Factors Associated with Cyberbullying Victimisation among Colombian High-School Adolescents

Abstract

The study’s purpose was to examine the prevalence and explore some factors associated with cyberbullying in high-school adolescents at ​​Santa Marta, Colombia. A cross-sectional study was designed, which participated in students between 13 and 17 years. Participants completed the item for being a victim of cyberbullying of the Youth Risk Behavior Survey for high school students designed by the Center for Disease Control of the United States. A total of 1462 students answered the questionnaire (M = 14.4 years, SD = 0.8). 19.6% (n = 287) reported lifetime cyberbullying victimisation. Cyberbullying victimisation was associated with post-traumatic stress disorder risk (OR = 2.05, 95%CI 1.51–2.79), lifetime cigarette smoking (OR = 1.91, 95%CI 1.42–2.57), female gender (OR = 1.68, 95%CI 1.25–2.26), family dysfunction (OR = 1.68, 95%CI 1.18–2.41), and poor-fair health condition (OR = 1.45, 95%CI 1.08–1.95). Being a victim of cyberbullying is frequent among Colombian high-school adolescents. It is associated with post-traumatic stress disorder risk, cigarette smoking, female gender, family dysfunction, and poor-fair general health condition. Longitudinal investigations are needed to measure better the dynamics of cyberbullying and the nature of such associations.

Introduction

Cyberbullying or virtual intimidation is intentional and continuous intimidation or assault, through electronic means, such as mobile phones or the Internet, with a power imbalance between the aggressor and the victim (Olweus, 2012). Unlike traditional school bullying, cyberbullying presents the possibility of happening anytime and in anyplace. It can potentially be subject to a broader audience and the anonymity of the aggressor or identity theft (Ferreira & Deslandes, 2018). Furthermore, the perpetrator’s anonymity makes it more difficult for the victim to identify the aggressor (Olweus & Limber, 2018). Currently, cyberbullying is a public health problem (Guarini et al., 2019), and it represents a severe threat to the development (González-Cabrera et al., 2018), general well-being (Carvalho et al., 2018), and mental health (Bottino et al., 2015) of children and adolescents.

The prevalence of cyberbullying is highly variable in the studies due to the definitions used and the samples studied’ demographic characteristics (Kowalski et al.,  2014). In a systematic review that included 159 prevalence studies, Brocahdo et al. show that, in the last year, the percentages for cases of victimisation were between 1.0 and 61.1%. In the same period, the proportion of perpetration was between 3.0% and 39.0%. Lifetime prevalence for cyberbullying was between 4.9% and 65.0%, while those of perpetration throughout life varied between 1.2% and 44.1% (Brochado et al., 2017). The review showed that the proportion of victims/perpetrators of last-year cyberbullying were between 1.5% and 72.0%, and lifetime cyberbullying ranged between 5.0% and 64.3% for perpetration throughout life (Brochado et al.,2017). However, researchers noticed that one in five children was involved in cyberbullying (Hamm et al., 2015; Modecki et al., 2014).

Studies are fewer in Latin America and highly heterogeneous in their methods and the findings, possibly due to the use of different instruments to measure the phenomenon. For example, in Argentina, the prevalence of cyberbullying was around 49% (Laplacette et al., 2011), while in Mexico (Castro., 2013), Uruguay (Lozano et al., 2011), or Chile (Varela et al., 2014), it did not exceed 15%.

Cyberbullying is a phenomenon that has not been widely studied in Colombia. The few studies conducted here reported the prevalence of between 13.6 and 60% (Aranzales Delgado et al., 2014; Mura & Diamantini, 2013; Yudes-Gómez et al., 2018). Regardless of the country or culture, cyberbullying is a social problem whose prevalence has increased considerably in recent years (Brochado et al., 2017). The theoretical perspective of bullying and, consequently, cyberbullying has expanded in the last two decades to a set of risk factors that mediate victimisation, such as the environment (proximity) and individual characteristics of the potential victims (ability to protect themselves). Among the victim’s characteristics, such as physical weakness or mental health problems, on the one hand, they can favour victimisation since they compromise the ability to face or resist (Finkelhor, 2007). On the other hand, these particularities may partially explain the negative psychosocial consequences of different bullying forms, such as poor academic performance, substance use, or high suicide risk (Brailovskaia et al., 2018; Fisher et al., 2016).

