Exploring protective factors in a high-risk subsample: the pivotal role of paternal support in preventing depression in a cohort of young adults.

BACKGROUND: Major depressive disorder (MDD) is a global concern due to its widespread prevalence and morbidity. Identifying protective factors in high-risk individuals, including those with a familial predisposition, maltreatment history, and socio-economic vulnerabilities, is crucial. METHODS: We assessed a high-risk subsample within a young adult population cohort (n = 791; mean age = 31.94 [SD = 2.18]) across three waves. Using multiple regression models to analyse higher education, feeling supported, spirituality, psychotherapy access, higher socioeconomic status, involvement in activities, cohabitation, and family unity in Waves 1 and 2, and their association with MDD resilience at Wave 3. RESULTS: In the high-risk group, MDD incidence was 13.7% (n=24). Paternal support had a protective effect on MDD incidence (OR = 0.366; 95% CI [0.137 to 0.955], p = 0.040) and suicidal attempt risk (OR = 0.380; 95% CI [0.150 to 0.956], p = 0.038). Higher resilience scores were also protective (OR = 0.975; 95% CI [0.953 to 0.997], p = 0.030), correlating with reduced BDI (r = 0.0484; B = -0.2202; 95% CI [-0.3572 to -0.0738]; p = 0.003) and MADRS scores (r = 0.0485; B = -0.2204; 95% CI [-0.3574 to -0.0741]; p = 0.003). CONCLUSIONS: Our paper emphasizes reorienting the MDD approach, focusing on positive prevention strategies. It highlights fathers' crucial role in family-based interventions and promoting resilience in high-risk populations.

Early exposure to cannabis and bipolar disorder incidence: Findings from a 22-year birth cohort study in Brazil.

BACKGROUND AND OBJECTIVES: Bipolar disorder is a chronic condition affecting millions of people worldwide. Currently, there is some evidence to suggest that cannabis use during adolescence may be an environmental risk factor for its onset, however inconsistencies have been observed across the literature. Considering this, we aimed to assess whether early lifetime cannabis is associated with subsequent bipolar disorder in young adults between 18 and 22 years of age. METHODS: Using data from the 1993 Pelotas (Brazil) birth cohort (n = 5249), cannabis exposure was examined at age 18 by self-report, and bipolar disorder diagnosis was measured at age 22 using the Mini International Neuropsychiatric Interview (MINI). In order to control the analysis, we considered socioeconomic status index, sex, skin color, physical abuse by parents and lifetime cocaine use. RESULTS: A total of 3781 individuals were evaluated in 2015 aged 22 years, of whom 87 were diagnosed with the bipolar disorder onset after the age of 18. Lifetime cannabis use predicted bipolar disorder onset at 22 years old (OR 1.82, 95% CI [1.10, 2.93]), and the effect remained after adjusting for socioeconomic status, sex, skin color, and physical abuse by parents (OR 2.00, 95% CI [1.20, 3.25]). However, this association was attenuated to statistically non-significant after further adjustment for all available covariates, including lifetime cocaine use (OR 1.79, 95% CI [0.95, 3.19]). We also found similar results for early cocaine use, where the association with bipolar disorder onset did not maintain significance in the multivariate model (OR 1.35, 95% CI [0.62, 2.86]). Otherwise, when we considered cannabis or cocaine lifetime use as a unique feature, our findings showed that the adolescent exposure to cannabis or cocaine increased the odds by 1.95 times of developing bipolar disorder at 22 years age, even when controlling for all other study variables (OR 2.14, 95% CI [1.30, 3.47]). Finally, our models suggest that cocaine use may potentially exert a major influence on the effect of lifetime cannabis use on bipolar disorder onset, and that physical abuse by parents and sex may modify the effect of cannabis use for later bipolar disorder onset. CONCLUSION: Based on our findings, early cannabis exposure predicted bipolar disorder onset in young adults, but this association was confounded by cocaine use. Contrary to schizophrenia, cannabis as a sole exposure was not associated with bipolar disorder onset after adjusting for control variables.

