Predicting suicide risk in real-time crisis hotline chats integrating machine learning with psychological factors: Exploring the black box.

BACKGROUND: This study addresses the suicide risk predicting challenge by exploring the predictive ability of machine learning (ML) models integrated with theory-driven psychological risk factors in real-time crisis hotline chats. More importantly, we aimed to understand the specific theory-driven factors contributing to the ML prediction of suicide risk. METHOD: The dataset consisted of 17,654 crisis hotline chat sessions classified dichotomously as suicidal or not. We created a suicide risk factors-based lexicon (SRF), which encompasses language representations of key risk factors derived from the main suicide theories. The ML model (Suicide Risk-Bert; SR-BERT) was trained using natural language processing techniques incorporating the SRF lexicon. RESULTS: The results showed that SR-BERT outperformed the other models. Logistic regression analysis identified several theory-driven risk factors significantly associated with suicide risk, the prominent ones were hopelessness, history of suicide, self-harm, and thwarted belongingness. LIMITATIONS: The lexicon is limited in its ability to fully encompass all theoretical concepts related to suicide risk, nor to all the language expressions of each concept. The classification of chats was determined by trained but non-professionals in metal health. CONCLUSION: This study highlights the potential of how ML models combined with theory-driven knowledge can improve suicide risk prediction. Our study underscores the importance of hopelessness and thwarted belongingness in suicide risk and thus their role in suicide prevention and intervention.

Suicide-related calls to a national crisis chat hotline service during the COVID-19 pandemic and lockdown.

BACKGROUND: A COVID-19 pandemic-related rise in suicide rates has been predicted due to social isolation, fear, uncertainty, economic turndown and grief. Detecting an increase in suicide rates is difficult in the absence of real-time data. Alternative data sources for such trends in psychopathology and suicidal behavior must be sought. METHODS: Data from a national chat-based crisis hotline for the first half of 2019 (pre-COVID-19), were compared to data from the first half of 2020 (during COVID-19). Chat sessions were classified by content and demographics and the data compared between the two time periods. OUTCOME: Total chats (n = 6756) were 48% higher during COVID-19 (p < .05). Suicide-related chat (SRC) number was also higher, although the proportion relative to all chats was slightly lower during COVID-19, compared to pre-COVID-19 (p < .05). SRCs increased during the COVID-19 lockdown. The number of severe SRCs resulting in urgent police intervention, increased during the lockdown (April-May 2020) compared with the same period in 2019 (p = .04). Issues of anxiety were higher in 2020 (19.4%) vs. 2019 (16.5%) (p < .00001) while issues of depression were lower (22.4% vs 33%, respectively) (p < .00001). The overall use of chats among adults aged >50 yrs increased during COVID-19 and likewise, the rate of SRCs in this age-group increased 30-fold in this period when compared to pre-COVID-19 (p < .00001). SRCs included more women than men (p < .0001) in both pre-COVID-19 and during the COVID-19 period, when the proportion of women increased from 62% in 2019 to 73% during COVID-19 (p < .0001). INTERPRETATION: The rise in total chats, SRCs and SRCs resulting in police action, commenced during lockdown and was ameliorated by end of the lockdown, indicating that distress created by the lockdown was more impactful than mourning deaths of loved ones, fear and uncertainty, because all these factors persisted beyond the end of the lockdown. Older populations were probably more distressed due to greater risk and less adaptability to isolation, social media and staying home. More calls by women may reflect women's better help-seeking capacity. The increase in SRCs indicates the potential for more suicides and the need for bolstering mental health services and reach-out to older people during pandemic lock-downs.