Introduction

Mental health problems are as prevalent among university students as they are in their non-university-attending same age peers (Blanco et al., 2008). In a study of surveys carried out in 21 countries, an estimated 20% of university students met criteria for a mental disorder (Auerbach et al., 2016). In the case of the Mexican university population, Melo-Carrillo, Van Oudenhove and López-Avila (2012) found that among Mexican medical students 36.29% of the students scored positive for depressive symptoms. In another study, there was a significant increase in the consumption of both alcohol and tobacco in transitioning from secondary school to university (Alvear-Galindo et al., 2015). Other mental health problems that have been researched among Mexican university students are eating disorders, suicide ideation, drinking problems, and post-traumatic stress disorder (Chávez-Rosales, Camacho, Maya, & Márquez, 2012; Díaz et al., 2008; Mendoza, Márquez, Guadarrama, & Ramos, 2013; Moreno & Ortiz, 2009; Rosales, Córdova, & Ramos, 2012). However, as far as we know, non-suicidal self-injuries (NSSI) have not been studied in the Mexican university population.

NSSI, conceptualized as the deliberate destruction or alteration of body tissue without the intent to die (Klonsky, 2007), like cutting or burning oneself, is a growing public health problem worldwide. NSSI are used for a variety of functions or motivations, like affect-regulation or self-punishment (Klonsky, 2007). It has also been hypothesized that the brain structures involved in processing the cognitive and emotional components of pain are altered in those that engage in NSSI indicated by an anomaly in the integration of pain as a conscious experience conceived of as unpleasant (Mendoza & Pellicer, 2002). Most research in this area has been conducted with adolescents (Jutengren, Kerr, & Stattin, 2011; Whitlock, Eckenrode, & Silverman, 2006; You, Leung, & Fu, 2011) and university students (Tsypes, Lane, Paul, & Whitlock, 2016; Whitlock et al., 2011; Yurkowski et al., 2015). The international prevalence in young adults (from Asia, Australia/New Zealand, Canada, Europe, United Kingdom, and the USA) is 13.4% (Swannell, Martin, Page, Hasking, & St John, 2014). NSSI can have negative consequences, like scars and infections, guilt, shame, and possible rejection of parents and peers (Wilkinson & Goodyer, 2011). Also, high rates of NSSI have been associated with severe psychopathology, such as depression, anxiety, borderline personality, drug abuse, and eating disorders as well as an increased risk of suicidal attempts (Bresin & Gordon, 2013; Klonsky & Olino, 2008). Some studies have reported that few people seek professional help for NSSI, which limits the possibility of intervention (Whitlock et al., 2006).

Some of the research on college students has found relations between NSSI and different psychological variables. Yurkowski et al. (2015) found that the quality of parent-child relationships had a more relevant impact on prediction of NSSI than the quality of peer relations, and that the feelings of alienation of both parents and peers had indirect effects on NSSI through deficits of emotion regulation. Another study in university students found that there is a continuum between NSSI and suicidal self-injury, and the differences are in degree rather than in kind (Orlando, Broman-Fulks, Whitlock, Curtin, & Michael, 2015). Whitlock et al. (2011) report that in a sample of college students the lifetime prevalence of NSSI was 15.3%. Females were more likely to self-injure because they were upset or in the hopes that somebody would notice them, whereas males were more likely to report anger or intoxication as a motivation. Also, disclosure of NSSI was low among both sexes, because only 8.9% disclosed their self-injuries to a mental health professional. Kharsati and Bhola (2014) found that the most common method of NSSI was self-hitting (15.2%), followed by cutting or carving skin (13.2%). Self-efficacy has been found to be related to depressive and anxiety symptoms and in one study to mediate the relation of stressful life events with depression and anxiety symptoms (Maciejewski, Prigerson, & Mazure, 2000; Muris, 2002). We therefore consider that self-efficacy could be a relevant variable in the study of NSSI.

Currently, there is limited data from developing countries, which hinders the efforts to have a broader picture of the problem and coordinate efforts to address this health issue (Muehlenkamp, Claes, Havertape, & Plener, 2012). There is scarce information from Latin America in general and Mexico in particular; in Mexican adolescents some studies have found prevalences between 5.6 to 17.1% (Albores-Gallo et al., 2014; González-Forteza et al., 2005). One study in a clinical sample with adolescents found that self-injury was present in high frequencies in female patients, with low socioeconomic status and a history of family violence (Ulloa, Contreras, Paniagua, & Figueroa, 2013). Currently we do not have any estimates of the prevalence of NSSI in Mexico in the general population of young adults, or in the college population in particular. It is important to have more information about how NSSI are present in the Mexican population in order to guide relevant strategies of intervention.

