Association between attempted suicide and academic performance indicators among middle and high school students in Mexico: results from a national survey

According to the Global Burden of Disease Study, suicide is the leading cause of death for children and adolescents from 10 to 19 years of age living in developing countries. Among the 10–14 year old population, suicide has gone from the 14th place in 1990 to the 10th in 2013, increasing 17%; among young people aged 15–19, suicide has remained the second cause of death, but has increased by 18% [1]. In Mexico, completed suicide rates have been constant and steadily increasing, being of particular concern among the young population, increasing rapidly in the group of 15–29 year olds [2]. Population surveys have estimated that one in every 100 Mexican students made a suicide attempt in the previous year [3].

Peer relationships, teachers and families have a significant impact on academic performance, as well as on mental health and suicidal behaviors during school years [4]. Previous studies [5] show that mental health is associated with academic performance, as the latter is an important source for the development of identity, the development of social relationships between peers, the improvement of skills such as critical thinking and problem solving, and because it contributes to better opportunities for the future.

Cohort studies with vital statistics in Sweden have estimated that the odds of a serious suicide attempt in students decreased 60% for each point increase in its grading system (range 1–5) [5]. Some cross-sectional studies have reported an association between low grades and statistically significant increases of twice the odds of suicidal ideation and suicidal plan, but not with suicide attempts [6]. Other studies have established a fivefold increased likelihood of a suicide attempt among students with low perceived academic performance compared to those who rated their achievement as above average [7].

Epidemiologic studies in Mexican students have a long tradition [8, 9], mainly through local surveys of students living in Mexico City, but also through national ones. A study in 2000 found that, among 802 females students in Mexico City who had attempted suicide, 5% did it because of poor academic performance [10]. A national study in 2007, which included public schools only (n = 12,424), estimated that the prevalence of attempted suicide among high school students who reported low academic recognition was 12 and 8% among those with high academic recognition with an adjusted Odds Ratio (OR) of 1.04 (0.84–1.30) [11]. However, academic recognition is only one indicator of academic performance, and studies are needed which focus on identifying other indicators which may be associated with suicidal behaviors, to inform how to better implement effective suicide prevention programs in schools. Such policies are needed since the goal of member States of the World Health Organization (WHO)—including developing nations—is to reduce suicide rates by 10% by 2020 [12].

The purpose of this paper is to describe the national prevalence of suicide attempts among Mexican students, their distribution through different population groups and to estimate the magnitude of the association between suicide attempts and four indicators of academic performance, independent of other sociodemographic variables. We analyze a recent, large national epidemiologic survey (n = 28,519) that covered both public and private schools in rural and urban areas. Our hypothesis is that students with worse indicators of academic performance have a higher prevalence of suicide attempts.

The analysis of the sociodemographic composition of Mexican students shows slightly more women in high school (51.2%) than in middle school (49.5%) (Table 1). Just over 5% of middle school students were 15 years or older and no young people under 14 attended high school. Approximately two out of ten students worked either full or part time during the previous year, and having a part-time job was reported more frequently by high school students than middle school students (p < 0.001). Less high school students spoke an indigenous language and lived most of their lives in small towns or rural areas as compared to their middle school counterparts. 75% of middle school students lived with both parents, decreasing to 72% for high school students. Level of education both for father and mother (or their surrogates) was higher in high school students than in middle school ones. In terms of school characteristics, three quarters of middle schoolers attended at morning shift (78%), as well as 58% of high school students; at both levels the highest proportion of students was concentrated in the 7th and 10th year (41.1% for middle school students and 42.2% for high school students).

The lifetime prevalence of attempted suicide was 3% for middle school students and 4.2% for high school students. In both middle and high school students, the prevalence of attempts in women (5.2 and 6.8%) was higher than in men (1.1 and 1.5%) with a statistically significant difference (Table 2). In relation to other variables, the highest prevalence rates in middle school students were estimated among students who were enrolled in their second or third year. For the variables of academic performance, the only statistically significant difference was observed among middle school students, with a higher proportion of suicide attempts among those who rated their academic performance as fair or poor (3.8%) compared to those who perceived it as good (3.2%) or very good (1.8%), p = 0.011.

Table 3 shows the estimates of adjusted ORs from multiple logistic regression models for middle school students. Significant predictors of suicide attempt related to academic performance were: not being a student the year before, worse self-perceived performance and having failed three or more courses. Compared to those who perceived themselves to have very good academic performance, those who reported only good performance had almost twice the odds of attempted suicide (OR = 1.86; 95% CI = 1.16–2.99), whereas those who reported having fair or poor performance had 2.35 times the odds (95% CI = 1.56–3.54), controlling for all other variables in the model (sex, age, shift, grade, etc.), further pairwise comparisons did not show significant differences in these two last estimates (p = 0.25). Regarding the number of failed courses, the only statistically significant association was observed between those who reported three or more failed courses compared with those with none, with an OR = 2.41 (95% CI = 1.26–4.60).

Adjusted estimates for high school students are shown in Table 4. After controlling for sociodemographic and school characteristics variables, the statistically significant predictors were having failed two courses compared to none (OR = 1.78; 95% CI = 1.10–2.86) and self-perceived academic performance, with associations of 1.65 (95% CI = 1.08–2.52) and 1.96 (95% CI = 1.25–3.06) for the categories of good and fair/poor respectively, compared to those who reported very good performance. Again, further pairwise comparisons did not reveal significant differences in these two last estimates (p = 0.22).

