inhibitors (SSRIs) versus bupropion. METHODS/PROCEDURES: Archival data (N = 8457) from patients receiving psychiatric care from a national tele-mental health company were used. Propensity matching was used to create SSRI and bupropion groups using 17 covariates. These samples were then compared using repeated measures analysis of variance on Generalized Anxiety Disorder Scale 7 scores at start of treatment, 6 weeks, and 12 weeks. FINDINGS/RESULTS: The SSRI and bupropion groups were significantly different across a number of variables. In the entire sample, the bupropion group had significantly greater anxiety levels. However, for propensity-matched comparisons, there were no significant interactions between group and time (ie, groups did not differ and improved comparably over time). IMPLICATIONS/CONCLUSIONS: Using propensity matching, there were no differences in anxiety outcome between those prescribed selective serotonin reuptake inhibitor versus bupropion across 12 weeks of treatment.

Early Response to Antidepressant Medications in Adults With Major Depressive Disorder: A Naturalistic Study and Odds of Remission at 14 Weeks.

PURPOSE/BACKGROUND: Early response after 2 to 4 weeks of antidepressant therapy has been shown to predict remission by 8 to 12 weeks. Most of the work to date on early response has been done using data from randomized controlled trials. METHODS/PROCEDURES: This naturalistic study uses archival data from a national tele-mental health company. The positive and negative predictive values as well as sensitivity and specificity were calculated using different drops in baseline Patient Health Questionnaire 9 scores at various periods. Demographic and clinical characteristics were compared between early responders versus those lacking early response. Binary logistic regression analyses determined if early response was predictive of remission, response, and greater than minimal improvement at 14 weeks. For those who do not show early improvement, treatments were investigated using binary logistic regression to see if changes predicted later outcomes. FINDINGS/RESULTS: Positive predictive values for all endpoints improved with the strength of early response but did not improve much with the time allowed for that response to occur. In contrast, negative predictive values increased substantially with time. Using a definition of 30% drop in Patient Health Questionnaire 9 score at week 4, 56.5% of patients were early responders. Early responders were ~3.2 times more likely to achieve remission than those lacking early response. Of nonresponders by week 4, those prescribed atypical antipsychotics (+SSRI) had significantly reduced odds of response at week 14, whereas those prescribed a norepinephrine and dopamine reuptake inhibitor had increased odds. IMPLICATIONS/CONCLUSIONS: Early response may be associated with better outcomes at 14 weeks. In those with lack of response by week 4, patients prescribed a norepinephrine and dopamine reuptake inhibitor may achieve superior outcomes.

Telehealth-Supported Decision-making Psychiatric Care for Suicidal Ideation: Longitudinal Observational Study.

BACKGROUND: Suicide is a leading cause of death in the United States, and suicidal ideation (SI) is a significant precursor and risk factor for suicide. OBJECTIVE: This study aimed to examine the impact of a telepsychiatric care platform on changes in SI over time and remission, as well as to investigate the relationship between various demographic and medical factors on SI and SI remission. METHODS: Participants included 8581 US-based adults (8366 in the treatment group and 215 in the control group) seeking treatment for depression, anxiety, or both. The treatment group included patients who had completed at least 12 weeks of treatment and had received a prescription for at least one psychiatric medication during the study period. Providers prescribed psychiatric medications for each patient during their first session and received regular data on participants. They also received decision support at treatment onset via the digital platform, which leveraged an empirically derived proprietary precision-prescribing algorithm to give providers real-time care guidelines. Participants in the control group consisted of individuals who completed the initial enrollment data and completed surveys at baseline and 12 weeks but did not receive care. RESULTS: Greater feelings of hopelessness, anhedonia, and feeling bad about oneself were most significantly correlated (r=0.24-0.37) with SI at baseline. Sleep issues and feeling tired or having low energy, although significant, had lower correlations with SI (r=0.13-0.14). In terms of demographic variables, advancing age and education were associated with less SI at baseline (r=-0.16) and 12 weeks (r=-0.10) but less improvement over time (r=-0.12 and -0.11, respectively). Although not different at baseline, the SI expression was evident in 34.4% (74/215) of the participants in the control group and 12.32% (1031/8366) of the participants in the treatment group at 12 weeks. Although the participants in the treatment group improved over time regardless of various demographic variables, participants in the control group with less education worsened over time, after controlling for age and depression severity. A model incorporating the treatment group, age, sex, and 8-item Patient Health Questionnaire scores was 77% accurate in its classification of complete remission. Those in the treatment group were 4.3 times more likely (odds ratio 4.31, 95% CI 2.88-6.44) to have complete SI remission than those in the control group. Female participants and those with advanced education beyond high school were approximately 1.4 times more likely (odds ratio 1.38, 95% CI 1.18-1.62) to remit than their counterparts. CONCLUSIONS: The results highlight the efficacy of an antidepressant intervention in reducing SI, in this case administered via a telehealth platform and with decision support, as well as the importance of considering covariates, or subpopulations, when considering SI. Further research and refinement, ideally via randomized controlled trials, are needed.