Therefore, the objective of the current investigation is to evaluate a sample of university students in Mexico City, a group for which there is no prior data that we are aware of with regards to NSSI, in terms of frequency, types of self-injury, gender differences, and correlations with psychological variables like depressive symptoms, anxiety symptoms, impulsivity, self-efficacy, and emotional regulation.

Method

Sample

A convenience sample of 564 undergraduate students (aged 17 to 26 years) from eight universities and different areas of study (mostly from psychology) in the Mexico City metropolitan area participated in a survey administered in their classrooms between 2012 and 2013. Two universities were public and six private. To calculate the sample, we used the golden rule, i.e., 10% of the population of approximately 6,000 undergraduate students (Ávila et al., 2009).

Measures

Non-suicidal self-injury: NSSI were assessed using the 16 questions about self-injurious behaviors of the Deliberate Self-Harm Inventory (DSHI) by Gratz (2001). The answer format was modified to indicate, for each NSSI behavior, the lifetime frequency of the behavior, on a 5-point scale (0, 1, 2-4, 5-10, 11, or more occasions). To adapt the scale to Mexico, the scale was translated by two bilingual psychologists, one a native Spanish speaker and the other a native English speaker. They reached a consensus for possible discrepancies in the translations. Also, two questions were added: one asked how many times during the last 12 months they had self-injured, and the other asked their age the first time they engaged in self-injury (which was then categorized as 10 years of age or less, between 11-17 years of age, and 18 years of age or older in order to evaluate if onset was before, during, or after adolescence given that the preponderance of the literature on NSSI is from adolescent samples).

Depressive symptoms: Depressive symptoms were assessed with the Center for Epidemiologic Studies Depression Scale (CES-D) which measures depressive symptomatology during the previous week (Radloff, 1977). The CES-D consists of 20 items measuring on a 4-point Likert scale the numbers of days in which the symptom was present: 0, 1-2, 3-4, or 5-7 days. In Mexico, this scale has been used extensively in research with adolescents and adults, showing adequate reliability (Cronbach’s alphas between .79 and .86) and concurrent validity (Benjet, Hernandez-Guzman, Tercero-Quintanila, Hernández-Roque, & Chartt-Leon, 1999; González-Forteza, Wagner-Echeagaray, & Jiménez-Tapia, 2012). In this study, Cronbach’s alpha was .90.

Anxiety symptoms: The Generalized Anxiety Disorder 7-item Scale (GAD-7) by Spitzer, Kroenke, Williams, and Lowe (2006) was employed to measure generalized anxiety in the prior two weeks. It is a 7-item Likert type scale with four response options: never, several days, over half the days and nearly every day. The Spanish language version was adapted by García-Campayo et al. (2010), finding good internal consistency (Cronbach’s alpha of .93) and strong content, criteria, and concurrent validity as assessed by interrater agreement, ROC curve analysis, and associations with other measures of anxiety and impairment. Cronbach’s alpha for the scale in this study was .87.

Impulsivity: We measured impulsivity with the Plutchik Impulsivity Scale (IS) by Plutchik and Van (1989). This 13-item questionnaire assesses impulsivity on a 4-point Likert scale (never, sometimes, often, almost always). In Mexico, this scale has been used in clinical populations who self-harm, with a reported internal consistency of .66 and adequate validity (Páez et al., 1996). In our sample Cronbach’s alpha was .79.

Emotion regulation: We used the Emotion Regulation Scale to assess differences in two emotion regulation strategies: cognitive reappraisal and expressive suppression (Gross & John, 2003). 10 items are measured on a 7-point Likert scale, from 1 (strongly disagree) to 7 (strongly agree). Cabello, Salguero, Fernández-Berrocal, and Gross (2013) translated and adapted the scale to Spanish. They reported a two-factor structure with adequate internal consistency; the Cronbach’s alphas were .75 for Expressive Suppression and .79 for Cognitive Reappraisal. Test-retest reliability over three months was .66 (p < .0001). Cronbach’s alphas in this sample were .78 for cognitive reappraisal and .71 for expressive suppression.

Self-efficacy: The General Self-Efficacy Scale (Bäßler & Schwarzer, 1996) is a 10-item questionnaire measured on a 4-point Likert scale (not at all true, hardly true, moderately true, exactly true) and adapted to Spanish by Sanjuán, Pérez, and Bermúdez (2000) who reported good internal consistency (.87) and predictive validity. Cronbach’s alpha in this study was .84.

