Determining shot-term and long-term risk factors for suicide using machine learning methods
Abstract
Long-term risk factors for suicide (mental disorders, social factors, suicide attempt) have been fairly well studied, but they are poor short-term predictors of suicide.
Materials and methods. Аnalyzed the socio-psychological, hormonal and biochemical data of individuals who attempted suicide, as well as the frequency of occurrence of genotypes and alleles of 9 genes more or less associated with the risk of suicide. The study was conducted during 2016—2022. The obtained data were processed using classical methods of mathematical statistics and machine learning methods, carrier vector algorithms (SVC ROC) were used; random forest method (RandonForest ROC); nearest neighbor method (KNeighborsClassifier ROC) regression analysis (LogisticRegression ROC). Results. The study made it possible to distinguish between short-term and long-term risk factors for suicide. Short-term risk factors include the presence of depression, being raised in a single-parent family, the presence of character traits: “hyperthymia” and “demonstrativeness,” as well as low levels of serotonin and norepinephrine in the peripheral blood. For long-term risk factors, the presence of frequent punishment in childhood, lack of higher education, phlegmatic type of temperament, accentuation of the character trait “demonstrativeness”, as well as polymorphism of the 5HTT gene. The resulting models had good predictive value (0.91 for short-term factors; 0.95 for long-term factors).
Conclusion. A distinction should be made between short-term and long-term risk factors for suicide. When taking into account short-term risk factors, socio-psychological and hormonal-biochemical risk factors should be taken into account. When taking into account long-term risk factors, molecular genetic and socio-psychological risk factors.
About the Author
S. V. Davidouski
ИПК и ПКЗ УО «Белорусский государственный медицинский университет»
Belarus
Belarus
References
1. Toward a biosignature for suicide / М. А. Oquendo [et al.] // American J. of Psychiatry. — 2014 — Vol. 17, № 12. — P. 1259—1277.
2. Heeringen van, K. The neurobiology of suicide / K. van Heeringen, J. J. Mann // Lancet Psychiatry. — 2014. — Vol. 1, № 1. — P. 63—72.
3. Time-related predictors of suicide in major affective disorder / J. Fawcett [et al.] // The Am. J. of Psychiatry. — 1990. — Vol. 147, № 9. — P. 1189—1194.
4. Simon, R. I. Imminent suicide: The illusion of shortterm prediction / R. I. Simon // Suicide and Life Threatening Behavior. — 2006. — Vol. 36, № 3. — P. 296—301.
5. Hopelessness and eventual suicide: A 10-year prospective study of patient hospitalized with suicidal ideation / Т. А. Beck [et al.] // American J. of Psychiatry. — 1985 — Vol. 142, № 5. — P. 559—563.
6. Relationship between hopelessness and ultimate suicide: A replication with psychiatric outpatients / Т. А. Beck [et al.] // American Psychiatric Association. — 1990 — Vol. 147, № 2. — P. 190—195.
7. Oquendo M. A. Prospective studies of suicidal behavior in major depressive and bipolar disorders: What is the evidence for predictive risk factors? / М. А. Oquendo, D. Currier, J. J. Mann // Acta Psychiatrica Scandinavica. — 2006 — Vol. 142, № 5. — P. 151—158.
8. Risk factors for suicide in psychiatric outpatients: a 20-year prospective study / G. K. Brown [et al.] // J. of Consult. and Clin. Psychol. — 2000. — Vol. 68, № 3. — P. 371—377.
9. O`Connor, R. C. The Integrated Motivational-Volitional Model of Suicidal Behavior / R. C. O`Connor // Crisis. — 2011. — Vol. 32, № 6 — P. 295—298.
10. Prospective study of risk factors for increased suicide ideation and behavior following recent discharge / P. Links [et al.] // General Hospital Psychiatry. — 2012. — Vol. 34, № 1 — P. 88—97.
11. Applying Machine Learning Approaches to Suicide Prediction Using Healthcare Data: Overview and Future Directions / Edwin D. Boudreaux [et al.] // Front Psychiatry. — 2021. — Vol. 3, № 12. — P. 707916.
12. Burke, T. A. The use of machine learning in the study of suicidal and non-suicidal self-injurious thoughts and behaviors : а systematic review. / Т. А. Burke, В. А. Ammerman, R. Jacobucci // J. оf Affective Disorders. — 2019. — Vol. 245, № 15 — P. 869—884.