Several investigations have shown that cyberbullying is associated with various negative consequences for victims (Kowalski et al., 2019; Ortega et al., 2012). These include high levels of anxiety, depression, suicidal ideation and acts, low self-esteem levels, poor social and psychological adjustment, isolation, anger, absenteeism from school, and low academic performance (Albdour et al., 2019; Chu et al., 2018; Cole et al., 2016; Didden et al., 2009; Hamm et al., 2015; Kowalski & Limber, 2013; Landoll et al., 2015; Tian et al., 2018; Wang et al., 2019; Yen et al., 2014).

From another perspective, in recent years, researchers have analyzed the variables that predict that a person is involved in cyberbullying as a victim at the individual, school, family, and community level (Kowalski et al., 2019). The probability of someone being a victim of cyberbullying depends on several variables: the time spent on the Internet (Sasson & Mesch., 2017; Shapka et al., 2017), gender (Caravaca et al., 2016; Morales & Rueda, 2018), race/ethnicity (Lapidot-Lefler & Hosri, 2016; Mohammed & Bellamy, 2015), sexual orientation (Cénat et al., 2015; Elipe et al., 2017), suffering a disability or chronic condition (Bevilacqua et al., 2017; Carrington et al., 2017), body weight (Kenny et al., 2017), low self-esteem (Chu et al., 2019), perception of peer support (Kowalski et al.,2019), family structure (Bevilacqua et al., 2017), family environment (Chen et al., 2018), parental bond (Sampasa-Kanyinga, Lalande, & Colman., 2018), sexual abuse history (Hébert et al., 2016), and neighbourhood climate (Khoury-Kassabri et al., 2019).

Factors associated with involvement in cyberbullying perpetration is also of vital importance. In this regard, a cross-sectional survey of 2590 middle-school students in Korea using a binomial regression model assessed predictors of cyberbullying perpetration; using a logit model, they showed that being male, offline delinquency, aggression, weekend smartphone use, weekend computer use and hostile parenting increased the likelihood of engaging in cyberbullying as a perpetrator (Kang et al., 2021).

In recent years, one of the focuses of cyberbullying has been establishing the prevalence and associated factors in different countries to determine cultural comparisons (Brochado et al., 2017; Hamm et al., 2015; Kowalski et al., 2019; Modecki et al., 2014). The topic has to be more widely examined in different contexts to design specific preventive actions based on evidence from a public health perspective. This work could increase the current level of knowledge about cyberbullying factors in adolescents in Colombian schools. This low and middle-income country usually generates little research.

The study’s purpose was to examine the prevalence of cyberbullying victimisation and analyze its association with demographic variables, general health condition, family dysfunction, alcohol drinking, cigarette, cocaine and cannabis smoking, smoking, post-traumatic stress disorder risk, depression risk, and suicide risk in a probabilistic sample of high-school adolescents at public and private institutions in the urban area of ​​Santa Marta, Colombia.

Method

Design

A secondary analysis of a cross-sectional study is presented in which the prevalence and factors associated with poly-victimisation are analyzed. Situations of victimisation included cyberbullying. This report focuses exclusively on cyberbullying.

Participants

The adolescents in the sample were high-school students aged between 13 to 17 years of institutions in the urban area of ​​Santa Marta, Colombia. Sampling calculations showed that in 2016, there were 10,810 students in grades ten and eleven. A multi-stage probabilistic sampling for clusters was conducted. Each classroom represented a cluster with 25 students. A sample size of 1948 students was estimated. This number of students included a 20% replacement in possible losses (García-García et al., 2013). Seventy-eight schools were taken out of 157 possible. It was excluded students with limitations that impeded them from completing the self-administered booklet.

Instruments

Participants completed information about age, gender, educational level, socioeconomic status, perceived academic performance, perceived general health condition, lifetime substance use, and the item for cyberbullying victimisation of the Youth Risk Behavior Survey Questionnaire for high school students designed by the Center for Disease Control of the United States of America (Have you ever been electronically bullied? Count being bullied through texting, Instagram, Facebook, or other social media). The survey is a standardized instrument widely used in a global context which allows more accurate comparisons (Brener et al., 2002). This instrument has shown acceptable reliability in different populations (Bae et al., 2010; Zullig et al., 2006), and it had been previously adapted and applied to Colombian adolescents (Campo-Arias et al., 2008). The use of specific questionnaire items does not affect its reliability (Bae et al., 2010; Campo-Arias et al., 2008).