Perceived need and help-seeking: Three questions were asked to measure whether those with NSSI perceived the need for professional psychological or psychiatric attention (Do you consider that your behaviors require psychological or psychiatric attention?), whether they ever sought professional psychological or psychiatric attention for NSSIs (Have you ever sought psychological or psychiatric help for your self-injury behavior?), and whether they were ever hospitalized or required medical attention for NSSIs (Did your self-injury behavior ever result in hospitalization or was serious enough to require medical treatment?). Answers were dichotomous (yes or no).

Procedure

After receiving authorization from the university, the students were contacted during school time. Following a brief introduction about the main aim of the study and obtaining informed consent, students completed the questionnaire battery collectively in their classrooms. Afterwards, information regarding services for psychological support was provided to each student.

Ethical considerations

Participants gave written informed consent. The study was approved by the Ethics committee of the Instituto Nacional de Psiquiatría Ramón de la Fuente Muñiz (National Institute of Pychiatry).

Statistical analyses

All analyses were conducted with SPSS software, version 22 (SPSS Inc., Chicago, IL). First, descriptive analyses were conducted to estimate frequency for each individual type of NSSI and for any lifetime NSSI. Sex differences in having engaged in each type of NSSI were evaluated with the Chi square test. Participants were then categorized as having never engaged in NSSI, having experimented with NSSI (those engaging in the behavior one to four times), or having recurrent NSSI (five or more incidents of NSSI). The classification of experimental and recurrent NSSI is for the sum of all NSSI behaviors. This classification of NSSI was based on that of Sarno, Madeddu, and Gratz (2010).

Second, we conducted ANOVAs to analyze differences among those who have never engaged in NSSI, those who had experimented with NSSI, and those with recurrent NSSI with regards to psychological factors: depressive symptomatology, anxiety, impulsivity, self-efficacy, and emotional regulation (cognitive reappraisal and expressive suppression). The interaction of gender with NSSI group was evaluated for each psychological factor. When effects were found to be significant, post hoc comparisons (Tukey) were conducted to determine which groups differed.

Third, we performed a multinomial logistic regression to estimate the probability of NSSI given relevant clinical characteristics (using a one standard deviation cut-off point on the psychological factors). The multivariate model included all the variables simultaneously in the same model. Finally, we performed chi squared tests to determine differences in perceptions of need and help seeking between experimental and recurrent self-injurers.

Results

As shown in Table 1, the sample contained more female subjects (76.5%), most were psychology students (85.4%), and were largely young adults between 20 and 26 years of age (56.4%).

With respect to NSSI, during the prior 12 months, 10.9% of the sample self-injured between 1 and 50 times. The age of onset for NSSI among those with at least one self-injury was: 18.5% before the age of 11, 74.1% between 11 and 17 years of age, and 7.4% after 17 years of age. The frequencies of lifetime NSSI are shown in Table 2. Of the total sample, 42.1% never had any self-injury, whereas 30.9% experimented with NSSI at some point in their life and 26.9% engaged in the behavior recurrently. The most frequent NSSI behaviors in the sample among those who self-injure (Table 3) were self-cutting (48.0%), carving words into skin (44.6%), severe scratching (39.1%), punching self (39.1%), sticking sharp objects into skin (37.6%), and carving pictures into skin (33%). While there were no gender differences in overall self-injurious behaviors, more females engaged in self-cutting than males (χ² = 7.13; p < .01) and more males engaged in burning themselves with cigarettes than females (χ² = 6.65; p < .01).

Table 4 shows the means of the psychological characteristics (depressive symptoms, anxious symptoms, impulsivity, self-efficacy, cognitive reappraisal, and expressive suppression) by NSSI group and sex. The recurrent group had the highest levels of depressive symptomatology, anxiety, and impulsivity. Females had higher levels of depression and anxiety than males in all groups. Results of the ANOVAs are also presented on Table 4 showing a significant group effect for depressive symptoms, anxiety, impulsivity, and self-efficacy, a gender effect for depressive symptoms, anxiety, and self-efficacy, and an interaction effect for depressive symptoms only. Post hoc analyses reveal that those with recurrent NSSI had more elevated levels of anxiety, depression, and impulsivity and lower self-efficacy compared with those with experimental NSSI and those without NSSI. There were no significant differences between the experimental and non-NSSI groups for these characteristics. With regards to the interaction effect for depressive symptoms, only females, but not males in the recurrent group had higher levels of depressive symptoms. The three groups exhibited no significant differences in emotional regulation (either cognitive reappraisal or expressive suppression).

