Interação entre trauma infantil, percepção de vínculo parental e características antissociais na vida adulta : um estudo transversal de machine learning em amostra de homens usuários de cocaína

• Ano de defesa: 2020
• Autor(a) principal: Schorr, Manuela Teixeira
• Orientador(a): Hauck, Simone
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Palavras-chave em Português: Usuários de drogas; Aprendizado de máquina; Cocaína; Trauma psicológico; Criança; Estudos transversais; Adulto; Poder familiar
• Link de acesso: http://hdl.handle.net/10183/212635

Background : Antisocial Personality Disorder (ASPD) describes a pattern of disregard for the rights of others that begins in childhood and continues into adulthood. The relationship between parental bonding, childhood trauma, and ASPD is well established. However, w hich types of child trauma and which pa rental bonding patterns are more associated with ASPD and how these variables interact remain uncertain. Objectives: Initially, we aimed to systematically review the literature about the topic. Subsequently, we aimed to better understand this complex relationship by analyzing these variables using traditional statistical analysis and machine learning (ML) method in a sample of male cocaine users. Methods: The systematic review was conducted by searching on Embase, Scielo , Web of Science and PubMed databases according to Prisma Sta t e ment guideline. We selected articles that assessed t he relationship between ASPD or antisocial traits and childhood trauma through Childhood Trauma Questionnaire ( CTQ ) which assesses five trau ma subtypes or parental bondi ng through the Parental Bonding Instrument ( PBI ) which assesses overprotection ’ and care ’ Subsequently, a cross sectional study was conducted with a sample of 346 male cocaine users selected from public psychiatric hospitalizations. Exclusion criteria were intelligence coefficient below 70, inability to participate or not agreement with the consent form. CTQ, PBI and MINI were applied. Statistical analysis was performed using a logistic regression method and ML approach. Results: In the systematic review, we found 357 abstracts with the defined search terms. Of these, 18 met all inclusion criteria. High hetero geneity can be observed between studies. Regarding CTQ, the most consistent finding was the association between physical abuse and neglect and ASPD. Sexual abuse was the variable least consistently associated with ASPD. Regarding PBI, the most consistent f inding was the relationship between low maternal and paternal care and ASPD. The overprotection dimension showed more heterogeneous findings and it is not possible to conclude how these patterns interact. In the study conducted with the sample of 346 male cocaine users, in the logistic regression analysis, the variables statistically significant associated with ASPD were emotional abuse (OR 1.95, p = 0.001), physical abuse (OR = 1.45, p = 0.032), parental overprotection (OR = 1.42, p = 0.027) and physical neglect (OR = 0.61, p = 0.017). In the ML analysis, the variables with statistical significance were emotional and physical abuse (importance of 100 and 69.6%, respectively) and paternal care and overprotection (importance of 11.2 and 6.7%, respectively). Conclusion: There is little evidence in literature of how these variables interact with each other. Further studies are needed in terms of guiding public health interventions and initiatives that can prevent factors associated with ASPD in adulthood. Our f indings contribute to a better understanding of this phenomenon and suggest specificities in relation to the stud ied population. Our results indicate a greater contribution of emotional and physical abuse as predictors of ASPD, in this population, as well as paternal parenting factors. There was a high rate of comorbidities with psychiatric illness among individuals with ASPD, suggesting a specific profile that may be more susceptible to particularly interventions than individuals with ASPD in general. Fina lly, a causal relationship cannot be affirmed from the findings of the present study. Longitudinal studies and studies with other population are needed to advance in the understanding of the phenomena under study.