The following scales were also used to measure different variables: a family dysfunction scale, risk of depressive episode scale, suicide risk scale, and post-traumatic stress disorder risk scale.

The Family APGAR questionnaire was used to quantify family dysfunction. It consists of five items that investigate the perception of family functioning during the last six months. For each item, five response options, on a scale of zero to four, are provided, with possible total scores ranging, which are rated from zero to four. The total score can be between zero and twenty. A score below fifteen indicated family dysfunction (Smilkstein, 1978). This questionnaire showed adequate internal consistency in Colombian adolescents, with a Cronbach’s alpha of 0.79 (Forero-Ariza et al., 2006).

The WHO-5 questionnaire was used to estimate the depressive disorder risk. This five-item instrument explores symptoms that may have occurred during the previous two weeks. The WHO-5 presented five response options on one to four, with possible total scores ranging from five to twenty. A score below nine indicated a risk of depressive disorder (World Health Organization Regional Office for Europe, 1998). This inventory has shown excellent psychometric performance in Colombian adolescents (Campo-Arias et al., 2015).

The Brief Davidson Trauma Scale was used to explore post-traumatic stress disorder risk. This scale consists of four items that offer five response options on a scale of zero to four. A score higher than twelve indicated a high risk of post-traumatic stress disorder (Davidson et al., 1997). In Colombia, in a sample of people older than 15 years, the extended version showed excellent internal consistency, with a Cronbach’s alpha of 0.97 (Pineda Salazar et al., 2002).

The Center for Epidemiological Study on Depression’s Suicidal Ideation Scale was used to measure suicide risk. This scale has four items focusing on suicidal manifestations during the previous week. The instrument offers four response options on a scale of one to four, with total scores ranging between four and sixteen. A score higher than eight indicated a high suicide risk (Roberts, 1980). This scale has shown excellent performance in a previous study with Colombian adolescents (Pineda-Roa et al., 2018). Given that the questionnaire was anonymous, it was impossible to know the high-risk students; Therefore, global reports were sent to the participating schools and the municipality’s education secretariat to inform the findings and the need to carry out the indicated actions. Additionally, reports were sent to school directors who expressly requested the results of their schools.

Procedures

Students completed the research booklet in the classroom under the supervision of a research assistant assigned by the educational institution’s board. Completing the questionnaire took them an average of 30 min.

Analysis of Data

Descriptive statistics included observation of the frequency distributions of the variables studied. Odds ratios (OR), with a 95% confidence interval (95% CI), estimated the association between the variables considered and cyberbullying. The associations were adjusted using logistic regression. This process followed Greenland’s recommendation to retain only the following variables within the model: those that showed probability values of less than 20% and those significantly associated with cyberbullying or produced a variation more significant than 10% (Greenland, 1989). The Hosmer-Lemeshow goodness of fit estimated the best final model (Hosmer et al., 1991). The authors computed internal consistency for them with Cronbach’s alpha (Cronbach, 1951). The analysis was performed in IBM-SPSS version 23 (IBM Corp. (2015).

Ethical Issues

The Research Ethics Board at Universidad del Magdalena, Santa Marta, Colombia, approved the project. The parents of the participants signed an informed consent form, and the students gave the respective voluntary consent as required by Colombian law (Ministerio de Salud de Colombia, 1993).

Results

A group of 1547 students returned the completed questionnaire-survey, of which the authors discounted 85 for being incomplete (5.5%). Thus, the report summarizes the responses of 1462 students. The mean age of the participants was 14.4 years (SD = 0.8). A total of 287 students (19.6%) reported lifetime cyberbullying victimisation. Table 1 presents the other variables. See Table 2 for range, mean (SD), and Cronbach’s alpha for scales.

Table 1.