To estimate the odds of being in the experimental or recurrent groups of NSSI given high rates in the psychological correlates, a multivariate logistic regression analysis was conducted (Table 5). Those without NSSI were the comparison group for odds ratios (ORs). Those with elevated levels of depressive symptomatology had 1.7 times the odds of experimental self-injury (95% CI [.09, 2.76]) and 2.2 times of recurrent self-injury (95% CI [1.34, 3.60]). Subjects with high impulsivity were 2.78 times more likely to have engaged in recurrent self-injury (95% CI [1.74, 4.45]), while those with high self-efficacy had half the odds of recurrent self-injury (95% CI [.30, .78]). No significant associations for emotional regulation and anxiety were found.

Finally, we evaluated, the perception of need for psychological or psychiatric attention, actually seeking psychological or psychiatric attention, and hospitalization or medical attention for NSSI (Table 6). While 54.4% of those who engaged in self-injury perceived the need for psychological or psychiatric help, only 18.1% sought psychological or psychiatric attention, and 53.1% required hospitalization or medical attention. A greater proportion of those with recurrent NSSI sought psychological or psychiatric attention than those with experimental NSSI, whereas there was no difference between the recurrent and experimental groups with regards to perceived need or required hospitalization or medical attention.

Discussion and conclusion

We found high lifetime frequencies of NSSI in this sample of Mexican university students, with 57.8% having reported at least one occasion of self-injury. This lifetime estimate is higher than others reported in college samples from other countries that range from 7-35% (Duggan, Toste, & Heath, 2012; Gratz, 2001; Muehlenkamp et al., 2012; Whitlock & Knox, 2009). These studies in general have few or no Mexicans in their samples.

Given the emphasis on adolescent samples in much of the prior literature, it is interesting to note that the proportion of these emerging adults with NSSI in the prior 12 months, only 10.9%, was much lower than the lifetime proportion. These 12-month estimates are in the range found in other countries of 7.3% to 45.5% during the prior 12 months (Duggan et al., 2012; Muehlenkamp et al., 2012; Whitlock et al., 2006). This suggests that while a greater proportion of adolescents engage in NSSI than young adults, this continues to be a problem in emerging adulthood. This is further evidenced by the reported ages of onset. While most university students with a history of NSSI (74.1%) began to self-injure in adolescence (between 11 and 17 years of age), less (18.5%) reported initiating in childhood (before age 11) and even fewer (7.4%) initiated as young adults (18 years or older). As was the case in other studies, the most common self-injury behavior was cutting (Claes, Klonsky, Muehlenkamp, Kuppens, & Vandereycken, 2010; Klonsky, 2011).

Some have suggested that the number of persons that self-injure has risen in the last decade, especially among adolescents and young adults, although this may be due to greater attention on this behavior from the media and researchers (Auerbach et al., 2016; Klonsky, 2007). Multiple factors may account for the elevated rates of NSSI among young adults. In Western societies, there is an increase in the acceptance of tattoos and piercings, and even if these behaviors are not considered NSSI, body modification, along with chats and web sites dedicated to self-injury, may have an impact upon the “normalization” of self-injury behavior (Favazza, 2009). Peer learning and the media can be relevant for the increase of NSSI, because the number of references to self-injury has increased in songs, movies, press, and internet during the last decade (Favazza, 2009). Internet sites that share or promote experiences of NSSI have risen and this could influence emotionally vulnerable young adults (Favazza, 2009; Whitlock et al., 2006). The increased exposure to NSSI may increase the probability for self-injury in such a way that social influences in college students should be taken into account (Muehlenkamp, Hoff, Licht, Azure, & Hasenzahl, 2008). Multiple factors can affect the vulnerability of young adults, from emotional regulation problems, body image, depression, impulsivity, child abuse, anxiety, substance abuse, borderline personality disorder, and suicide attempts among other factors (Brown, 2009; Duggan et al., 2012; Klonsky, 2007; Klonsky & Olino, 2008; MacLaren & Best, 2010).

