FACTORS ASSOCIATED WITH SUICIDE ATTEMPT BY PEOPLE WITH MENTAL DISORDER

Objective: to identify the factors associated with a suicide attempt by people with mental disorders. Method: observational and cross-sectional study carried out with 300 people with mental disorders undergoing treatment in two Psychosocial Care Centers II in Curitiba-PR from April to June 2014. Descriptive, univariate and multivariate analyses were performed, using logistic regression. Results: the prevalence of a history of suicide attempt was 67.7%, with a significant association in the multivariate analysis, in people of the “female gender” (p = 0.004), with “education level between zero and eight years” (p = 0.016), with “education level between nine and 12 years” (p = 0.013), with “perception of health as bad and regular” (p = <0.001) and “non-adherence to drug therapy” (p = 0.037). Conclusion: female, with a low level of education, perceiving health as bad and regular and not adhering to drug therapy were the prominent factors associated with the history of attempted suicide.


INTRODUCTION
Data from the World Health Organization (WHO) showed that more than 800,000 people worldwide die by suicide each year, being the second cause of death among young people between 15 and 29 years old.It is estimated that for each consummated suicide, there are approximately more than 20 attempts, and most of them are not reported. 1onceptually, the suicide attempt has any non-lethal suicidal behavior related to an intentional act against one's own life. 1 Studies have shown that the risk of suicide increases proportionally to the number of attempts, tending to be recurrent, emphasizing that the history of attempted suicide is the main risk factor for a new attempt. 2,3This fact could be confirmed in a study developed in Taiwan, which followed 2,070 people who attempted suicide between 2006 and 2010 for at least one year, and 18. 1% of this sample had more than an attempted suicide. 2 Considering another study in Scotland, we found an even higher frequency of suicide attempts, in which 24.6% of patients hospitalized for attempted suicide returned to the service in up to four years because of this same cause. 4he risk of suicide is higher in people with mental disorders when compared to the psychologically healthy population since international records showed that approximately 90% of people who commit suicide have some type of mental disorder, with emphasis on mood, psychotic, anxiety, and personality disorders and the use of psychoactive substances. 1,5,6esearch addressing this theme contributes to the expansion of knowledge about the main factors associated with the history of suicide attempts in people with mental disorders, especially those patients undergoing out-of-hospital treatment.Also, they support the multidisciplinary health team in the preparation and planning of preventive health actions and assisting intervention in people at risk of suicide, to reduce the rates of suicide attempt and the act itself and its consequences. 7,8here are multiple factors associated with the suicide attempt of people with mental disorders, including social, biological, family and physical and mental health factors.The identification of these factors can be useful to recognize warning signs of possible suicidal behavior, favoring the anticipation of more effective interventions for its prevention. 3iscussing the factors associated with the suicide attempt is also in line with the strategies of the WHO regarding the expansion of knowledge on the topic, planning appropriate strategies to prevent suicides. 1Thus, the objective of this study was to identify factors associated with suicide attempts by people with mental disorders.