Perinatal factors as risk factors of bipolar disorder onset in young adulthood: a 22-year birth cohort.

BACKGROUND: Bipolar disorder (BD) is a leading cause of disability-adjusted life years in young adults. Complications during prenatal periods have been associated with BD previously. The study aims to examine the association between perinatal factors and BD in order to prevent the risk of developing BD. METHODS: 3,794 subjects from the 1993 Pelotas population-based birth cohort study were included. We assessed 27 initial variables at birth and modelled BD onset at 18 and 22 years. We performed bivariate analysis, using binomial logistic regression models. The variables with p-value smaller than 0.05 were included into a multiple regression with confounding variables. RESULTS: Maternal smoking was associated with a 1.42-fold increased risk of BD at 18 or 22 years old (95% CI: 1.091-1.841), and maternal passive exposure to tobacco with a 1.43-fold increased risk (95% CI: 1.086-1.875). No association was found between other perinatal factors and BD after controlling for confounding factors. CONCLUSION: The results of this cohort corroborate with previous findings in the literature that already indicate the negative outcomes of maternal smoking during pregnancy. They may now be linked to other studies to target these factors for preventing the development of BD.

Premorbid intelligence quotient and school failure as risk markers for bipolar disorder and major depressive disorder.

Mood disorders significantly impact global health, with MDD ranking as the second leading cause of disability in the United States and BD ranking 18th. Despite their prevalence and impact, the relationship between premorbid intelligence and the subsequent development of BD and MDD remains inconclusive. This study investigates the potential of premorbid Intelligence Quotient (IQ) and school failure frequency as risk factors for Bipolar Disorder (BD) and Major Depressive Disorder (MDD) in a birth cohort setting. We analyze data from the Pelotas population-based birth cohort study, comprising 3580 participants aged 22, who had no prior mood disorder diagnoses. Utilizing regression models and accounting for potential confounders, we assess the impact of IQ and school failure, measured at age 18, on the emergence of BD and MDD diagnoses at age 22, using individuals without mood disorders as comparators. Results reveal that lower IQ (below 70) at 18 is associated with an increased risk of BD (Adjusted Odds Ratio [AOR] 1.75, 95%CI: 1.00-3.09, p < 0.05), while higher IQ (above 120) is linked to MDD (AOR 2.16, 95%CI: 1.24-3.75, p < 0.001). Moreover, an elevated number of school failures is associated with increased BD risk (AOR 1.23, 95%CI: 1.11-1.41, p < 0.001), particularly for BD type 1 (AOR 1.36, 95% CI: 1.17-1.58, p < 0.001). These findings offer insights into the distinct premorbid intellectual characteristics of BD and MDD and contribute to a deeper understanding of their developmental trajectories, potentially informing the development of risk assessment tools for mood disorders.

Treatment Approaches for Problematic Pornography Use: A Systematic Review.

Problematic pornography use (PPU) is an emerging condition associated with several negative psychological and sexual outcomes. This study aimed to systematically review treatment approaches for PPU. Potentially eligible studies were searched for in PubMed/MEDLINE, Embase, PsycINFO, and Web of Science up to April 1, 2023. The quality of the evidence was assessed with the use of the Joanna Briggs Institute's checklists, the Cochrane risk-of-bias tools, and the GRADE approach. A total of 8936 references were retrieved, and 28 studies were included in the systematic review (n = 500 participants). Included studies were case reports (k = 16), case series (k = 1), quasi-experimental investigations (k = 7), and randomized clinical trials (k = 4). The majority of included studies presented overall low quality and significant risk of bias, with all interventions receiving a low or very low rating according to the GRADE approach. Most studies investigated psychological interventions, with the predominance of second and third wave cognitive behavioral therapy interventions. Pharmacological treatments included opioid antagonists (naltrexone in most cases) and antidepressants, while one study investigated a protocol that included rTMS. Several investigations described the combination of psychological and pharmacological approaches. Most studies did not report on side effects, with online and self-help interventions presenting significant issues in terms of treatment adherence. Even though there are promising options that may have efficacy for the treatment of patients with PPU, the literature in the topic still presents significant limitations that compromise the reaching of more definitive conclusions.

