Impact of Clinician Virtual Human Interaction Training in Emotional Self-Awareness on Patients Suicidal Ideation and Suicide Crisis Syndrome

Impact of Clinician Virtual Human Interaction Training in Emotional Self-Awareness on Patients Suicidal Ideation and Suicide Crisis Syndrome

Impact of Clinician Virtual Human Interaction Training in Emotional Self-Awareness on Patients Suicidal Ideation and Suicide Crisis Syndrome: a Randomized Controlled Trial

Working with patients at risk for suicide is highly stressful for clinicians and often elicits powerful negative emotional responses that may adversely affect suicidal outcomes. A proposed explanation has been that negative emotional responses may result in less empathic communication and unwitting rejection of the patient, which is liable to damage therapeutic alliance. Thus, there is a need for clinician training in effective management of negative emotions towards suicidal patients, which can result in tangible improvement in suicidal outcomes. The training must be web-based, scalable and easy to disseminate, making it available to clinicians everywhere. In this project, the study team will address this critical need by using Virtual Human Interaction (VHI) to train outpatient clinicians in emotional self-awareness (ESA), which includes both recognition of one's own negative emotional responses, and ability to engage in verbal empathic communication with acutely suicidal patients. The study team will conduct a prospective multisite, blinded, randomized trial comparing VHI ESA training with a Control condition, which will assess both clinician-level and patient-level outcomes in 80 outpatient clinician participants (CPs) and 400 of their participating patients (PPs). Using the same VHI scenarios, the ESA group will receive clinician-focused, comprehensive feedback in ESA, while the Control group will assess the VH's suicide risk without receiving the ESA feedback. The study team will measure efficacy of the VHI ESA training on clinicians by comparing ESA feedback and Control CPs' post-training (T2) ESA towards virtual humans, adjusting for pre-training (T1) ESA. The study team will measure the impact of the VHI ESA training on patients' suicidal outcomes by assessing the primary and secondary outcome variables in both CPs and PPs three times: at baseline (T0), after the first post-training treatment session (T3) and one-month post-training (T4). The study team will examine the role of therapeutic alliance as a possible mediator of the relationship between clinicians' ESA and their patients' Suicidal Ideation (SI) and Suicide Crisis Syndrome (SCS). To accomplish this goal, the study team will use the novel validated suicide risk assessment instruments developed in preliminary studies: the Therapist Response Questionnaire - Suicide Form (TRQ-SF), which assesses negative emotional responses to suicidal patients, and the Suicide Crisis inventory (SCI), which assesses the SCS severity and predicts near-term suicidal behavior. For web-based VHI training the study team will use the already tested and disseminated web-based empathy-teaching platform, coupled with an assessment of verbal empathy measured by the Empathic Communication Coding System (ECCS).

The study has three interconnected Specific Aims. In Aim 1 the study team will determine if, VHI ESA clinician training improves ESA when interacting with virtual humans. The study team hypothesizes that clinicians in the ESA group will report less negative emotional responses and greater empathic communication at the post-training (T2) assessment than clinicians in the Control group, adjusting for pre-treatment (T1) scores. In Aim 2 the study team will establish if, VHI ESA clinician training results in sustained improvement in patients' suicidal ideation and SCS. The study team's primary patient level outcomes will be SI as measured by the Beck Suicide Scale, and SCS severity as measured by the SCI. The study team's secondary outcomes will be treatment retention, patients' perception of the therapeutic alliance as measured by the Working Alliance Inventory, patients' perception of clinician empathy as measured by the Consultational and Relational Empathy scale, treatment adherence as measured by the Morisky Medication Adherence Scale, and suicidal behaviors as measured by the Columbia Suicide Severity Rating Scale. The study team hypothesizes that patient participants in the ESA group will report less severe SI and SCS at the first post-training clinical encounter with their clinician (T3) than the Control group, adjusting for baseline (T0) levels of SI and SCS, and that these differences will extend to the final post-training follow-up (T4). In Aim 3 the study team will assess the role of therapeutic alliance in improving SI and SCS. The study team hypothesizes that patients' therapeutic alliance at post-training assessments T3 and T4 will mediate the relationship between post-training ESA indices and post-training patient suicidal outcomes, adjusting for baseline (T0) levels of the outcome. If successful, the proposed work may have a broad impact on clinician training and practice by adding an ESA dimension to evidence-based suicide risk assessment. Suicide is preventable and the ultimate goal is to create an industry-standard platform to train clinicians in ESA, thereby improving their skills in interacting with suicidal patients, and, ultimately, saving lives.

