Boamente: A Natural Language Processing-Based Digital Phenotyping Tool for Smart Monitoring of Suicidal Ideation

Abstract

People at risk of suicide tend to be isolated and cannot share their thoughts. For this reason, suicidal ideation monitoring becomes a hard task. Therefore, people at risk of suicide need to be monitored in a manner capable of identifying if and when they have a suicidal ideation, enabling professionals to perform timely interventions. This study aimed to develop the Boamente tool, a solution that collects textual data from users’ smartphones and identifies the existence of suicidal ideation. The solution has a virtual keyboard mobile application that passively collects user texts and sends them to a web platform to be processed. The platform classifies texts using natural language processing and a deep learning model to recognize suicidal ideation, and the results are presented to mental health professionals in dashboards. Text classification for sentiment analysis was implemented with different machine/deep learning algorithms. A validation study was conducted to identify the model with the best performance results. The BERTimbau Large model performed better, reaching a recall of 0.953 (accuracy: 0.955; precision: 0.961; F-score: 0.954; AUC: 0.954). The proposed tool demonstrated an ability to identify suicidal ideation from user texts, which enabled it to be experimented with in studies with professionals and their patients.

Passive Sensing for the Prediction of Suicidal Thoughts and Behaviors: A Systematic Review and Recommendations for Future Research

“…Third, the current results on social signals are limited to quantitative features, such as numbers of calls or social media application usage. However, contents of social media posts might be highly relevant for the prediction of STB, as indicated in language processing models [51][52][53].…”

Application of Natural Language Processing (NLP) in Detecting and Preventing Suicide Ideation: A Systematic Review

“…Also, persistent, and passive observation of individuals with a confirmed diagnosis of mental health issues is essential and has been shown to reduce suicide and self-harm incidence (Xu et al, 2021;Diniz et al, 2022). It has been reported previously that up to ninety percent of suicides are associated with mental health issues (Brådvik, 2018), therefore passively monitoring persons with a confirmed diagnosis using an NLP or other ML/AI-based suicide risk assessment tool might be useful and advantageous.…”

“…Similar to (Diniz et al, 2022), (Cook et al, 2016) employ NLP and machine learning to predict suicide ideation and elevated mental symptoms among adults (18+) recently released from psychiatric inpatient or emergency hospital settings in Spain. They used NLP and ML (logistic regression) on participant-sent text messages.…”

Application of Natural Language Processing (NLP) in Detecting and Preventing Suicide Ideation: A Systematic Review

“…Further, [15] developed the Boamente program, which gathers textual data from users' smartphones and detects the presence of suicidal ideation. They created an Android virtual keyboard that can passively gather user messages and transfer them to a web service using NLP and Deep Learning.…”

“…These include limiting access to the means of suicide (such as pesticides, weapons, and certain medicines), training and education of healthcare professionals in recognising suicidal behaviour, responsible media reporting, raising awareness, and the use of mobile apps and online counselling tools, amongst other potential solutions. However, the screening tools that are now available may not be sensitive enough to enable person-centred risk detection consistently [15]. Consequently, there is an urgent need for novel approaches that focus on the individual when identifying people who may be at risk for suicide.…”