Withdrawal-like Symptoms in Problematic Pornography Use: A Scoping Review.

OBJECTIVES: There is still debate whether problematic pornography use (PPU) can be understood as an addictive disorder, and little is known about withdrawal-like symptoms in these patients. Therefore, our aim was to summarize the existing scientific literature about this topic in the format of a scoping review. METHODS: Potentially eligible references were retrieved from 6 databases (MEDLINE/PubMed, Web of Science, Embase, PsycINFO, LILACS, and SciELO) up to February 7, 2023. Studies were included if they reported original data about withdrawal-like symptoms in individuals with PPU. Studies investigating child pornography use, or in which it was not possible to extract data of interest were excluded. Data extracted from included studies and their quality assessment were described narratively. RESULTS: A total of 937 references were retrieved, and 14 studies were included (n = 31,009 participants). According to our results, cravings were intense in most cases and frequent reasons for relapse. The proportion of participants experiencing other withdrawal-like symptoms (mental, sexual, and physical symptoms) reached up to 72.2% in the cross-sectional studies with this information available. Severity of withdrawal symptoms was associated with severity of PPU and frequency of pornography use. In some studies, measures of withdrawal symptoms were considered central symptoms of PPU and were significantly higher in individuals considering treatment. Masturbation was reported to alleviate withdrawal-like symptoms in some cases. CONCLUSIONS: There is preliminary evidence for the existence of withdrawal-like symptoms in PPU. Future studies should investigate onset, characteristics, duration, and proportion of withdrawal-like symptoms in PPU with robust methodology, also investigating potential confounding variables.

Characteristics of deaths by suicide in postmortem studies in Brazil: A systematic review and meta-analysis.

INTRODUCTION: Suicide is one of the leading causes of death, with a trend for its increase in Brazil in past decades. This study aimed to review the characteristics of suicides in Brazilian postmortem studies. METHODS: Studies investigating suicide deaths in Brazil, and based on autopsy or psychological autopsy were included. Proportions were pooled across studies with the use of random and fixed effects models. RESULTS: 6777 references were retrieved from six databases (searches up to January, 2023), and 45 studies included. In autopsy studies (k = 37, n = 16,231), substance use at toxicological analysis was found in 36.42% of cases (95% CI: 30.05-43.32), previous suicide attempts in 23.92% (95% CI: 6.73-57.78). In psychological autopsy studies (k = 8, n = 139), previous suicide attempts were reported in 28.09% (95% CI: 19.74-38.28), psychiatric conditions/symptoms in 90.67% (95% CI: 67.79-97.82), family history of suicidality in 21.33% (95% CI: 13.5-32.03). Most suicide deaths were reported in males and took place at the victim's home, hanging was the most frequent suicide method. Included studies presented significant limitations in quality assessment. CONCLUSION: Future studies should present more robust methodology, including bigger samples, the use of controls, and validated methodology.

Validation of the Ten-Item Internet Gaming Disorder Test (IGDT-10) and Its Association with Functional Impairment in Brazilian Gamers.

INTRODUCTION: Despite the growing recognition of gaming disorder as a mental disorder, there is still debate about how it should be best screened for. This is especially relevant in countries where prevalence studies that could support evidence-based policymaking are still to be conducted. This study aims to evaluate the psychometric properties of the Brazilian Portuguese version of the Ten-Item Internet Gaming Disorder Test (IGDT-10) and to explore its association with functional impairment. METHODS: An online convenience sample of 805 Brazilian adults who reported playing games completed the adapted version of IGDT-10 and World Health Organization Disability Assessment Schedule 2.0, as well as the Problematic Internet Use Questionnaire, the Center for Epidemiologic Studies-Depression Scale, the Rosenberg Self Esteem Scale and socio-demographic questions. RESULTS: The Brazilian Portuguese version of IGDT-10 demonstrated a unidimensional structure in both confirmatory and exploratory factor analysis, with satisfactory internal consistency and adequate temporal stability. Participants who scored five or more on IGDT-10 presented higher levels of functional impairment compared to those who scored positive for four symptoms or less. The difference between the two groups was statistically significant and showed a moderate effect size. Network analysis showed a direct connection between IGDT-10 and functional impairment, and identified "negative consequences" as the most relevant item connecting these variables. CONCLUSION: The IGDT-10 is a brief, easy-to-understand, valid, and reliable instrument, proving to be a suitable candidate for screening gaming disorder in future epidemiological studies.

