Prof. Bingxiang Yang

Wuhan University, China

Speech Title： Construction and Evaluation of Proactive Stratified Intervention Program for Emotional Crisis—Based on an Automatic Crisis Balance Analysis Model

Authors：Bing Xiang Yang, Lin Xia

Objectives (1) To construct an automatic crisis balance analysis model, which can

proactively identify and stratify early warning of emotional crisis, analyze psychological characteristics such as cognitive distortion and interpersonal relationships, and use it as a basis for targeted intervention and an effect evaluation tool. (2) To design a stratified intervention program: ①for people with general emotional crisis (with negative emotions), develop a computerized cognitive behavioral therapy (CCBT) system and add online peer support on the basis of CCBT intervention; ② for people with extreme emotional crisis (with high suicide risk), one-to-one crisis intervention based on the “six-step crisis intervention model” was implemented, and CCBT intervention was assisted. (3) To carry out empirical research on general emotional crisis intervention and extreme emotional crisis intervention to verify the effect of the interventions.

Methods Part I: the construction and evaluation of the automatic crisis balance analysis model: the model is formed by constructing the multi-task fine-grained deep learning model based on attention mechanism by using the computer deep learning technology. The model consists of suicide risk assessment model, cognitive distortion and interpersonal relationship analysis model. F1 score, precision rate, recall rate and accuracy rate were used as the evaluation indicators of the model performance, and the prediction efficiency of suicide risk, cognitive distortion and interpersonal relationship was verified in the “Zou Fan Tree Hole” microblog group and college student group. Part II: stratified intervention of emotional crisis under the guidance of the automatic crisis balance analysis model: (1) for people with general emotional crisis: based on the principles of cognitive behavioral therapy, the CCBT program was developed, and online peer support was added on the basis of CCBT intervention. The recruited 104 college students were randomly divided into three groups. The first group received CCBT intervention (CCBT group, 35 participants), and the second group received CCBT intervention combined with online peer support (CCBT + peer support group, 35 participants). The intervention lasted for 6 weeks in both groups. The third group served as a blank control group (34 participants) without any intervention for 6 weeks. Depression, anxiety, insomnia, cognitive distortion and interpersonal relationship were investigated using Patient Health Questionnaire-9, 7-item Generalized Anxiety Disorder Scale, Insomnia Severity Index, Cognitive Bias Questionnaire and Interpersonal Relationship Assessment Scale. Meanwhile, cognitive distortions and interpersonal relationship changes of participants before and after the intervention were assessed using the automatic crisis balance analysis model. The measurement data were described by mean ± standard deviation, ANOVA was used to compare between-group differences in scale scores before and after intervention among the three groups, and paired[1]samples t-tests were used to compare within-group differences in scale scores before and after intervention among the three groups. The count data were described as frequency or percentage, and the paired Chi-square test was used to compare the differences in changes in cognitive distortions and interpersonal relationships assessed by the model before and after the intervention. (2) For people with extreme emotional crisis: the recruited 39 patients with mood disorders at risk of suicide received one-to-one crisis intervention based on the “six-step crisis intervention model” and CCBT intervention. Suicide risk, suicide attitude, cognitive distortions and interpersonal relationship level were investigated using the Suicidal Behaviors Questionnaire-Revised, Suicide Attitude Questionnaire, Cognitive Bias Questionnaire and Interpersonal Relationship Assessment Scale. Meanwhile, cognitive distortions and interpersonal changes of participants before and after the intervention were assessed using the automatic crisis balance analysis model. The measurement data were described by mean ± standard deviation, paired-samples t-tests were used to compare differences in scale scores before and after intervention. The count data were described as frequency or percentage, and the paired Chi-square test was used to compare the differences in changes in cognitive distortions and interpersonal relationships assessed by the model before and after the intervention.