Do older adults benefit from telepsychiatric care: Comparison to younger adults.

Background: Telemental health platforms may increase access to care for older adults. Historically, older adults have tended to adopt new technologies at a slower rate which creates a perception that they may not be able to benefit from them. The purpose of this study was to determine whether or not older adult patients receiving psychiatric care for depression via a telemental health platform achieve the same outcomes as younger adults. Method: Participant data utilized in the current investigation were obtained from a national mental health telehealth company (i.e., Brightside) and consisted of 12,908 U.S.-based adult patients receiving psychiatric care for depression between October, 2018 and January, 2022. Propensity matching was used to create an older and younger sample (n = 141 in each) using 23 covariates. These samples were then compared using repeated measures ANOVA on Patient Health Questionnaire-9 (PHQ-9) scores at start of treatment, 6 weeks, 8 weeks, 10 weeks, 12 weeks, 14 weeks, and 16 weeks. Results: Despite matching, the groups still significantly differed on prior mental health treatment, such that more older adults reported having had prior mental health treatment. There were no other differences between the groups on assessed variables. Both younger and older adults had decreasing scores over time with no significant differences between them. Conclusion: Older adults have similar improvement in depression symptom severity over time following initiation of psychiatric treatment via a telehealth platform. These findings suggest that age is not a barrier to benefitting from telepsychiatric care.

Feasibility and acceptability of a novel telepsychiatry-delivered precision prescribing intervention for anxiety and depression.

BACKGROUND: Major Depressive Disorder and Generalized Anxiety Disorder are pervasive and debilitating conditions, though treatment is often inaccessible and based on trial-and-error prescribing methods. The present observational study seeks to describe the use of a proprietary precision prescribing algorithm piloted during routine clinical practice as part of Brightside's telepsychiatry services. The primary aim is to determine the feasibility and acceptability of implementing this intervention. Secondary aims include exploring remission and symptom improvement rates. METHODS: Participants were adult patients enrolled in Brightside who completed at least 12 weeks of treatment for depression and/or anxiety and received a prescription for at least one psychiatric medication. A prescription recommendation was made by Brightside's algorithm at treatment onset and was utilized for clinical decision support. Participants received baseline screening surveys of the PHQ-9 and GAD-7, and at weeks 2,4,6,8,10 and 12. Intent-to-treat (ITT) sensitivity analyses were conducted. Feasibility of the implementation was measured by the platform's ability to enroll and engage participants in timely psychiatric care, as well as offer high touch-point treatment options. Acceptability was measured by patient responses to a 5-star satisfaction rating. RESULTS: Brightside accessed and treated 6248 patients from October 2018 to April 2021, treating a majority of patients within 4-days of enrollment. The average plan cost was $115/month. 89% of participants utilized Brightside's core medication plan at a cost of $95/month. 13.4% of patients in the study rated Brightside's services as highly satisfactory, averaging a 4.6-star rating. Furthermore, 90% of 6248 patients experienced a MCID in PHQ-9 or GAD-7 score. Remission rates were 75% (final PHQ-9 or GAD-7 score < 10) for the study sample and 59% for the ITT sample. 69.3% of Brightside patients were treated with the medication initially prescribed at intake. CONCLUSIONS: Results suggest that the present intervention may be feasible and acceptable within the assessed population. Exploratory analyses suggest that Brightside's course of treatment, guided by precision recommendations, improved patients' symptoms of anxiety and depression.

Improving Well-being With a Mobile Artificial Intelligence-Powered Acceptance Commitment Therapy Tool: Pragmatic Retrospective Study.

BACKGROUND: Research and dissemination of smartphone apps to deliver coaching and psychological driven intervention had seen a great surge in recent years. Notably, Acceptance Commitment Therapy (ACT) protocols were shown to be uniquely effective in treating symptoms for both depression and anxiety when delivered through smartphone apps. The aim of this study is to expand on that work and test the suitability of artificial intelligence-driven interventions delivered directly through popular texting apps. OBJECTIVE: This study evaluated our hypothesis that using Kai.ai will result in improved well-being. METHODS: We performed a pragmatic retrospective analysis of 2909 users who used Kai.ai on one of the top messaging apps (iMessage, WhatsApp, Discord, Telegram, etc). Users' well-being levels were tracked using the World Health Organization-Five Well-Being Index throughout the engagement with service. A 1-tailed paired samples t test was used to assess well-being levels before and after usage, and hierarchical linear modeling was used to examine the change in symptoms over time. RESULTS: The median well-being score at the last measurement was higher (median 52) than that at the start of the intervention (median 40), indicating a significant improvement (W=2682927; P<.001). Furthermore, HLM results showed that the improvement in well-being was linearly related to the number of daily messages a user sent (ß=.029; t81.36=4; P<.001), as well as the interaction between the number of messages and unique number of days (ß=-.0003; t81.36=-2.2; P=.03). CONCLUSIONS: Mobile-based ACT interventions are effective means to improve individuals' well-being. Our findings further demonstrate Kai.ai's great promise in helping individuals improve and maintain high levels of well-being and thus improve their daily lives.

