Cross-sectional study, with data derived from the second wave of the prospective cohort titled Vigilância em saúde mental e trabalho: uma coorte da população de Feira de Santana (BA) “Surveillance on Mental Health and Work: a cohort of the population of Feira de Santana (BA)”, conducted by the Epidemiology Center of the Universidade Estadual de Feira de Santana and funded by the Fundação de Amparo à Pesquisa do Estado da Bahia (FAPESB, Bahia State Research Support Foundation/SESAB/CNPQ/MS PPSUS/BA – 028/2018).

The sample for this study consisted of 4,170 individuals, aged 15 years or older, residing in urban areas of the municipality of Feira de Santana, Bahia, Brazil. The selection was done by complex sampling, stratified according to the five subdistricts of the municipality, establishing cluster sampling in two stages (census tracts and streets).

Eligibility criteria: being 15 years of age or older and residing in the selected household. All eligible individuals were invited, excluding those with hearing or verbal impairments and/or any sensitive psychoemotional condition that would prevent them from responding reliably to the survey.

Individual interviews were conducted in the participants’ own homes by a previously trained research team, with as much privacy as each household offered. Data collection took place between December 2018 and November 2019, during daytime, on all days of the week.

Suicidal ideation was assessed using item 17 of the Self-Reporting Questionnaire (SRQ-20) (Santos et al., 2009) instrument: “ Has the thought of ending your life been on your mind?” categorized as 0 (not exposed) and 1 (exposed). The independent variables of the study are described in Table 1.

The theoretical model was constructed based on a comprehensive review of the international literature, establishing an ordered structure through blocks of variables arranged by proximity to suicidal ideation. These characteristics are displayed, each at its level, in Figure 1.

Analyses were conducted using Stata software, version 16.0, with the use of the survey module to perform weighting for complex samples. First, absolute and relative frequencies of the categorical variables of interest were calculated, followed by measures of central tendency and dispersion for continuous variables, and overall prevalence estimates of suicidal ideation and according to independent variables. The Odds Ratio (OR) with 95% Confidence Intervals (CI) were calculated as the estimated measure of association, considering that the outcome’s prevalence was less than 10%.

In the bivariate analysis, Pearson’s Chi-square test with Rao-Scott correction was used to test the association between the outcome – suicidal ideation, and the independent variables, with a statistical significance level of 5%. For the processing of ordered inter- and intra-block modeling (Hosmer et al., 2013), Logistic Regression analysis with backward procedure was conducted. First, intra-block analysis was performed, with all variables from each block that had a p-value ≤ 0.20 in the bivariate analysis being included, and retaining those with a p-value ≤ 0.17 (Hosmer et al., 2013). Next, inter-block analysis was conducted, with gradual entry of blocks according to the conceptual structure of the ordered model (Figure 1), retaining in the model the variables kept in the previous block and those from the newly inserted block with a p-value ≤ 0.05. To assess the accuracy of the models, the F-adjusted mean residual test was applied (Archer & Lemeshow, 2006).

The study was approved by the Research Ethics Committee of the Universidade Estadual de Feira de Santana (protocol No. 2,420,653), following Resolutions 466/2012 and 510/2016 of the Brazilian National Health Council, as it involves research with human subjects.

Gender differences are widely established in global publications. There is ample evidence indicating that being a woman poses a greater risk for suicidal ideation (Campisi et al., 2020; WHO, 2021). In a study involving over twenty thousand French workers, being a woman was highlighted as a determinant of suicidal ideation (Niedhammer et al., 2020). This association was also observed in China, in a study involving over two thousand workers (Liu et al., 2019). Family stress, manifested in household work overload and work-related stress, increased the chances of women experiencing SI by five times compared to men (Lin et al., 2020).

Educational level is a structuring factor of general living conditions, enabling or creating barriers to job quality, unemployment, income, social valuation, and access to goods and services. Low educational attainment may favor experiences in social contexts of marginalization and violence, leading to mental distress. Therefore, it is a predominant factor in explaining SI, as evidenced in global studies (Aslan et al., 2019; Chen et al., 2019; Lange et al., 2023; Lin et al., 2020; Pereira et al., 2020). Suicidal ideation associated with low or medium education occurred in multiple age groups: in adults (Chen et al., 2019; Pereira et al., 2020), older individuals (Aslan et al., 2019), and young people (Sousa et al., 2020).