Characteristics of adolescent students

Variable	n	%
Age (years)
Between 13 and 15	425	29.1
Between 16 and 17	1037	70.9
Gender
Female	882	60.3
Male	580	39.7
Grade
Tenth	809	55.3
Eleventh	653	44.7
Income
Low	725	49.6
Middle	508	34.7
High	39	2.7
No response	190	13.0
Ethnic/race (self-identification)
Native	19	1.3
Afro-Colombian	151	10.3
Mulato	85	5.8
Caucasian	349	23.9
Mestizo	785	53.7
Other	55	3.8
No response	18	1.2
Academic achievement
Excellent-Good	1012	69.2
Fair-Poor	450	30.8
General health condition
Excellent-Good	1101	75.3
Fair-Poor	361	24.7
Family dysfunction
Yes	1112	76.1
No	350	23.9
Lifetime alcohol drinking
Yes	330	22.6
No	1132	77.4
Lifetime cigarette smoking
Yes	328	22.4
No	1134	77.6
Lifetime cannabis smoking
Yes	170	11.6
No	1292	88.4
Lifetime cocaine smoking
Yes	40	2.7
No	1422	97.3
Post-traumatic stress disorder risk
Yes	283	19.4
No	1179	80.6
Depression risk
Yes	103	7.0
No	1359	93.0
Elevated suicide risk
Yes	194	13,3
No	1268	86,7

Table 2.

Range, mean (SD), and Cronbach’s alpha for scales

Scale	Range	M (SD)	Cronbach’s alpha
Family APGAR questionnaire	0–20	12.4 (4.2)	0.82
WHO-5	5–20	12.6 (3.2)	0.82
Davidson’s brief scale	4–20	9.7 (3.3)	0.66
CES-D-SI	4–16	5.9 (2.6)	0.75

In the bivariate analysis, cyberbullying victimisation was associated with female gender, poor-fair general health condition, family dysfunction, lifetime alcohol drinking, lifetime cigarette smoking, elevated suicide risk, and high depression risk. See Table 3. However, the multivariate analysis, lifetime alcohol consumption, and high suicide risk lost statistical significance. The model fits appropriately (Hosmer-Lemeshow’s test chi-square = 8.39, df = 7, p = 0.30). See Table 4.

Table 3.

Variables associated with cyberbullying victimization in adolescent students

Variable	OR	95%CI	p value
Age between 16 and 17	1.28	0.95–1.72	0.780
Female gender	1.88	1.42–2.49	0.001
Grade eleventh	1.03	0.79–1.33	0.810
Low income	1	–
Middle income	1.09	0.47–2.53	0.834
High income	1.31	0.56–3.04	0.533
Fair or poor academic achievement	1.12	0.85–1.48	0.419
Fair or poor general health condition	1.95	1.47–2.58	0.001
Family dysfunction	2.01	1.42–2.85	0.001
Lifetime alcohol drinking	1.78	1.26–2.52	0.001
Lifetime cigarette smoking	2.04	1.54–2.71	0.001
Lifetime cannabis smoking	1.92	1.34–2.74	0.001
Lifetime cocaine smoking	3.05	1.86–6.63	0.001
Depression risk	1.19	0.74–1.93	0.475
Elevated suicide risk	1.77	1.25–2.49	0.001
Post-traumatic stress disorder risk	2.66	1.99–3.55	0.001

Table 4.

Multivariate model 1 for variables associated with cyberbullying victimization in adolescent students

Variable	OR	95%CI	p value
Post-traumatic stress disorder risk	2.05	1.51–2.79	0.001
Lifetime cigarette smoking	1.91	1.42–2.57	0.001
Female gender	1.68	1.25–2.26	0.001
Family dysfunction	1.68	1.18–2.41	0.005
Fair or poor general health condition	1.45	1.08–1.95	0.014

Discussion

This study indicates a high lifetime prevalence of cyberbullying victimisation (19.6%) among high-school adolescents included in this sample. Being a victim of cyberbullying was statistically associated with the risk of post-traumatic stress disorder, lifetime cigarette smoking, female gender, family dysfunction, and poor-fair general health condition.

As shown in studies involving similar populations, having been a victim of cyberbullying at some time in life is a frequent phenomenon among adolescents. For example, González-Cabrera et al. (González-Cabrera et al., 2018) found a prevalence of cyberbullying of 19.5%, which coincides with the figure found in this study of 19.6%. This high prevalence of cyberbullying victimisation raises the need to develop preventive actions for this phenomenon and identify possible risk and protective factors (Espelage & Hong, 2016).

