Profiling reasons for non-attendance in psychological assessments in adolescent suicide at-risk group using natural language processing.

BACKGROUND

Non-attendance at follow-up psychological evaluations is a major barrier to care for adolescents at risk of suicide. Identifying the factors associated with non-attendance is essential to improving engagement with mental health services.

OBJECTIVE

This study examines reasons for non-attendance among adolescents identified as at-risk for suicide.

PARTICIPANTS AND

SETTING

The sample consisted of 189 adolescents (72% female; mean age = 14.11, SD = 1.44) from the EPISAM-SCHOOL project who did not attend scheduled psychological evaluations.

METHODS

A quantitative approach utilizing both structured psychometric data and NLP-assisted content analysis of unstructured text data was employed to classify reasons for non-attendance. Machine learning algorithms and inferential statistics were used to identify patterns and associations. Statistical analyses assessed the association between non-attendance and psychological, familial, and behavioral factors, including experiences of abuse.

RESULTS

A Naive Bayes classifier achieved 93% accuracy in predicting non-attendance categories, which included parental refusal (most common), no response, lack of contact information, and adolescent refusal. Statistical analyses revealed significant differences in physical abuse (p = .03) and child-to-parent violence towards fathers (p = .044) among non-attendance categories.

CONCLUSIONS

Family dynamics, particularly instances of intrafamilial violence and abuse, played a crucial role in non-attendance. Stigma and adolescents' reluctance to engage with mental health services were identified as key barriers to accessing care. Findings underscore the need for multifaceted intervention strategies, including parental engagement and addressing family violence. This research contributes to the development of targeted approaches to improve engagement in mental health services for at-risk adolescents, potentially reducing suicide risk in this vulnerable population.