Predicting functional impairment in euthymic patients with mood disorder: A 5-year follow-up.

Major Depressive Disorder and Bipolar Disorder are psychiatric disorders associated with psychosocial impairment. Despite clinical improvement, functional complaints usually remain, mainly impairing occupational and cognitive performance. The aim of this study was to use machine learning techniques to predict functional impairment in patients with mood disorders. For that, analyzes were performed using a population-based cohort study. Participants diagnosed with a mood disorder at baseline and reassessed were considered (n = 282). Random forest (RF) with previous recursive feature selection and LASSO algorithms were applied to a training set with imputed data by bagged trees resulting in two main models. Following recursive feature selection, 25 variables were retained. The RF model had the best performance compared to LASSO. The most important variables in predicting functional impairment were sexual abuse, severity of depressive, anxiety, and somatic symptoms, physical neglect, emotional abuse, and physical abuse. The model demonstrated acceptable performance to predict functional impairment. However, our sample is composed of young participants and the model may not generalize to older individuals with mood disorders. More studies are needed in this direction. The presented calculator has clinical, sociodemographic, and environmental data, demonstrating that it is possible to use such information to predict functional performance.

Haloperidol versus Second-generation Antipsychotics on the cognitive performance of individuals with schizophrenia and related disorders: pairwise meta-analysis of randomized controlled trials.

INTRODUCTION: Despite previous literature, the superiority of Second-generation Antipsychotics (SGAs) relative to First-generation Antipsychotics- especially haloperidol - on cognitive management in schizophrenia is still controversial. Thus, we aimed to compare the effects of haloperidol versus SGAs on the cognitive performance of individuals with schizophrenia or related disorders. METHODS: We conducted an updated systematic review and nine pairwise meta-analyses of double-blinded randomized controlled trials published up to October 30th, 2022, using Medline, Web of Science, and Embase. RESULTS: Twenty-eight trials were included, enrolling 1,932 individuals. Compared to SGAs, haloperidol performed worse on cognitive composite (MD -0.13; 95% CI: -0.33 to -0.03; MD = mean difference, CI = confidence interval), processing speed (MD -0.17; 95% CI: -0.28 to -0.07), attention (MD -0.14; 95% CI: -0.26 to -0.02), motor performance (MD -0.17; 95% CI: -0.31 to -0.03), memory and verbal learning (MD -0.21; 95% CI: -0.35 to -0.08), and executive function (MD -0.27; 95% CI: -0.43 to -0.11). In contrast, there were no significant differences between SGAs and haloperidol on working memory (MD 0.10; 95% CI: -0.08 to 0.27), visual learning (MD 0.08; 95% CI: -0.05 to 0.21), social cognition (MD 0.29; 95% CI: -0.30 to 0.88), and visuoconstruction (MD 0.17; 95% CI: -0.04 to 0.39). CONCLUSION: Haloperidol had poorer performance in global cognition and in some cognitive domains, but with small effect sizes. Therefore, it was not possible to conclude that haloperidol is certainly worse than SGAs in the long-term cognitive management of schizophrenia.

Suicide risk classification with machine learning techniques in a large Brazilian community sample.

