Research teams from Laval University, Dalhousie University and the University of Montreal have designed models linked to artificial intelligence (AI) for the analysis and prediction of suicide risks. Thanks to the collaboration with the National Institute of public health du Québec (INSPQ), researchers had access to a mountain of data.
“This first major project thus constitutes a great demonstration of the potential contribution of AI to prevention in mental health and dependencies”, indicates Christian Gagné, professor at the Faculty of Science and Engineering of theLaval University and director of the Intelligence and Data Institute.
Fatemeh Gholi Zadeh Kharrat, postdoctoral fellow at Laval University, integrated ecological data, linked to demography orenvironmentand anonymized data from individuals, listed between 2000 and 2019. She analyzed statistics linked to population, insurance medicineaccessibility of health systems and much more.
Understanding to better prevent
The initiative made it possible to confirm already existing hypotheses on the subject, in addition to bringing new knowledge to the surface. “For example, we saw that people who had had mental health follow-up in the preceding 60 days had an increased risk of suicide. Idem for drug use. This is the kind of relationship we expected to see, but the analysis by machine learning clearly showed us their impact”, indicates Christian Gagné, who worked closely with Alain Lesage, professor at the Faculty of medicine of theUniversity of Montreal.
Models using AI have also shown that mental health and addiction disorders are important factors in predicting suicide. They also highlighted the cumulative effect of risk factors, both linked to the individual and to the context in which he or she operates. What happens on an individual level is also determined by ecological factors, such as the regional budget for mental health and addictions.
Analyzes have also shown that the rate of deaths by suicide among men is higher in regions where the per capita budget for situations linked to addiction is lower. “It is therefore a clear relationship between the levels of public investment in mental health and addictions and the risk of suicide that has been established. Conversely, if we increase funding, then there is a real effect on reducing this risk”, underlines Professor Gagné.
600 variables studied
This type of relationship was able to be put forward thanks to the contribution of a quantity significant number of variables. Sociodemographic situation, diagnoses and hospitalizations, physical or mental health history, regional mental health budget, some 600 clinical or societal variables were considered. “We were able to see how rich the data provided by the INSPQ was!” says Fatemeh Kharrat. Two AI models quickly emerged, defined according to gender. “The differentiation of female and male risk factors is something that is already well understood by the clinical community. By developing models by sex, it allowed us to highlight other variables of interest , or even to identify variables which would be specific according to sex”, notes Christian Gagné.
Throughout the simulations, the team measured the impact of different variables to target the factors with the most influence. “If we play with the social deprivation of the neighborhood where the person lives, what effect does that have on the risk factors?, illustrates Fatemeh Kharrat. We were able to understand the relationships between the variables and their effect on the level of risk.”
To interpret the results, the researchers worked with specialists in the field. “They could check if the relationship exists or is probable from a clinical point of view,” adds Christian Gagné.
This project, supported by funding from the New Frontiers in Research program of the three research councils of Canada, was the subject of scientific publications in journals PLOS One et JMIR Public Health and Surveillance.
Related News :