Post-traumatic stress disorder (PTSD) symptoms are common in adolescent victims of cyberbullying (Ranney et al., 2016). Likewise, in the present study, post-traumatic stress disorder risk was associated with being a victim of cyberbullying. However, this association’s direction is not clear; thus, further studies are required to analyze this association. According to DSM-5, PTSD follows exposure to a “traumatic event,” which can generate clinical PTSD symptoms (American Psychiatric Association, 2013). Since psychiatric interviews did not confirm the diagnosis, the suggestive symptoms of PTSD reported by victims of cyberbullying may not correspond to PTSD. They may perhaps be better interpreted as unspecified anxiety disorders or adjustment reactions (Bottino et al., 2015).

The findings of the present study show an association between lifetime cigarette smoking and cyberbullying victimisation. Other studies in similar populations have found that being a victim of cyberbullying is associated with lifetime cigarette smoking (Sourander et al., 2010). This association could be explained by the fact that cyberbullying victims use cigarettes as a coping strategy for their psychological distress (self-medication hypothesis) (Case et al., 2016).

The study data showed an association between the female gender and having been a victim of cyberbullying. The results presented are in the same line as previous studies showing that women are more likely to report being victims of cyberbullying than men (Bor et al., 2014; Sampasa-Kanyinga, Lalande, & Colman., 2018). Additionally, the gender variable mediates the outcomes associated with victimisation. Thus, it has been observed that exposure to cyberbullying increases the risk of psychological distress, suicide ideation, and substance use in female adolescents more than in their male peers (Kim et al., 2018).

Because female adolescents present an increased risk for depression in the context of interpersonal relationships, it is plausible that this susceptibility may extend to cyberbullying victimisation (Bor et al., 2014; Hammen, 2009). Other works have shown that female adolescents are more vulnerable to interpersonal stress (Hammen, 2009). Their greater vulnerability to interpersonal anxiety and depression may mean that the risk of psychological distress in adolescents may exacerbate in the context of cyberbullying in females more than in males (Bor et al., 2014). The difference in exposure to cyberbullying between men and women can be partially explained using gender-specific social network patterns. For example, evidence indicates that female adolescents are more likely to use social media for their relationships than male adolescents (Barlett & Coyne, 2014), and they tend more often to be the victims of relational aggression (Carbone-Lopez et al., 2010). However, there is likely under-reporting of exposure to cyberbullying among men, which may partly contribute to the lower prevalence of cyberbullying among men (Sampasa-Kanyinga et al., 2018). These findings support the need to develop preventive interventions with a differential gender approach in adolescents.

According to our data, there is an association between family dysfunction and being a victim of cyberbullying. Similarly, previous studies have found a significant association between cyberbullying victimisation and family functioning variables (Elgar et al., 2014). Adequate family functioning can act as a protective factor since it facilitates parents’ guidance and support of adolescents, more open communication between parents and children, and offers adolescents the possibility to express their problems (Espelage & Hong, 2016). Therefore, good family dynamics can promote adolescent health and cushion the impact of stressful situations such as cyberbullying on adolescent functioning (Musick & Meier, 2012).

Adolescents involved in cyberbullying victimisation, like other victims, are at increased risk of developing adverse outcomes in their general health condition (Nixon, 2014). The present study documented an association between being a victim of cyberbullying and poor-fair health conditions. Dimensions of general health often include physical and mental well-being. Longitudinal studies indicate that cyberbullying victims are more likely to report fair well-being conditions (Fahy et al., 2016). For example, Kowalski et al. (2013) also found a strong association between an individual’s general health condition and cyberbullying victimisation. A more recent study identified a significant interaction between physical symptoms and cyberbullying (Ranney et al., 2016). Furthermore, several other investigations suggest that cyberbullying victims frequently experience somatic symptoms, such as sleep problems, headaches, fatigue, poor appetite, skin problems, abdominal pain, and enuresis (Kowalski & Limber, 2013; Mohammed & Bellamy, 2015). These findings indicate that cyberbullying affects adolescents’ mental and physical health.