Even though suicide is a relatively preventable poor outcome, its prediction remains an elusive task. The main goal of this study was to develop machine learning classifiers to identify increased suicide risk in Brazilians with common mental disorders. With the use of clinical and sociodemographic baseline data (n = 4039 adult participants) from a large Brazilian community sample, we developed several models (Elastic Net, Random Forests, Naïve Bayes, and ensemble) for the classification of increased suicide risk among individuals with common mental disorders. 1120 participants (27.7%) presented increased suicide risk. The Random Forests model achieved the best AUC ROC (0.814), followed by Naive Bayes (0.798) and Elastic Net (0.773). Sensitivity varied from 0.922 (Naive Bayes) to 0.630 (Random Forests), while specificity varied from 0.792 (Random Forests) to 0.473 (Naive Bayes). The ensemble model presented an AUC ROC of 0.811, sensitivity of 0.899, and specificity of 0.510. Features representing depression symptoms were the most relevant for the classification of increased suicide risk. Some of our models presented good performance metrics in the classification of increased suicide risk in the investigated sample, which can provide the means to early preventive interventions.

Depression in mothers at childhood: Direct and indirect association with problematic gaming in late adolescence/young adulthood.

OBJECTIVES: Little is known about the relationship between depression in mothers and problematic gaming in their children. Therefore, we aimed to investigate the potential role of mothers' depression in childhood as a risk factor for problematic gaming in their offspring in late adolescence/young adulthood. METHODS: We assessed data from 1557 participants on three waves (T0 collected in 2010/2011, T1 in 2013/2014, and T2 in 2018/2019) of a large Brazilian school-based cohort. Mother's depression at T0 was tested as a predictor of problematic gaming at T2 in a logistic regression model. In mediation analyses, we individually assessed internalizing or externalizing disorders at T1 as mediators in this association, with participants' sex being tested as a moderator in both models. Inverse probability weights were used to account for sample attrition at T2. All models were adjusted for maternal and participant-related covariates. RESULTS: Mother's depression at T0 was significantly associated with problematic gaming at T2 (OR = 2.09, p < 0.001) even after adjusting for multiple confounding factors. The presence of any internalizing disorder at T1 was a partial mediator of this relationship, accounting for 8.18% (p = 0.032) of the total effect, while the presence of any externalizing disorder at T1 was not a significant mediator. Participants' sex was not a significant moderator in mediation models. CONCLUSIONS: These findings suggest mother's depression in childhood as a risk factor for problematic gaming in later developmental stages, which may be partially mediated by internalizing psychopathology.

Prevalence and risk factors of psychiatric symptoms and diagnoses before and during the COVID-19 pandemic: findings from the ELSA-Brasil COVID-19 mental health cohort.

BACKGROUND: There is mixed evidence on increasing rates of psychiatric disorders and symptoms during the coronavirus disease 2019 (COVID-19) pandemic in 2020. We evaluated pandemic-related psychopathology and psychiatry diagnoses and their determinants in the Brazilian Longitudinal Study of Health (ELSA-Brasil) São Paulo Research Center. METHODS: Between pre-pandemic ELSA-Brasil assessments in 2008-2010 (wave-1), 2012-2014 (wave-2), 2016-2018 (wave-3) and three pandemic assessments in 2020 (COVID-19 waves in May-July, July-September, and October-December), rates of common psychiatric symptoms, and depressive, anxiety, and common mental disorders (CMDs) were compared using the Clinical Interview Scheduled-Revised (CIS-R) and the Depression Anxiety Stress Scale-21 (DASS-21). Multivariable generalized linear models, adjusted by age, gender, educational level, and ethnicity identified variables associated with an elevated risk for mental disorders. RESULTS: In 2117 participants (mean age 62.3 years, 58.2% females), rates of CMDs and depressive disorders did not significantly change over time, oscillating from 23.5% to 21.1%, and 3.3% to 2.8%, respectively; whereas rate of anxiety disorders significantly decreased (2008-2010: 13.8%; 2016-2018: 9.8%; 2020: 8%). There was a decrease along three wave-COVID assessments for depression [ß = -0.37, 99.5% confidence interval (CI) -0.50 to -0.23], anxiety (ß = -0.37, 99.5% CI -0.48 to -0.26), and stress (ß = -0.48, 99.5% CI -0.64 to -0.33) symptoms (all ps < 0.001). Younger age, female sex, lower educational level, non-white ethnicity, and previous psychiatric disorders were associated with increased odds for psychiatric disorders, whereas self-evaluated good health and good quality of relationships with decreased risk. CONCLUSION: No consistent evidence of pandemic-related worsening psychopathology in our cohort was found. Indeed, psychiatric symptoms slightly decreased along 2020. Risk factors representing socioeconomic disadvantages were associated with increased odds of psychiatric disorders.

