Suicidal ideation in mothers of asthmatic children and adolescents in a subspecialty outpatient practice

: We aimed to investigate prevalence and factors associated with Suicide ideation (SI) in mothers of asthmatic children. This cross-sectional study included 362 dyads of mothers and children with asthma aged 2 to 14 years who attended two pediatric outpatient clinics in Brazil. We assessed the presence of SI (Self-Report Questionnaire-20), the occurrence of stressful events and maternal social support. The prevalence of SI was 8.6%. Low maternal education, exposure to serious illness , and low perception of social support in its affective–social in- teraction dimension remained significantly associated with SI in the final model. Thus, life stressors, social support and low maternal education accounted for most of the variation in prevalence of maternal SI. There were no effects of child asthma severity on maternal SI in this study.


Suicidal ideation in mothers of asthmatic children and adolescents in a subspecialty outpatient practice
Ideação suicida em mães de crianças e adolescentes asmáticos em ambulatório especializado Ideación suicida en madres de niños y adolescentes asmáticos en consulta externa de subespecialidad Every year, over 800,000 people die from suicide worldwide (World Health Organization, 2014). Suicide is the 15 th most common cause of death and accounts for 1.4% of all deaths globally; 75.5% of these deaths occur in developing countries (Cha et al., 2018;Turecki & Brent, 2016;World Health Organization, 2014. Suicidal behavior is a spectrum that includes suicidal ideation (SI), planning, attempt, and the action of committing suicide itself (World Health Organization, 2014). However, suicidal ideation (SI) inclusion in suicidal behavior is controversial because the factors associated with SI may differ from the factors underlying suicide attempt and suicide itself (Klonsky et al., 2016;Wetherall et al., 2018;World Health Organization, 2014). The SI prevalence was 9.2% in a multicenter study involving 17 countries. Studying SI and its determinants is important because up to 60% of the transitions to suicide planning and attempt can occur in the first year after suicidal thoughts are developed (Nock et al., 2008).
As SI can be an important predictor of death by suicide (Nock et al., 2008;World Health Organization, 2014), it is extremely important to identify SI as a topic for planning suicide prevention. Some theories regarding suicide and suicidal behavior are rooted in the ideation-to-action framework.
These theories consider SI development and SI transition to suicide attempt as distinct processes (Joiner, 2005;Klonsky & May, 2015;Klonsky et al., 2017;O'Connor, 2011). The Three-Step Theory (3ST), for example, hypothesizes that: 1) SI results from the combination of pain (mainly psychological) and hopelessness; 2) among those who experience one or both, connectedness is a crucial protective factor against the escalating SI and 3) progression from ideation to attempts depends on dispositional, acquired, and practical contributors to the capacity to attempt suicide (Klonsky & May, 2015;Klonsky et al., 2017). Regarding the first and second steps of 3ST, stressful life events (SLE) may act as a trigger and social support as a protective factor on the escalating SI. In the first case, association between negative stressful situations and SI and behavior was generally consistent (Liu & Miller, 2014).
As caring for patients with chronic mental or physical illness can be a stressful event, a growing number of studies have investigated aspects of the caregiver's mental health, including SI, whose prevalence may range from 10.3 to 18% (Huang et al., 2018;Koyama et al., 2017;O'Dwyer et al., 2016;Park et al., 2013;Skeen et al., 2014). Studies revealed much heterogeneity, with SI prevalence depending on child's illness type and other factors associated with the caregiver, such as the presence of CMD, depression, and anxiety, social support, age, and associated chronic disease (Huang et al., 2018;O'Dwyer et al., 2016;Park et al., 2013;Skeen et al., 2014). Being single, female, and unemployed; having low perception of social support; and presenting a mental disorder have been frequently identified as factors for increased risk of SI among caregivers (Huang et al., 2018;O'Dwyer et al., 2016). In the particular case of children, studies evaluating the presence of suicidal thoughts in their caregivers, especially mothers (if we consider that they are the main caregivers in this age group), are lacking. (Lise et al., 2017).
Moreover, no studies have tested any type of socioeconomic and/or psychosocial model for SI in mothers of children with asthma, the most prevalent chronic disease in childhood. (Asher & Pearce, 2014;Global Initiative for Asthma, 2015) Thus, this study aims to investigate the socioeconomic and psychosocial factors that are associated with an increased prevalence of SI in mothers of asthmatic children in subspecialty outpatient practice. Since having a child with asthma can be a stressful event for the mother, we will consider the additional stress of severe asthma in the child as a determinant for SI in these mothers.

Participants
This study was conducted in two public pediatric pulmonology outpatient clinics that are reference for attendance of children and adolescents in the state of Alagoas, Brazil. Eligible participants were mothers of asthmatic children aged 2 to 14 years selected by convenience.
A total of 481 eligible mothers were invited to participate in the study. Once they agreed to participate and signed a consent form, a face--to-face interview was conducted in a private room before medical appointment. Seventeen mothers refused to participate, and 102 questionnaires were excluded due to inconsistencies in their completion or by noting that mothers were embarrassed to complete one or more items in the questionnaire.
Mother-child dyads included mainly male children, mothers with low education, and severe financial problems (see table 1 for detailed information on socio-demographic characteristics, maternal factors, and asthma severity in the sample).