Exploring social determinants of health: Comparing lower and higher income individuals participating in telepsychiatric care for depression.

Background: Telemental health may increase access to care; there has been little research on efficacy with those at the lower end of the income distribution. The purpose of this study was to determine whether lower vs. higher income patients receiving telepsychiatric care for depression achieve: (1) effective symptom reduction and (2) similar outcomes. Methods: Data utilized were obtained from a national mental health telehealth company and consisted of 5,426 U.S.-based patients receiving psychiatric care for moderate to severe depression between October, 2018 and January, 2022. Propensity matching was used to create lower and higher income samples (n = 379 in each) using 22 covariates. These samples were then compared using repeated measures ANOVA on Patient Health Questionnaire-9 (PHQ-9) scores at start of treatment, 6, 8, 10, 12, 14, and 16 weeks. Results: Both lower and higher income groups made significant improvement over time, with groups averaging mild symptom severity by week 16. There was a significant group x time interaction, such that the lower income group had significantly greater depression severity at the last two timepoints. Conclusion: Lower and higher income groups both made significant improvement in depression symptom severity over time following initiation of psychiatric treatment via a telehealth platform, though higher income individuals, all else being equal besides employment, tend to do better. These findings suggest that when lower income individuals do participate in care, good outcomes can be achieved. Further research is needed to better understand the role social determinants of health (SDOH) play in outcome disparities.

Adolescents' Well-being While Using a Mobile Artificial Intelligence-Powered Acceptance Commitment Therapy Tool: Evidence From a Longitudinal Study.

BACKGROUND: Adolescence is a critical developmental period to prevent and treat the emergence of mental health problems. Smartphone-based conversational agents can deliver psychologically driven intervention and support, thus increasing psychological well-being over time. OBJECTIVE: The objective of the study was to test the potential of an automated conversational agent named Kai.ai to deliver a self-help program based on Acceptance Commitment Therapy tools for adolescents, aimed to increase their well-being. METHODS: Participants were 10,387 adolescents, aged 14-18 years, who used Kai.ai on one of the top messaging apps (eg, iMessage and WhatsApp). Users' well-being levels were assessed between 2 and 5 times using the 5-item World Health Organization Well-being Index questionnaire over their engagement with the service. RESULTS: Users engaged with the conversational agent an average of 45.39 (SD 46.77) days. The average well-being score at time point 1 was 39.28 (SD 18.17), indicating that, on average, users experienced reduced well-being. Latent growth curve modeling indicated that participants' well-being significantly increased over time (ß=2.49; P<.001) and reached a clinically acceptable well-being average score (above 50). CONCLUSIONS: Mobile-based conversational agents have the potential to deliver engaging and effective Acceptance Commitment Therapy interventions.

Gabapentin augmentation therapy in bipolar depression.

BACKGROUND: Gabapentin (GBP) may be useful in bipolar disorders, including as adjunctive therapy for bipolar depression, although controlled studies suggest inefficacy as primary treatment for mania or treatment-resistant rapid cycling. METHODS: We performed a 12-week trial of open GBP (mean dose 1725 mg/day) added to stable doses of mood stabilizers or atypical antipsychotics in 22 (10 women, mean age 38.4 years) depressed (28-item Hamilton Depression Rating Scale (HDRS) > 18] bipolar (10 bipolar I, 12 bipolar II) disorder outpatients. Mean illness duration was 18.6 years, current depressive episode duration was 18.0 weeks. Prospective 28-item HDRS, Young Mania Rating Scale (YMRS) and Clinical Global Impression-Severity (CGI-S) ratings were obtained. RESULTS: Overall, HDRS ratings decreased 53% from 32.5 +/- 7.7 at baseline to 16.5 +/- 12.8 at week 12 (p < 0.0001). Twelve of 22 (55%) patients had moderate to marked improvement (HDRS decrease = 50%) with HDRS decreasing 78% from 27.9 +/- 6.2 to 6.2 +/- 4.5 (p < 0.0001). Eight of 22 (36%) patients remitted (HDRS > or = 8). In non-responders, HDRS decreased from 38.0 +/- 5.4 to 28.9 +/- 6.7 (p = 0.005). Ten of 13 (77%) mild to moderately depressed (baseline HDRS > 18 and <35) patients responded, while only two of nine patients (22%) with severe depression (HDRS > or = 35) responded (p < 0.03). Both groups, however, had similar, statistically significant HDRS decreases. GBP was well tolerated. CONCLUSION: Open adjunctive GBP was effective and well tolerated in patients with mild to moderate bipolar depression. This open pilot study must be viewed with caution, and randomized controlled studies are warranted.