Two experimental groups will complete the same Virtual Human Interaction (VHI) scenarios. Only the Emotional Self-Awareness (ESA) group will also receive clinician-focused, comprehensive feedback in ESA, while the Control group will assess the VH's suicide risk without receiving the ESA feedback. The study team will measure the VHI ESA training effectiveness by comparing clinician participants' (CPs) ESA towards virtual humans at post-training (T2), adjusting for pre-training (T1) levels. The study team will measure the impact of the VHI ESA on patient participants' (PPs) suicidal outcomes by assessing the primary and secondary outcome variables in both CPs and PPs three times: at baseline (T0), after the first post-training treatment session (T3) and after the final treatment session, one month post-training (T4). The study team will examine the role of therapeutic alliance as a possible mediator of the relationship between clinicians' ESA and their patients' SI and SCS.

The study will follow a double-blind procedure to minimize the effects of expectation bias. Prior to the start of the study, the Co-Investigators will assign different letters to the ESA training and control group; they will not reveal which letter corresponds to which condition to the coordinators and research assistants until the completion of the study. The Co-Investigators at each site will carry out the randomization procedure and record this information in a password-protected file. Following the baseline T0 assessments but prior to the T1 VHI, the Co-Investigators will inform the coordinators which condition each CP has been assigned to. Although the CPs will be aware that they are receiving feedback on their ESA ratings, they will not know that this is the active component of the training. To ensure PP blinding, CPs will be asked not to discuss the VHI experience with their colleagues or patients to minimize the potential that they will figure out which condition they are in.

Virtual Human Interaction (VHI) to train outpatient clinicians in emotional self-awareness (ESA) and receive clinician-focused, comprehensive feedback in ESA

The Control group CPs will engage in the same two VHI scenarios and will complete the TRQ-SF in response to each scenario without receiving the ESA feedback

Virtual Human Interaction (VHI) to train outpatient clinicians in emotional self-awareness (ESA) and receive clinician-focused, comprehensive feedback in ESA

A 49-item self-report instrument for the assessment of the severity of the Suicidal Crisis Syndrome (SCS), an acute state shown to correlate with a near-term suicide attempt. Items are rated by self-report on a five-point scale ranging from 0=not at all to 4=extremely. Full scale from 0-196, with higher score indicating increased suicide risk.

A 49-item self-report instrument for the assessment of the severity of the Suicidal Crisis Syndrome (SCS), an acute state shown to correlate with a near-term suicide attempt. Items are rated by self-report on a five-point scale ranging from 0=not at all to 4=extremely. Full scale from 0-196, with higher score indicating increased suicide risk.

A 49-item self-report instrument for the assessment of the severity of the Suicidal Crisis Syndrome (SCS), an acute state shown to correlate with a near-term suicide attempt. Items are rated by self-report on a five-point scale ranging from 0=not at all to 4=extremely. Full scale from 0-196, with higher score indicating increased suicide risk.

A well validated 21-item self-report measure of active and passive suicidal desires will be used to assess severity of patient SI pre- and post-training. Scores on each item range from 0 to 2, full scale from 0-42, with higher scores indicating greater severity of SI.

A well validated 21-item self-report measure of active and passive suicidal desires will be used to assess severity of patient SI pre- and post-training. Scores on each item range from 0 to 2, full scale from 0-42, with higher scores indicating greater severity of SI.

A well validated 21-item self-report measure of active and passive suicidal desires will be used to assess severity of patient SI pre- and post-training. Scores on each item range from 0 to 2, full scale from 0-42, with higher scores indicating greater severity of SI.

10-item, Likert-type scale designed to capture clinicians' emotional responses to acutely suicidal patients. TRQ-SF individual item scores range from 0 (Not at all) to 4 (Extremely). The scale probes suicide-related negative emotional response to patients with three sub-scales: Affiliation (5 items), Distress (3 items), and Hopefulness (2 items). The potential range for TRQ total scores is 0-40. Higher score indicate higher risk of suicide.