Measurements
Maternal SI. Information about maternal SI was obtained through the question "Have you had thoughts about ending your life in the past 30 days?" from the Self Report Questionnaire . This is a common mental disorder (CMD) screening questionnaire comprising 20 questions and dichotomous answers (yes/no) on symptoms over the previous 30 days and validated in Brazil (Harding et al., 1980;Mari & Williams, 1986). Participants were divided into groups of mothers with and without SI.

Maternal stressful life events (SLE).
The occurrence of maternal stressful life events (SLE) in the previous 12 months was measured through nine close-ended questions about events or unpleasant situations taking place over the previous 12 months, with dichotomized answers in yes or no -exposed or non-exposed, respectively (Lopes & Faerstein, 2001). Each one of the nine events was evaluated as an isolated variable (serious illness, hospitalization, death of close family member, severe financial problems, change of residence, separation/divorce, physical aggression, mugging/robbery). This questionnaire showed good test-retest reliability, with most of the questions having a good stability when reported by adults in a previous study in Brazil (Lopes & Faerstein, 2001).

Maternal Social Support.
Maternal social support was assessed using the Social Support Scale (Medical Outcomes Study Questions; MOS--SSS). The instrument consisted of 19 questions and answers on a five-item Likert scale ("never" -1, "rarely" -2, "sometimes" -3, "almost always" -4 and "always" -5), validated for the Brazilian population (GRIEP et al., 2005). For the purposes of this study, the items were organized to cover three dimensions of social support: 1) affectivepositive social interaction (7 items), 2) emotionalinformational (8 items), and 3) material support (4 items) (Griep et al., 2005). The higher the score, the greater the perception of social support. The scores in each dimension were dichotomized at high and low, using the first distribution quartile as a cut-off point. Support from relatives and friends. Support from relatives and friends was measured through the questions: 1) "How many relatives do you feel comfortable with and can talk about almost everything?" and 2) "How many friends do you feel comfortable with and can talk about almost everything?", also extracted from MOS-SSS mentioned above. Both variables were dichotomized based on presenting or not support from relatives or friends, at least one.
Covariates. Child's asthma severity was con-

Results
The maternal SI prevalence in the previous month was 8.6%. All mothers with SI also had evidence of common mental disorder (CMD), with eight or more positive answers to SRQ-20, a cut-off point defined in the Brazilian validation of the questionnaire (Mari & Williams, 1986). Half of the mothers in the sample had CMD. Almost half of the dyads belonged to economic classes D and E. Exposure to severe financial problems was the most frequently reported SLE in the previous year, followed by loss of a close relative (Table 1). CI95%: 1,09 -4,63, respectively). Child's asthma severity was not associated with maternal SI report (Table 2).  (Table 3). Interpersonal support provided by relatives was also inversely associated with maternal SI.

Discussion
This study including mothers of asthmatic children in a subspecialty outpatient practice regarding suicidal thoughts in the previous month found a SI prevalence of 8.6%, which was similar to that found in a previous multicenter study -9.2% (Nock et al., 2008 Connor et al., 2018). In Brazil, a study carried out with pregnant women and using a methodology like the methodology described here found SI prevalence of 6.3% (Huang et al., 2012).
Based on the assumption that context matters,  (Huang et al., 2018;O'Dwyer et al., 2016;Park et al., 2013;Skeen et al., 2014). is no questions about frequency, intensity, or duration of SI in the suicide item of the SRQ-20, thus limiting its ability to assess the severity of suicidality.
Despite these concerns, the present study has many important strengths, such as demonstrating a model where the psychosocial context explains much more IS in mothers of asthmatic children than the severity of the disease in the child. Moreover, it raises the possibility of identifying maternal SI and associated factors in a routine outpatient child consultation by means of a simplified instrument with rapid application (20-30 minutes). There was no need for the interviewers to go through complex training. The instrument could be useful mainly in the context of primary healthcare, could be applied in an outpatient waiting room, is easily accessible, and has low cost for the health system. Once the risk is identified, the caregiver should be referred to specialized care for appropriate diagnosis and treatment. The importance of identifying SI and associated factors could enable early intervention and prevention or block the process leading from ideation to suicidal behavior through active search in an interview with mothers. In addition, providing support to the mother will probably improve their children's asthma status.

Conclusions
The SI prevalence in asthmatic children's mothers was the same as in the general population.
Previously described psychosocial factors, as maternal education, exposure to stressors (in this case, serious illness) and low perception of social support in their affective-social interaction was also significantly associated with SI in mothers of asthmatic children. Our results points to the lack of association between severity of disease in children and SI in their mothers. In general, the presence of SI is better explained by the experience of life stressors and low perception of social support. Finally, the current study suggests that it is possible to identify maternal SI and associated factors with the aid of a simplified instrument during a routine visit to a child outpatient unit.