Gender differences, along with low and medium levels of education, historically impact employment conditions and income. These are social phenomena historically determined by an exclusionary model of life production, especially for women who, subjugated in a sexist and misogynistic society, occupy unfavorable and undervalued social and economic positions, compounded by domestic work and family care responsibilities (Pinho & Araújo, 2012).

Intermediate factors associated with SI included living in a context of violence (physical and/or emotional), perceiving one’s own health poorly, and expressing dissatisfaction with both oneself and one’s quality of life.

Manifestations of violence associated with suicidal ideation occur in various settings and are perpetrated by different actors, whether in the workplace or within the family. In Israel, emotional violence led to feelings of imprisonment and, consequently, suicidal ideation, planning, and attempts (Levi-Belz & Zerach, 2018). Psychological abuse poses a higher risk when compared to other types of abuse. Living in a violent environment causes suffering on various levels and exacerbates conditions that promote dissatisfaction, psychopathologies, low self-esteem, and consequently, conditions conducive to suicide.

Suicidal ideation was associated with negative perception of one’s own health. The self-perception that something is wrong with one’s health can reinforce feelings of self-depreciation. Aesthetics, self-esteem, socially referenced normative impositions, the presence of physical limitations and illnesses, the absence of leisure activities, poor social relationships, among others, may contribute to these conditions that favor the presence of SI, as observed in international studies (Adewuya & Oladipo, 2020; Kyung et al., 2020).

Global dissatisfaction, as a product of a life production model with multiple origins, is an important factor for understanding the suicidal context. Therefore, conformity in impressions about quality of life and self-understanding; as well as impressions about the physical and psychological body, shape the production of global (dis)satisfaction, which is a potent factor in explaining suicidal ideation. Understanding this dimension, which fluctuates between concrete facts and feelings, prior to psychopathologies and SI, can be an important interventional approach for secondary interventions aimed at reducing suicides.

Dissatisfaction can approach suicidal ideation through dynamic constructions stemming from feelings of devaluation, defeat, humiliation, and imprisonment. Social inability combined with a lack of moments of well-being adds to these factors, which together generate fluctuating levels of dissatisfaction (O’Connor & Kirtley, 2018). In this perspective, SI is configured as a distressing battle, crystallized by hopelessness and intense desires for death (Denneson et al., 2020; Martínez Baquero et al., 2017; Pereira et al., 2020).

Therefore, dealing with dissatisfaction is a significant task in confronting suicidal ideation. Measures involving the production of positive feelings and contentment can constitute protective factors. Such initiatives need to be aligned with actions that promote well-being (negatively associated with SI) (Kyung et al., 2020). Thus, forms of social support emerge as a structuring mechanism for the main suicide prevention actions and its context (Faria et al., 2019). Interventions that establish the strengthening of satisfaction with personal relationships, oneself, and quality of life can prevent or alleviate the problem (Pereira et al., 2018).

In this study, suicidal ideation was associated with behavioral/mental disorders: sleep disturbances, depression, and anxiety. In the international literature, mental disorders appear as the most prominent explanatory factor for SI. Among individuals who died by suicide in low/middle-income countries, psychopathologies were present in 58% of them and in 45% of those who exhibited ideation, plans, and suicide attempts.

Sleep disturbances are also widely discussed in the literature as predictors of suicidal ideation (Kainat & Ijaz, 2019; Owusu et al., 2020). They can be linked to various factors that imply psychological distress. Rumination of thoughts (psychological, melancholic, and reflexive) are common explanatory components of both sleep disturbances and the suicidal context (Holdaway et al., 2018).

Depression is the most frequently associated psychopathology with SI in the international literature (Campisi et al., 2020; Chen et al., 2019; Levi-Belz & Zerach, 2018). In a study conducted in China, it was evidenced that in individuals with suicidal ideation, the severity of depression is proportionally associated with suicide attempts (Wei et al., 2018). In a study conducted with Nepalese widows, depression, post-traumatic stress disorder, and prolonged grief were strongly associated with suicide, which reached 16.2% among these women (Garrison-Desany et al., 2020). When gender differences are associated with depression, there is a notable intensification of the suicidal context, especially the idea of killing oneself. In Canada, in a study with 397,299 adolescents, being a woman was associated with suicidal ideation (Campisi et al., 2020), as well as in the United States (Martínez Baquero et al., 2017), Africa (Owusu et al., 2020), and Australia (Resko et al., 2018). The most prevalent condition in individuals at risk of suicide is depression, with this psychopathology being an important mediator of the transition from ideation to attempts (Zakhour et al., 2021).