10-item, Likert-type scale designed to capture clinicians' emotional responses to acutely suicidal patients. TRQ-SF individual item scores range from 0 (Not at all) to 4 (Extremely). The scale probes suicide-related negative emotional response to patients with three sub-scales: Affiliation (5 items), Distress (3 items), and Hopefulness (2 items). The potential range for TRQ total scores is 0-40. Higher score indicate higher risk of suicide.

10-item, Likert-type scale designed to capture clinicians' emotional responses to acutely suicidal patients. TRQ-SF individual item scores range from 0 (Not at all) to 4 (Extremely). The scale probes suicide-related negative emotional response to patients with three sub-scales: Affiliation (5 items), Distress (3 items), and Hopefulness (2 items). The potential range for TRQ total scores is 0-40. Higher score indicate higher risk of suicide.

10-item, Likert-type scale designed to capture clinicians' emotional responses to acutely suicidal patients. TRQ-SF individual item scores range from 0 (Not at all) to 4 (Extremely). The scale probes suicide-related negative emotional response to patients with three sub-scales: Affiliation (5 items), Distress (3 items), and Hopefulness (2 items). The potential range for TRQ total scores is 0-40. Higher score indicate higher risk of suicide.

10-item, Likert-type scale designed to capture clinicians' emotional responses to acutely suicidal patients. TRQ-SF individual item scores range from 0 (Not at all) to 4 (Extremely). The scale probes suicide-related negative emotional response to patients with three sub-scales: Affiliation (5 items), Distress (3 items), and Hopefulness (2 items). The potential range for TRQ total scores is 0-40. Higher score indicate higher risk of suicide.

An expert-rated scale developed to code empathetic opportunities, defined as an explicit, clear, and direct statement of emotion, progress, or challenge by the patient. The ECCS also codes healthcare providers' verbal responses to these opportunities ranging from level 6 to level 0, denial of patient's perspective. The CPs' responses to the empathetic opportunities presented by the VHI will be coded with this scale. Full scale from 0-6, with higher score indicating more shared feeling or experience.

An expert-rated scale developed to code empathetic opportunities, defined as an explicit, clear, and direct statement of emotion, progress, or challenge by the patient. The ECCS also codes healthcare providers' verbal responses to these opportunities ranging from level 6 to level 0, denial of patient's perspective. The CPs' responses to the empathetic opportunities presented by the VHI will be coded with this scale. Full scale from 0-6, with higher score indicating more shared feeling or experience.

A semi-structured interview of current and past suicidal thoughts and behaviors. Full range from 0 to 9, with higher score indicating higher intensity suicidal ideation.

A semi-structured interview of current and past suicidal thoughts and behaviors. Full range from 0 to 9, with higher score indicating higher intensity suicidal ideation.

Well-validated 12 item, 7-point Likert scale, self-report patient and clinician questionnaire with demonstrated predictive value for psychotherapeutic treatment adherence and outcomes. Full scale from 12-84, with higher scores indicating better alliance. The WAI-SR is a patient-rated questionnaire. Patients rate items on a 5-point Likert scale anchored at each end with 'rarely or never' (1) and 'always' (5). The full score range from 5 to 20. Higher scores indicate a better therapeutic alliance.

Well-validated 12 item, 7-point Likert scale, self-report patient and clinician questionnaire with demonstrated predictive value for psychotherapeutic treatment adherence and outcomes. Full scale from 12-84, with higher scores indicating better alliance. The WAI-SR is a patient-rated questionnaire. Patients rate items on a 5-point Likert scale anchored at each end with 'rarely or never' (1) and 'always' (5). The full score range from 5 to 20. Higher scores indicate a better therapeutic alliance.

Well-validated 12 item, 7-point Likert scale, self-report patient and clinician questionnaire with demonstrated predictive value for psychotherapeutic treatment adherence and outcomes. Full scale from 12-84, with higher scores indicating better alliance. The WAI-SR is a patient-rated questionnaire. Patients rate items on a 5-point Likert scale anchored at each end with 'rarely or never' (1) and 'always' (5). The full score range from 5 to 20. Higher scores indicate a better therapeutic alliance.

A one-item 10-point Likert Scale will be used to assess clinicians' judgement of patient suicide risk. Full score from 0-10, with higher score indicating higher risk of making a suicide attempt in the next 6 months if untreated.