METHOD
This is an observational and cross-sectional study carried out in two Psychosocial Care Centers (Centro de Atenção Psicossocial -CAPS) in Curitiba (PR), from April to June 2014, with 300 people with mental disorders.The study included people over 18 years old who attended the service during the data collection period.Those who were in a situation of crisis, eventual assistance and who were not able to answer the questions were excluded, according to the evaluation of the multi-professional service team.
Three-hundred and seventy patients of the total of 510 with mental disorders registered in the two CAPS were approached by a verbal invitation, in which 300 of them accepted to participate after signing the informed consent form, 14 patients refused to participate and 56 did not fit the inclusion criteria.The 140 individuals who were not approached did not attend the CAPS during the data collection period.
We collected data through structured interviews and consultation of medical records.The collection was simultaneously in both services and carried out by 10 trained interviewers, who remained on the place during the service's opening hours.This allowed all participants who met the inclusion criteria to be addressed.
We used an instrument developed for this research, consisting of 23 questions related to the participant's demographic, socioeconomic, clinical and pharmacotherapeutic characteristics.The dependent variable was suicide attempt, and the independent variables were gender, religion, marital status, education, work situation, use of alcohol, use of illicit drugs, clinical comorbidity, family history of mental disorder, perception of their health, adherence to drug therapy, psychiatric diagnosis, psychiatric comorbidity, self-administration of medications for the treatment of mental disorders, family participation in treatment and other therapeutic activities.
For the statistical analysis, there was a double-check of the tabulation and codification of the questions.We inserted the data into an Excel® database and, after checking and correcting typing errors, we transferred them to the SPSS Statistics version 20® program.
Frequencies and percentages described the results of the qualitative variables.The Fisher's exact test, Williams's G test or the chi-square test assessed the association between suicide attempt and qualitative variables.For the multivariate analysis, a logistic regression model was adjusted considering suicide attempt as the answer variable and including explanatory variables those that presented p <0.25 in the univariate analysis.
We used the Wald test to assess the significance of the variables included in the final model.The estimated values of Odds Ratio with 95% confidence intervals showed the measures to drug treatment have a 1.82 chance of attempting suicide than those who adhere to drug therapy.
After combining the four variables included in the multivariate model, we identified that the profile of the person with mental disorder most likely to attempt suicide (95.7%) is female, with education up to 12 years, considering their health bad/regular and do not adhere to drug therapy.The profile with the least probability of attempting suicide (14.5%) is a male patient, with an education of 13 years or more, who considers his health very good/excellent and adheres to drug treatment in mental health.

DISCUSSION
During the analysis of the characteristics of people with mental disorders undergoing treatment at CAPS, there was a predominance of people with a history of previous suicide attempts in 67.7% of the sample, 20.3% were men and 47.3% were women.In the method used in the suicide attempt, there was a prevalence of exogenous drug intoxication.
Corroborating this finding, a study developed in emergency service in Teresina-PI showed that in 277 victims of self-inflicted violence due to exogenous intoxication, 59.5% used psychotropic drugs and 18.4% used rodenticides. 8The most used drug classes for the suicide attempt were drugs with priority action on the central nervous system. 9n the univariate analysis, there was an association with marital status with a history of attempted suicide, with a higher prevalence among divorcees, similar to the findings of research developed in South Korea, whose aim was to analyze the national suicidal tendency.This research found that unmarried people are more likely to commit suicide, specifically divorced people, sequenced by those who never married and by widowers. 10he relationship with intimate partners has direct influences on the intensification and development of suicide and suicide attempts, especially when considering that the low quality of relationship and situations of divorce and separation are also risked factors for suicidal behaviors and stimuli for have a suicide attempt. 11n the clinical comorbidities, a study aimed at investigating the risk of suicide among the American population with impaired physical health conditions showed that in 2,674 individuals who died from suicide, 62% had at least one clinical problem in the year before death.Among the health conditions associated with suicidal behavior are: sleep disorders, an acquired immunodeficiency syndrome (AIDS), brain trauma, migraine, epilepsy, chronic obstructive pulmonary disease (COPD), heart failure, cancer, and back pain. 12 association between each variable and the suicide attempt.From the estimated coefficients, we calculated the probabilities of suicide attempt according to patient profiles, defined from the variables that remained in the final multivariate model, with p <0.05 values being statistically significant.
The Research Ethics Committee of the Universidade Federal do Paraná approved the study under number 406, 158/2013, and it was conducted in accordance with Resolution 466/2012.