Adolescence and young adulthood are stages that involve new responsibilities, more demands and stressors, and physical, cognitive, and neurological changes, so it is possible that young people are particularly vulnerable to the development of NSSI when they cannot cope with environmental demands. NSSI appear to fulfill different functions such as the emotional regulation of affective states, like depression, because when they self-injure many people report a decrease of negative mood, and a temporary relief, which influences the repetition of the behavior (Klonsky, 2007). Persons with NSSI appear to have more risk factors and difficulties responding in more adaptive ways to problems (Favazza, 2009), presenting more psychological disorders, or problems than those without self-injuries (Brown, 2009; Cerutti, Manca, Presaghi, & Gratz, 2011; Klonsky & Olino, 2008; Whitlock et al., 2006).

In this research, we considered depressive symptoms, anxious symptoms, impulsivity, self-efficacy, and emotional regulation as relevant factors to understand NSSI. From the risk factors analyzed in this investigation, depression had the greatest association with experimental NSSI. Self-injuries among experimenters may be attempts to modify and find temporary relief to negative mood (Klonsky, 2009). Among those who engage in recurrent NSSI, depression, impulsivity, and low self-efficacy are significantly associated with self-injury. Contrary to what we expected, emotional regulation (cognitive reappraisal and expressive suppression) was not significant in any analyses. This is different from some studies (Klonsky, 2007; Nock, 2009) where deficits in emotional regulation were associated with NSSI. It may be that the emotional regulation areas of this scale (Gross & John, 2003) that we used are not particularly relevant for the study of self-injury in this sample. It is possible that the emerging adults in our sample who self-injure can think of adaptive strategies to solve problems, but tend to choose more often maladaptive strategies (Nock & Mendes, 2008), in such a way that the differences between those without self-injury and those with NSSI are in other areas of emotional regulation. These areas might be impulse control difficulties, awareness of one’s own emotions, or the difficulty to perform tasks when negative emotions are felt (Gratz & Roemer, 2004).

Limitations

The results presented here should be considered in light of some limitations. The sample contained more female participants and students of psychology than the overall college population. While females did not report more overall NSSI, we do not know whether psychology students have more or less NSSI than students from other areas. So, the next step is to replicate results with a representative sample to determine the prevalence of NSSI in Mexican university students. Furthermore, the cross-sectional design precludes interpretations of causality. Finally, the large proportion of missing data for the questions on perceived need and service use suggests the need for caution in interpreting these results. We have no information to determine whether this missing data contribute to an upward or downward bias in these estimates.

Implications

Despite these limitations, the results of this study are valuable for several reasons. It is the first approximation to the understanding of self-injury in university students in Mexico. In our sample, we found a very high lifetime frequency of NSSI, as well as a substantial group who continues to self-injure after adolescence, as well as a group who initiated NSSI as emerging adults. Because there is a lack of information about these behaviors in Mexico and other countries of Latin America compared with the USA and Europe, most teachers, university authorities, and health professionals in contact with young adult students do not have any strategies to detect, prevent, or treat this problem. This is relevant not only in Mexico, but also in countries with significant Mexican immigrant populations, like the USA and Canada, to determine service needs and to develop intervention strategies, because NSSI behavior may be higher in these populations. Also, these NSSI behaviors found in this sample could be similar to other Latin American populations of college students.

Another relevant finding, similar to what has been documented elsewhere (Whitlock et al., 2006), is that most university students that self-injure do not seek professional psychological or psychiatric help for this behavior despite that roughly half recognized the need. This can be problematic not only because of the injuries and emotional problems as a consequence of self-injury, but also for the higher risk of suicide attempts that have been documented in individuals with NSSI (Wilkinson & Goodyer, 2011). Because the tendency is to avoid help, there are poor levels of engagement with health professional among individuals who self-injure and many go undetected by friends, teachers, and health professionals (Whitlock et al., 2006; Whitlock & Knox, 2009). This may be even truer in the context of Mexico and other Latin American countries where there is less awareness of the NSSI problem. In our research, the authorities of these universities were not aware that NSSI could happen at the college level, believing that NSSI was a behavior that occurred primarily in adolescents. Few public policies to address NSSI among university students likely contribute to low rates of disclosure and help seeking.

The importance of NSSI as a public health problem must be addressed by the relevant authorities, especially in the Mexican context, in order to implement public policies to assess the real dimension of the problem through national and regional surveys, and to design prevention and treatment strategies that are culturally relevant. These findings suggest that depressive symptomatology, impulsivity, and self-efficacy are factors that should be considered when developing strategies for preventing or reducing self-injury, though further research is needed to understand whether they play a causal role in the development and/or maintenance of the behavior. Compared to other countries, like the USA, Canada or the European Union, Latin America needs to address this mental health problem and conduct more research in different groups and populations to have an accurate global view of these behaviors and how to address them.