Generalized Anxiety Disorder (GAD) is pointed in various studies as a factor present in individuals with suicidal ideation. This data is expected, as in the very nature of the suicidal context, there is a predominance of anxious behavior (Adewuya & Oladipo, 2020; Aslan et al., 2019; Kyung et al., 2020; Pereira et al., 2018; Zakhour et al., 2021). Higher levels of anxiety may relate to impulsivity traits, a behavior strongly associated with suicide attempts and their effectiveness (O’Connor & Kirtley, 2018). Generalized Anxiety Disorder was associated with SI among healthcare professionals in Spain, where within a span of less than 30 days, six professionals attempted suicide (Mortier et al., 2021). This association was similarly observed in diverse populations across different regions. For instance, Sapara et al. (2021) found a comparable link in a population-based survey. Likewise, studies conducted in Nigeria (Adewuya & Oladipo, 2020), Portugal (Duarte et al., 2019) among adolescents, and in Brazil (Palma et al., 2020), targeting the same demographic, also reported a similar association.

Thinking of psychopathology as the main indicator for decision-making represents a belated attempt to solve the complex problem that the suicidal context represents. It is necessary to understand psychopathology not only as a proximal event but as an element that participates in a causal chain with multiple events, these being fundamental for health planning.

Protective systems need to focus on the early identification of conditions that generate mental distress; therefore, it is in a much more sensitive and difficult intervention context, which questions how social fabric is structured and sustains its value systems, how consumption relationships are established: of oneself, of others, and of the world. The current model of society is built on exclusionary social production processes historically crystallized in systems of multiple oppressions. These oppressions mark entire populations’ lives with pain, and without seeing a way out, they plunge into deep suffering, from which the solution found is to cease to exist. Therefore, although it prevails in common sense that suicide is a personal/individual event, it was observed in this study that there is a context of production of this event which, when it encounters personal weaknesses, consolidates itself. It is, in this sense, a marker of a profound need to review how social relations are established.

Therefore, suicidal ideation is not necessarily dependent on psychopathologies as determinants of its occurrence (O’Connor & Kirtley, 2018). It is important to reinforce the understanding that no factor, in isolation, has the ability to explain its context. Understanding this complex event as a system that takes into account behavioral, cognitive, and affective processes is key to its comprehension. It is necessary to understand that there are multiple factors that jointly and procedurally define this outcome. A careful look at the most distal factors represents a necessary and challenging counter-hegemonic task that needs to be faced.

This complex system associated with suicidal ideation goes beyond what can be measured, not only by individual subjective characteristics tangential to mental health but also because there are established forces in the way of living, producing, and consuming oneself and the world around us that are much more potent. There was an ordered association with suicidal ideation from the most distal to the closest factors: gender, lower educational attainment, experience of emotional or physical violence, negative perception of health, dissatisfaction with quality of life and with oneself, sleep disturbances, depression, and GAD.

From the identified factors, it is necessary to flex them to the life production model that potentiates multiple sufferings. Public policies should be proposed towards the identification and early intervention of factors that cause suffering. Such policies need to be anchored in social support, through better training of caregivers, in order to be more sensitive to the signs and symptoms of suicide. It is important for the entire community to be engaged in this process so that interventions become increasingly less delayed.

Identifying and establishing interventions in factors that are statistically associated with the suicidal context and/or societal illness, in general, will still be incipient, but it is an important contributory path to demystify the complexity of this phenomenon.

As this is a cross-sectional study, causal identification is compromised. Therefore, inferences in this regard need to be understood not only by empirical data but by the evidence from the discussed literature. Moreover, the outcome itself, of a subjective and complex nature, may trigger reservations in its recognition, producing underestimated indicators of SI. Assuming such a condition could generate social stigma, possibly resulting in information bias due to its concealment. Possibly, SI is even more prevalent than can be estimated.

Despite these limitations, the results obtained are consistent with evidence from international literature, providing a set of useful and coherent data that highlight the need for interventions in a range of distal, intermediate, and proximal conditions.