13. Suicide detection in Chile: proposing a predictive model for suicide risk in a clinical sample of patients with mood disorders / J. Barros [et al.] // Brazilian J. of Psychiatry. — 2017. — Vol. 39, № 1. — P. 1—11.
14. Walsh, C. G. Predicting risk of suicide attempts over time through machine learning / C. G. Walsh, J. D. Ribeiro, J. C. Franklin // Clinical Psychological Science. — 2017. — Vol. 5, № 3. — P. 457—469.
15. Давидовский, С. В. Особенности суицидального поведения среди жителей г. Минска / С. В. Давидовский // Здравоохранение. — 2016. — № 3. — С. 72—77.
16. Давидовский, С. В. Суицидальное поведение жителей г. Минска согласно данным 9-летнего наблюдения (с 2005 по 2014 г.) / С. В. Давидовский // Психиатрия, психотерапия и клиническая психология : материалы конф., посвящ. 90-летию кафедры психиатрии и мед. психологии Белорус. гос. мед. ун-та [Минск, 17 дек. 2015 г.]. — Минск, 2015. — С. 26—30.
17. Ассанович, М. Анализ психометрических свойств шкалы оценки депрессии Монтгомери — Асберг на основе модели Раша / М. Ассанович // Актуальные проблемы психиатрии, наркологии, психотерапии и клинической психологии : сб. науч. тр. / Саратов. гос. мед. ун-т. — Саратов, 2017. — Вып. 15. — С. 60—65.
18. Давидовский, С. В. Оценка выраженности мотивации к совершению суицида как метод выявления лиц с истинным суицидальным поведением / С. В. Давидовский, Ю. В. Мещеряков // Психиатрия, психотерапия и клин. психология. — 2022. — Т. 13, № 2. — С. 121—128.
19. Анализ социальных факторов и индивидуально-психологических особенностей лиц, совершивших парасуицид / С. В. Давидовский [и др.] // Психиатрия, психотерапия и клин. психология. — 2020. — Т. 11, № 3. — С. 531—545.
20. Анализ показателей гипоталамо-гипофизнонадпочечниковой системы у лиц, совершивших парасуицид / С. В. Давидовский [и др.] // Мед. журн. — 2021. — № 3. — С. 54—59.
21. Особенности генотипа лиц, совершивших парасуицид / С. В. Давидовский [и др.] // Психиатрия, психотерапия и клин. психология. — 2019. — Т. 10, № 3. — С. 417—427.
22. Анализ взаимосвязей между содержанием серотонина, норадреналина в периферической крови и суицидальным поведением / С. В. Давидовский [и др.] // Веснік Нац. акад. навук Беларусі. Сер. мед. навук. — 2022. — Т. 19, № 2. — C. 230—239.
23. Халилова, З. Л. Анализ ассоциаций генов, участвующих в реализации стресс-реакции, с суицидальным поведением : автореф. дис. ... канд. биол. наук : 03.02.07 / З. Л. Халилова ; Ин-т биохимии и генетики Уфим. науч. центра РосРАН. — Уфа, 2013. — 22 с.
24. Genome-wide association study of suicide death and polygenic prediction of clinical antecedents / A. R. Docherty [et al.] // The Am. J. of Psychiatry. — 2020. — Vol. 177, № 10. — P. 917 — 927.
25. SKA2 methylation is involved in cortisol stress reactivity and predicts the development of post-traumatic stress disorder (PTSD) after military deployment / M. P. Boks [et al.] // Neuropsychopharmacology. — 2016. — Vol. 41, № 5. — P. 1350—1356.
26. Voracek, M. Genetics of suicide : а systematic review of twin studies / M. Voracek, L. M. Loibl // Wien. Klin. Wochenschr. — 2007. — Vol. 119, № 15/16. — P. 463 — 475.
27. Nora Hansson Bittár. How intelligence and emotional control are related to suicidal behavior across the life course — A register-based study with 38-year follow-up / Nora Hansson Bittár, Daniel Falkstedt, Alma Sörberg Wallin // Psychol. Med. — 2020. — Vol. 50, № 13. — P. 2265—2271.
Review
For citations:
Davidouski S.V. Determining shot-term and long-term risk factors for suicide using machine learning methods. Healthcare. 2024;1(8):12—19. (In Russ.)
Views:
8
Email this article 