Results (1) The automatic crisis balance analysis model showed better performance in predicting suicide risk, cognitive distortion and interpersonal relationships. For the early warning of suicide risk, the F1 score, precision rate, recall rate and accuracy rate of the model were 77.98%, 80.75%, 75.41% and 78.68%, respectively. For the prediction of cognitive distortion and interpersonal relationships, the F1 score, precision rate and recall rate of the model were 77.26%, 78.22% and 76.33%, respectively. Chi-square test was used to compare the results of model analysis with those of scale evaluation, and there was no statistically significant difference in cognitive distortion and interpersonal relationships (P>0.05). (2) Results of general emotional crisis intervention: during the intervention process, the number of dropouts in the CCBT group, CCBT + peer support group and blank control group were 9, 6, 5, respectively, and the dropout rate was 19.2%. The final sample size was 84 cases,

including 26 cases in the CCBT group, 29 cases in each of CCBT + peer support group and blank control group. After 6 weeks of intervention, the scores of depression, anxiety, insomnia and cognitive distortion in CCBT group decreased significantly. Depression, anxiety, insomnia, cognitive distortion and interpersonal relationship scores were significantly decreased in CCBT + peer support group. Compared with CCBT group, CCBT + peer support group had a higher completion of the CCBT program and fewer dropouts. The automatic crisis balance analysis model was used to evaluate the changes of cognitive distortion and interpersonal relationships in the intervention group before and after intervention. The results showed that the proportion of cognitive distortion in CCBT group decreased after intervention, and the difference was statistically significant (P<0.05), while the difference of proportion of interpersonal relationships was not statistically significant (P>0.05). The proportion of cognitive distortion and interpersonal relationships in CCBT + peer support group decreased after intervention, and the differences were statistically significant (P<0.05). (3) Results of extreme emotional crisis intervention: during the intervention process, 16 cases dropped out naturally, so the final sample size was 23. The suicide risk level, cognitive distortion and interpersonal relationships score of patients with extreme emotional crisis were significantly decreased after 6 weeks of intervention (P<0.05). The automatic crisis balance analysis model was used to evaluate the changes of cognitive distortion and interpersonal relationships before and after the intervention, and the results showed that the proportion of cognitive distortion and interpersonal relationships decreased after intervention, and the differences were statistically significant (P<0.05).

Conclusions (1) The automatic crisis balance analysis model showed better performance in predicting suicide risk, cognitive distortion and interpersonal relationships. In addition, the prediction results of suicide risk, cognitive distortion and interpersonal relationships by the automatic crisis balance analysis model were highly consistent with the expert artificial judgment and scale evaluation results. The model can be used as a guiding basis for targeted intervention and effect evaluation tool. (2) Single CCBT intervention can improve college students’ depression, anxiety, insomnia and cognitive function. (3) CCBT combined with online peer support intervention can not only improve college students’ depression, anxiety, insomnia and cognitive function, but also improve the interpersonal relationships level and reduce the dropout rate. (4) One-to-one crisis intervention based on the “six-step crisis intervention model” combined with CCBT intervention can effectively alleviate the suicide risk of patients with mood disorders, and improve cognitive function and interpersonal relationships level.

Keywords：Emotional crisis; computerized cognitive behavioral therapy; peer support; six-step crisis intervention model

Biography

Bing Xiang Yang RN, Ph.D., FAAN,Young Top-notch Talent Cultivation Program of Hubei Province. She is currently a Full Associate Professor with the School of Nursing, and a Researcher with the Department of Psychiatry, Renmin Hospital, Wuhan University, Wuhan, China. She got three national funding to support the mental health promotion researches and published at least ten international papers, some papers were published at the top medicine journals, such as The Lancet Global Health and American Journal of Psychiatry. She also worked as a Counsellor and a Professional Volunteer for the volunteer group Tree Hole Rescue Team, which utilizing AI to assist suicide monitoring and intervention among social network users. Her research interest is interdisciplinary study of mental health and artificial intelligence, especially focused on AI-based suicide monitoring, and intervention.