In Mexico, prior estimates of the lifetime prevalence of attempted suicide among students vary from 1.4% of middle school students and 2% of high school students, [21], up to 9% in high school students [11]. In this paper we estimated a prevalence of 3% in students from middle school and 4.2% for those attending high school. While other national studies have used a single question to identify suicide attempts, in our study we used a battery of questions that increased the instrument’s sensitivity to detect young people with a genuine suicide attempt. The results are very similar to those reported by Mexican adolescents in the general population (3.1%) obtained through other instruments like the WHO Composite International Diagnostic Interview [22].

On the other hand, while the prevalence of attempted suicide increases with school year in students attending middle school (probably due to stress related to adjustments to adolescence), as it goes from 1.8 in the 7th year to practically 4 in 8th and 9th grade, among the high school population the prevalence decreased. Because the data comes from a survey, it is possible that school dropout plays a role in the prevalence of attempts, especially in the high school level: in Mexico, only 57% of the population between 15 and 18 attends school [23], with a dropout rate of 15.9% [24]. The latter could be explained because young people with major mental health problems, including suicide, are most likely to leave school at this level, thus, the prevalence diminishes through this selection effect.

Of the four indicators of academic performance we studied, only perceived academic performance was associated to suicide attempt in middle school students in bivariate analysis. After adjustment for potential confounders, self-perceived academic performance was identified as a risk factor for suicide attempt, suggesting a dose–response for both school levels. This is consistent with other findings reported in both longitudinal and cross-sectional studies [5, 7, 25]. In middle school students, those who did not attend school the previous year had higher odds of suicidal attempt. Regarding the number of failed courses, we found a significantly higher prevalence among middle school students who failed three or more, and in high school students among those with two. Therefore, it would be appropriate to identify students that have a higher number of failed courses in order to screen them for suicidal behaviors. Since the number of failed courses was self-reported, caution must be exerted in the interpretation of this results, since students might conceal that they failed a course or, quite the contrary, to over-report them by interpreting the question as not doing well.

The fourth indicator that we studied, being 2 years older for their school grade, was not associated with suicidal attempt among this population. Nevertheless, this indicator could be related with other suicidal behaviors such as ideation or suicidal plan, which at the same time are precursors to more serious behaviors [17, 26, 27]. Given that our subjective indicator (perceived academic performance) was consistently more associated with attempts than the objectives ones (like age inconsistency with grade level), it is possible that cognitive distortions resulting from depressive states change student’s perceptions of academic achievement, being this a consequence of poor mental health instead of a real decline of academic performance. Future investigations should research this area.

In middle school students three sociodemographic risk factors were identified in the self-perceived academic performance models: sex, having worked part time and the type of family structure. Women are generally at greater risk of suicidal ideation plan and attempt [28] and in our study being female was the largest predictor of suicide attempt. It is noteworthy that students who are studying and working part-time rather than full-time are at the greatest risk, yet this association has been documented elsewhere [29]. It is likely that families’ financial stress is the main driving force that makes middle schoolers to look for a job, putting them at increased burden. Our results suggest that prevention programs in middle schools may screen students for suicidal behaviors, among those who share this burden, or who have left school for a year and came back.

Because this is a cross-sectional study, its main limitation is the impossibility to estimate the incidence of suicide attempts in the student population since the first follow-up year. It is likely that students with mental health problems and suicide attempts abandon school [30], so it is very important to identify and treat currently enrolled students who have these behaviors, since only half of adolescents who reported suicide attempts received mental health care once in a lifetime [22]. Furthermore, with this design we are not able to determine the direction of the association (timing) between our academic performance indicators and suicidal attempts, and it is possible that some mental health problems, such as depression, are risk factors for low academic performance [31].

Another limitation of the study was measuring suicide attempts: despite the use of the PIDS, a scale proved and used throughout the years in Mexico, in this study we incorporate the criterion that students confirm their suicidal action, with the intention of increasing the sensitivity of the measurement of an actual attempt and not only self-harming behavior (deliberate self-harm) [15]. The effect of this criteria could (providing that this effect it is non-differential for dichotomous variables) underestimate the extent of association measures (i.e. OR) [32] so the magnitude of the relationship of suicide attempts with academic performance variables could be even higher than estimated.

Finally, this work does not take into account the role that psychiatric disorders have on suicide attempts, since they are one of its main risk factors. Studies in Mexico [33] indicate that young people with depression have a 16-fold greater risk of suicidal ideas and 5 times higher for suicide attempts compared with those without. Also, because the questionnaire was divided and applied in four different sub-samples, with the sections on academic performance and depression (which also included suicide thoughts) being applied separately, it was not possible to include any of these last measures in the analysis.

Our results show that suicide prevention efforts in México’s schools may include assessing adolescents’ perception about their own academic performance. This recommendation could be implemented through “gatekeepers” such as teachers and school personnel, who can be trained in suicide prevention and in identifying people at risk in order to direct them to an evaluation and appropriate treatment. Moreover, suicide prevention efforts in the public education system should consider comprehensive interventions at the individual, selective and universal levels, as recommended by the WHO [34] with support from other branches of the government, such as the health and public security sectors, in order to consolidate a national suicide prevention program, with the intention to cover all the way from the adequate registration of suicidal behavior to the adequate reference for treatment of students with suicide attempts.