Predicting Adverse Childhood Experiences -Childhood Trauma level

Predicting Adverse Childhood Experiences via Machine Learning Ensembles

Recently, we collaborated with the Indian Institute of Technology Mandi team to understand if predicting exposure to childhood trauma (Adverse Childhood Experiences-ACEs) is possible without directly asking questions about specific ACEs. We found that by understanding specific internalization experiences (such as well-being, depression and anxiety, sleep quality) and externalizing behaviours (such as suicide behaviour, ability to focus, history of self-harm, etc.), it is possible to understand if the individual has been exposed to low or high ACE level. The findings affect psychotherapists, psychologists and psychiatrists since many people take the time or find it challenging to discuss childhood trauma! We continue to study further. However, a summary of the findings presented at Petra ’23 is highlighted below.

The work was led by Akash Rao, and the rest of the members’ name is highlighted below.

ABSTRACT

Adverse Childhood Experiences (ACEs) have been linked to negative health outcomes later in life, including depression, anxiety, insomnia, and suicidal behaviour. Recent studies have explored machine learning methods to classify individuals based on their ACE scores and predict their mental health outcomes. However, an extensive prediction of ACE via novel machine-learning ensembles based on several measures is yet to be undertaken. In this study, we used machine learning algorithms to classify individuals into high and low ACE groups and predict their mental health outcomes using various measures, including the Major DepressiveInventory, Generalized Anxiety Disorder, Insomnia Severity Index, World Health Organization Well-Being Index (WHO-5), suicide behaviour, irrational decisions, self-harm, ability to focus, and suicidal thoughts. The study results showed that novel machine learning ensemble algorithms like a support-vector-decision tree ensemble and a support-vector-decision tree-random forest ensemble could accurately classify individuals into high and low ACE groups and predict their mental health outcomes. The study highlights the potential of using machine learning methods to identify individuals at high risk for mental health issues and provide targeted interventions to prevent the long-term negative consequences of ACEs.
(PDF) Predicting Adverse Childhood Experiences via Machine Learning Ensembles. Available from: https://www.researchgate.net/publication/371641171_Predicting_Adverse_Childhood_Experiences_via_Machine_Learning_Ensembles [accessed Aug 09 2023].