RESULTS
Two-hundred and three (67.7%) of the 300 participants reported having attempted suicide at any time, 63 (31%) reported more than three attempts, with exogenous intoxication being the most used method (132, 41%), and among the substances of choice, medication use was reported 125 times (79.2%) by the participants.
Table 1 shows the variables related to the suicide attempt presented by gender since the scientific literature shows that suicide attempts differ between men and women.In this study, the chi-square test showed a statistically significant difference (p = 0.0003) between suicide attempts and gender; however, there was no statistical significance in the association between gender and number of attempts, method, and substance used.
The highlighted variables statistically associated with treatment in mental health and attempted suicide were: adherence to drug therapy (p = 0.001), diagnosis of depression (p = 0.001) and diagnosis of schizophrenia (p <0.001).
In the univariate analysis, we included the variables that were significantly associated with the outcome in a logistic regression model.For each of them, the null hypothesis of no association between suicide attempts and the hypothesis that there is an association was tested.Thus, in the final model, the variables remained: gender, education, perception of their health and adherence to drug therapy.
Table 4 shows that a woman with a mental disorder has 2.26 chances of a suicide attempt than a man with a mental disorder.Also, those people with an education level of zero to eight and nine to 12 years have a chance of 3.37 and 3.70, respectively, of attempting suicide when compared to those with more years of study.
The results also indicated that perceiving their health as bad/regular increases the probability of attempting suicide by 8.37 times than those who perceive their health as very good.Also, people with mental disorders who do not adhere    In the multivariate analysis, we identified an association between attempted suicide during life and socio-demographic and clinical variables.There was a significance for females, low education level, bad/regular health condition and low adherence to drug therapy.In this sense, women are more likely to have a history of attempted suicide than men, including the likelihood of subsequent suicide attempts. 1,2,5 national study carried out in a city in Minas Gerais to identify the epidemiological profile of suicide attempts and the act itself showed that women were 16.5 times more likely to go against their own lives than men. 17he literature shows that women are more likely to show suicidal ideation and attempted suicide and, consequently, they need hospitalization for this act.When considering the total burden of mortality and morbidity together, the burden of suicidal acts by women is higher than in men. 18ontrarily to this finding, a study in CAPS in the state of Minas Gerais based on the analysis of the medical records of 410 people with mental disorders showed rates lower than the history of attempted suicide in women, with 22. 1% having this history. 7e believe that women are more at risk of committing suicidal acts due to gender vulnerability associated with psychopathology and biological, psychological and social stressors. 18According to WHO, the reasons for distinguishing between suicidal behavior rates in men and women are due to gender inequality, the access and the preference for suicide methods, the pattern of use of psychoactive substances and, mainly, to the different rates of demand for psychiatric treatments between men and women. 1 As the findings of this investigation, the literature showed that low education is directly associated with a history of previous suicide attempts. 13,19  study with the Chinese population reported that people with a history of previous suicide attempts with some psychiatric diagnosis are more prone to social and economic disadvantages. 13n a competitive society, the educational level affects the means of promoting work and the individual's financial condition, thus, education is a relevant predictor for obtaining work. 19Thus, we suggest that low socioeconomic conditions In addition to clinical conditions, psychiatric diagnoses are also frequently associated with suicide attempts, predominantly mood disorders, psychotic disorders, substance-related disorders and anxiety disorders. 13Similar to the findings of this study, national and international studies showed that the diagnosis of depression represents a positive relationship with suicide attempts. 5,14 ata from a study carried out in China with 409 people with a history of attempted suicide and 409 controls demonstrated that the diagnosis of major depressive disorder is associated with a high risk of suicide attempt in people between 35 and 70 years old and, mainly, female. 13nother study developed in the city of Recife-PE seeking to understand the psychiatric and socio-demographic aspects related to suicide attempt by exogenous intoxication explained that in 120 patients who attempted suicide, 55.4% also had symptoms corresponding to the depressive episode, while 16.9% of the control group -who reported no attempt history -had these symptoms. 14onsistent with this perspective, Turecki and Brent contextualize that, in the interpersonal factors that intensify the risk for the development of suicidal behaviors, the intense feeling of hopelessness, helplessness, failure, and devaluation were highlighted, which can be characteristic of diagnostic depressive symptoms. 5esides the mood disorders, the data in this study show a significant correlation between schizophrenia diagnosis and a history of attempted suicide.Similar to this, the literature mentioned that 25 to 50% of people with schizophrenia try against their own lives at some point in their lives, while 4 and 13% commit suicide. 15 similar characteristic was found in a study with the Canadian population, in which 39.2% of the total number of people diagnosed with schizophrenia (n = 101) had a history of attempted suicide, while 2.8% of the population without the diagnosis (n = 21,643) had this history, showing that people with schizophrenia are six times more likely to attempt suicide. 16Scientific advances have revealed that people with schizophrenia, mental comorbidities, substance abuse, suicidal behaviors and who do not adhere to therapy proposal have an increased risk of committing suicide. 15.continued  are associated with increased attempted suicide and suicide, especially when considering the possible exacerbation of psychological suffering caused by social disadvantages. 20n this study, individuals with self-perceived health as bad or regular were more likely to have a history of attempted suicide.Considering this perspective, a cohort study conducted with the French population to assess the population's self-perceived health up to 15 years before death statistically proved that individuals who died by suicide over 15 years before death were more likely to rate their health as bad. 21he literature explained that self-rated health represents an individual perception, predominantly subjective, which includes diverse intrinsic and extrinsic factors to the individual.A study developed in the United States of America that investigated changes in self-perceived health and mortality identified that self-perceived health is one of the prominent predictors of population mortality, representing a relevant aspect in the psychosocial approach on which health systems are based. 22he low adherence to drug therapy in this study was associated with a history of a suicide attempt during life.This finding corroborates research conducted with 150 patients with diagnostic criteria for type I bipolar disorder in South India, which reported that approximately 40% of people with a history of attempted suicide did not satisfactorily adhere to psychopharmacological therapy. 23herefore, drug therapy causes the minimization and remission of psychiatric symptoms and active thoughts of death.Although when they do not adhere to the proposed therapy, the worsening of the clinical condition intensifies and, consequently, favors the development of multiple suicide attempts and need for hospitalization. 24