Brazilian Psychiatric Association guidelines for the management of suicidal behavior. Part 3. Suicide prevention hotlines.

Crisis hotlines are direct communication systems, usually telephone-based, set up to prevent suicide. However, few studies have evaluated their effectiveness. The present study aims to perform a systematic review, using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework, of the effectiveness of interventions through direct communication systems to reduce the number of suicides or suicide attempts. We searched the MEDLINE, Cochrane, SciELO, and ClinicaTrials.gov databases, and used the 2011 Oxford Centre for Evidence-Based Medicine Levels of Evidence classification. The literature search yielded 267 studies, of which 35 fulfilled the selection criteria. Although significant heterogeneity was found among studies, there is evidence that direct telephone interventions are effective when included in broader preventive protocols and provided by trained staff. Despite the limitations, which included heterogeneity of samples, designs, and outcome measures, we were able to design a protocol for the use of remote services to prevent suicide and suicide attempts. A hotline or similar system could be an effective intervention for broader suicide prevention programs. However, further research is necessary to specify which protocol components are key to enhance effectiveness.

Systematic Review of Psychiatric Adverse Effects Induced by Chloroquine and Hydroxychloroquine: Case Reports and Population Studies.

OBJECTIVE: To perform a systematic review on the psychiatric adverse effects of chloroquine (CQ) and hydroxychloroquine (HCQ); to summarize what is known about psychiatric adverse effects of these drugs; to compare clinical trials, populational studies, and case report studies; and to increase awareness of the potential psychiatric adverse effects of these drugs. DATA SOURCES: A literature search of PubMed, Scopus, and Web of Science was performed to identify manuscripts published between December 1962 and June 2022. Search terms included CQ, HCQ, psychiatry, psychosis, depression, anxiety, bipolar disorder, delirium, and psychotic disorders. STUDY SELECTION AND DATA EXTRACTION: Relevant studies included reports of adverse effects after CQ or HCQ ingestion. DATA SYNTHESIS: The current literature presents evidence for a risk of short-term psychiatric adverse effects induced by either CQ or HCQ. However, the populational-level studies presented some limitations regarding the voluntary response in survey data, self-report adverse effects, and placebo group reporting similar symptoms to the case group. Thus, populational-level studies addressing the discussed limitations and the nature and extent of possible psychiatric adverse effects are needed. RELEVANCE TO PATIENT CARE AND CLINICAL PRACTICE: Most of the patients who developed such adverse effects did not report a family history of psychiatric disease. The frequency of psychiatric adverse effects depends on the patient's biological sex, age, and body mass index, but not on the drug dosage. CONCLUSIONS: Based on clinical trials and case reports, the current literature presents evidence for a risk of short-term psychiatric adverse effects induced by either drug.

Prenatal and perinatal risk factors for bipolar disorder: A systematic review and meta-analysis.