Several studies have found a significant association between depressive symptoms and cyberbullying victimisation in adolescents (Cole et al., 2016; Roberts, 1980; Sasson & Mesch, 2017; Wang et al., 2019). However, in this work, no statistically significant association was found between depressive symptoms and being a victim of cyberbullying. One possible explanation for this difference is that adolescents’ emotional responses to cyberbullying vary in intensity and quality. Thus, in a multicenter study that evaluated cyberbullying’s emotional impact, 24.5% of adolescents did not report depressive symptoms (Shapka et al., 2017). The effect of cyberbullying victimisation generates depressive symptoms for most victims; however, factors such as resilience, positive assessment of the stressful situation, self-esteem, self-efficacy, and self-pity could minimize emotional damage in some victims of cyberbullying (Shapka et al., 2017). There may be other moderating factors of depression in adolescents exposed to cyberbullying, such as social anxiety (Landoll et al., 2015; Sourander et al., 2010) and hopelessness (Chu et al., 2018). This study avoided the measurement of these symptoms. As such, future research should assess resilience factors that can cushion the impact of negative experiences, such as cyberbullying in adolescents (Fahy et al., 2016).

In this sample, lifetime alcohol drinking was associated with being a victim of cyberbullying (OR = 1.78, 95%CI 1.26–2.52), but this association was not significant in the multivariate model. These findings are consistent with those of Chang et al. (2013); they found that cyberbullying victimisation was not associated with excessive alcohol consumption, nor was alcohol consumption related to cyberbullying over time. Likewise, a study in high school students found that being a victim of cyberbullying was not associated with alcohol consumption after controlling for confounding variables (Chan & La Greca, 2015). In contrast, being a victim of cyberbullying is related to various alcohol consumption patterns in adolescents (Fisher et al., 2016). These differences reveal a complicated relationship between alcohol consumption and cyberbullying. Some observations suggest that there could be other variables that moderate this complicated relationship (Chan et al., 2019; Rodríguez-Enríquez et al., 2019). For example, depression may be a moderating factor in the association between alcohol use and being a victim of cyberbullying, explaining the lack of association in some studies (Kim et al., 2019). So, the complexity of the relationship between alcohol drinking and cyberbullying merits further research with more robust statistical models.

This study’s strengths included an appropriate sample size to detect differences and robust statistical analysis methods that allowed strong associations. Finally, the study’s youth risk behaviour survey allows comparisons with other worldwide studies; the questionnaire is broadly accepted and used extensively.

The following are some of the limitations that should be considered when interpreting the findings of this study. Firstly, studying cyberbullying with a simple question measure does not allow the detection of all the commonly used criteria for evaluating cyberbullying, such as status, victim and aggressor, types of victimisation, etcetera. Second, the study’s cross-sectional nature does not establish a causal relationship between the variables studied and cyberbullying. Third, the present study only included self-report measures, which may introduce non-measurable biases. Furthermore, cyberbullying used only one item and did not consider other tests, such as reports from family members or teachers, which could have increased the results’ validity. Finally, doctors, psychiatrists, or psychologists did not confirm the physical and psychological self-reported complaints.

The findings have several important implications. Rapid advances in information technology and the increasing popularity of social media pose a crucial need for efficient strategies to assess cyberbullying and associated physical and mental health problems. Preventive interventions start from a context-specific needs assessment, and a study of this nature provides critical inputs for developing prevention strategies.

Additional research that expands on these findings is needed to corroborate the association of cyberbullying with other psychological problems such as anxiety, depression, and suicidal behaviours that have been associated with being a victim of cyberbullying in other studies (Bottino et al., 2015; Carvalho et al., 2018) and that are a growing concern in Colombian adolescents. Longitudinal investigations will determine whether cyberbullying victimisation is associated with adverse health conditions over time. Considering that cyberbullying is related to traditional bullying (Herrera-López et al., 2017), future studies are required to explore whether traditional bullying in schools is associated with cyberbullying and vice-versa. More research is required to broaden our understanding of cyberbullying dynamics since knowledge of risk and protective factors can develop capable assessment instruments and cyberbullying prevention strategies.

Conclusions

Cyberbullying victimisation is frequent in the sample of school adolescents from Santa Marta and is associated with post-traumatic stress disorder risk, cigarette smoking, female gender, family dysfunction, and poor-fair general health condition. Longitudinal investigations and more rigorous methods are needed to measure better the dynamics of cyberbullying and establish the nature of such associations.

Funding

The Research Vice-rectory of the University of Magdalena financed this research through Resolution 0347 of 2018 (Call Fonciencias 2017).

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Conflict of Interest

The authors have no conflicts of interest to declare.