CONCLUSION
We concluded that being female, with a low level of education, perceiving health as bad and regular and not adhering to drug therapy were the prominent factors associated with the history of a suicide attempt during life in people with mental disorders undergoing treatment in a CAPS from Curitiba.While the profile of men, with more education level, who consider their health to be very good or excellent and who adhere to drug therapy were less likely to have attempted a previous suicide.
We emphasize that the cross-sectional design of this research does not enable to make causal inferences about the factors associated with suicide attempt in people with mental disorders.Thus, the results obtained must be applied with weight in this population.Longitudinal studies are recommended to better examine the risk factors for attempted suicide, considering different contexts and the specificity of each psychiatric diagnosis.

Table 1 -
Distribution of patients with mental disorders by gender according to variables related to a suicide attempt (N = 300), Curitiba, Paraná, Brazil, 2014 Source: the authors (2014).Notes: *Chi-square test; **Williams G test; ***The question could have more than one possibility of an answer.Continue...

Table 2 -
Univariate analysis of the association between suicide attempt and demographic, socioeconomic and clinical variables of people with mental disorders (N = 300), Curitiba, Paraná, Brazil, 2014

Table 2 -
Univariate analysis of the association between suicide attempt and demographic, socioeconomic and clinical variables of people with mental disorders (N = 300), Curitiba, Paraná, Brazil, 2014 Source: the authors (2014).Notes: *Chi-square test; † Fisher' s exact test; **Two people could not answer.

Table 4 -
Logistic regression model of the association between suicide attempt and gender, education level, perception of health and adherence to drug therapy (N = 300), Curitiba, Paraná, Brazil, 2014

Table 4 -
Logistic regression model of the association between suicide attempt and gender, education level, perception of health and adherence to drug therapy (N = 300), Curitiba, Paraná, Brazil, 2014