BACKGROUND: Perinatal and prenatal risk factors may be implicated in the development of bipolar disorder, but literature lacks a comprehensive account of possible associations. METHODS: We performed a systematic review and meta-analyses of observational studies detailing the association between prenatal and perinatal risk factors and bipolar disorder in adulthood by searching PubMed, Embase, Web of Science and Psycinfo for articles published in any language between January 1st, 1960 and September 20th, 2021. Meta-analyses were performed when risk factors were available in at least two studies. FINDINGS: Twenty seven studies were included with 18 prenatal or perinatal factors reported across the literature. Peripartum asphyxia (k = 5, OR = 1.46 [1.02; 2.11]), maternal stress during pregnancy (k = 2, OR = 12.00 [3.30; 43.59]), obstetric complications (k = 6, OR = 1.41 [1.18; 1.69]), and birth weight less than 2500 g (k = 5, OR = 1.28 [1.04; 1.56]) were associated with an increased risk for bipolar disorder. INTERPRETATION: Perinatal and prenatal risk factors are implicated in the pathogenesis of bipolar disorder, supporting a role of prenatal care in preventing the condition.

Differential impact on suicide mortality during the COVID-19 pandemic in Brazil

Objectives: To compare suicide rates observed in Brazil after the onset of the COVID-19 pandemic with the estimated rate based on suicide deaths between 2010 and 2020, and identify sociodemographic variables associated with this outcome. Methods: Ecological time-series study. Data were obtained from Brazilian Unified Health System Department of Information Technology (DATASUS), with the structural break of the data set in March 2020. The number of actual suicides observed and the number of expected suicides if there were no COVID-19 pandemic were analyzed through bayesian structural time series modeling. Results: The overall incidence of suicides in Brazil remained stable after the start of the COVID-19 pandemic compared to what would be expected. However, there was a significant increase in suicide deaths among women (6.9%) and older adult (9.1%). Analysis by macro-regions of the country showed significant increases in suicide deaths in the Center-West (7.4%), Northeast (5.7%), and Southeast (10%). Stratified analyses revealed differences according to age, sex, education, and skin color. Conclusions: Despite stability in the overall number of suicides, this phenomenon occurs heterogeneously among different population groups and regions of Brazil. Rates have increased in populations with a history of poor access to health, which may have been more severely impacted by the pandemic.

Predicting criminal and violent outcomes in psychiatry: a meta-analysis of diagnostic accuracy.

Although reducing criminal outcomes in individuals with mental illness have long been a priority for governments worldwide, there is still a lack of objective and highly accurate tools that can predict these events at an individual level. Predictive machine learning models may provide a unique opportunity to identify those at the highest risk of criminal activity and facilitate personalized rehabilitation strategies. Therefore, this systematic review and meta-analysis aims to describe the diagnostic accuracy of studies using machine learning techniques to predict criminal and violent outcomes in psychiatry. We performed meta-analyses using the mada, meta, and dmetatools packages in R to predict criminal and violent outcomes in psychiatric patients (n = 2428) (Registration Number: CRD42019127169) by searching PubMed, Scopus, and Web of Science for articles published in any language up to April 2022. Twenty studies were included in the systematic review. Overall, studies used single-nucleotide polymorphisms, text analysis, psychometric scales, hospital records, and resting-state regional cerebral blood flow to build predictive models. Of the studies described in the systematic review, nine were included in the present meta-analysis. The area under the curve (AUC) for predicting violent and criminal outcomes in psychiatry was 0.816 (95% Confidence Interval (CI): 70.57-88.15), with a partial AUC of 0.773, and average sensitivity of 73.33% (95% CI: 64.09-79.63), and average specificity of 72.90% (95% CI: 63.98-79.66), respectively. Furthermore, the pooled accuracy across models was 71.45% (95% CI: 60.88-83.86), with a tau squared (τ2) of 0.0424 (95% CI: 0.0184-0.1553). Based on available evidence, we suggest that prospective models include evidence-based risk factors identified in prior actuarial models. Moreover, there is a need for a greater emphasis on identifying biological features and incorporating novel variables which have not been explored in prior literature. Furthermore, available models remain preliminary, and prospective validation with independent datasets, and across cultures, will be required prior to clinical implementation. Nonetheless, predictive machine learning models hold promise in providing clinicians and researchers with actionable tools to improve how we prevent, detect, or intervene in relevant crime and violent-related outcomes in psychiatry.