A one-item 10-point Likert Scale will be used to assess clinicians' judgement of patient suicide risk. Full score from 0-10, with higher score indicating higher risk of making a suicide attempt in the next 6 months if untreated.

CPs ratings of satisfaction with the VHI through a satisfaction survey completed online at the 2nd VHI interaction. The survey has six items adapted from the Maastricht Assessment of Simulated Patients. Total score from 0-36 with higher score indicating higher satisfaction with virtual human interaction.

A well-validated 10-item questionnaire that measures patient's perception of healthcare provider's empathy, with each item measured on a 5-point scale: Poor =1, Fair=2, Good=3, Very Good=4, Excellent=5. Full score from 10-50, with higher score indicating better healthcare provider's empathy perceived by the patient.

A well-validated 10-item questionnaire that measures patient's perception of healthcare provider's empathy, with each item measured on a 5-point scale: Poor =1, Fair=2, Good=3, Very Good=4, Excellent=5. Full score from 10-50, with higher score indicating better healthcare provider's empathy perceived by the patient.

A well-validated 10-item questionnaire that measures patient's perception of healthcare provider's empathy, with each item measured on a 5-point scale: Poor =1, Fair=2, Good=3, Very Good=4, Excellent=5. Full score from 10-50, with higher score indicating better healthcare provider's empathy perceived by the patient.

Well-validated 8-item self-report instrument measuring patient's medication adherence and barriers to it. Full scale is from 0-10, with higher score indicating better adherence.

Well-validated 8-item self-report instrument measuring patient's medication adherence and barriers to it. Full scale is from 0-10, with higher score indicating better adherence.

Well-validated 8-item self-report instrument measuring patient's medication adherence and barriers to it. Full scale is from 0-10, with higher score indicating better adherence.

The Facial Action Coding System (FACS) is a comprehensive, anatomically based system for describing all visually discernible facial movement, breaking down facial expressions into individual components of muscle movement. FACS, using Noldus FaceReader Software, will be applied to video recordings of clinicians interacting with the VHs. The system will evaluate clinicians' facial affective behavior to detect the presence of basic emotions, including sadness, happiness, anger, disgust, surprise, and fear. Duration of analyzed video segments in minutes and seconds, as well as computerized character values of emotions detected by the software, will be provided for analysis.

The Facial Action Coding System (FACS) is a comprehensive, anatomically based system for describing all visually discernible facial movement, breaking down facial expressions into individual components of muscle movement. FACS, using Noldus FaceReader Software, will be applied to video recordings of clinicians interacting with the VHs. The system will evaluate clinicians' facial affective behavior to detect the presence of basic emotions, including sadness, happiness, anger, disgust, surprise, and fear. Duration of analyzed video segments in minutes and seconds, as well as computerized character values of emotions detected by the software, will be provided for analysis.

The number of nonsensical and sensical VH responses will be manually coded. Members of the research team will code nonsensical and sensical VH responses through assessment clinician transcripts with the VH, which are generated automatically after each VH interaction. The proportion of nonsensical to sensical counts per transcript will be calculated, and items are rated nominally as 0 = nonsensical and 1 = sensical.

The number of nonsensical and sensical VH responses will be manually coded. Members of the research team will code nonsensical and sensical VH responses through assessment clinician transcripts with the VH, which are generated automatically after each VH interaction. The proportion of nonsensical to sensical counts per transcript will be calculated, and items are rated nominally as 0 = nonsensical and 1 = sensical.

Clinician Participant Inclusion Criteria: a staff clinician or trainee in one of the Psychiatry outpatient clinics of the MSHS or Florida International University/Citrus Health Network outpatient clinics have a caseload of 5 - 10 patients meeting PP inclusion and exclusion criteria. Exclusion Criteria: - non-approval for study participation by department chairman, training director or clinical supervisor. Patient Participant Inclusion Criteria: currently in ongoing outpatient treatment at one of the recruitment clinics have SI in the past month as defined by the CSSRS, have a lifetime history of actual SA as defined by the CSSRS or to have a confirmed psychiatric emergency room visit or hospitalization within the past two weeks. Exclusion Criteria: a need for hospitalization at the time of study enrollment acute medical illness, cognitive impairment, linguistic limitation precluding understanding of the consent or research questions and lack of contact information needed for follow up assessment.