Bullying Victimization at School and Headache: A Meta-Analysis of Observational Studies


  • Conflict of Interest: None.
  • Funding Source: No external funding was secured for this study.


Background and objectives

Being bullied at school is a risk factor for a variety of negative consequences, including somatic problems. The purpose of this meta-analysis is to determine the association between peer victimization and headache in the school-age population.


A systematic literature search was conducted in September 2013 to identify observational studies that examined the association between being bullied and headache in children and adolescents. Odds ratios (OR) were pooled by using a random-effects model. Moderator and sensitivity analyses were conducted.


Twenty studies, including a total of 173,775 participants, satisfied the pre-stated inclusion criteria. Fourteen studies reported data on the prevalence of headache, which was on average 32.7% (range: 9.1-71.7%) in the bullied group and 19.1% (range: 5.3-46.1%) in the control group. Two separate meta-analyses of the association between being bullied and headache were performed on 3 longitudinal studies (OR = 2.10, 95% confidence interval = 1.19-3.71) and 17 cross-sectional studies (OR = 2.00, 95% confidence interval = 1.70-2.35), respectively. Results showed that bullied children and adolescents have a significantly higher risk for headache compared with non-bullied peers. In the cross-sectional studies, the magnitude of effect size significantly decreased with the increase of the proportion of female participants in the study sample. No further moderators were statistically significant.


The positive association between bullying victimization and headache was confirmed. Further research on the environmental factors that may influence this symptom is needed.

95% CI

95% confidence interval


Meta-analysis of Observational Studies in Epidemiology


odds ratio


tension-type headache


socioeconomic status

Recurrent headache is the most frequent neurological symptom during school age and one of the most frequent manifestations of pain in childhood and adolescence.1-3 A recent systematic review[4] showed that headache is very common across the world with about 60% of children and adolescents reporting this symptom over at least a 3-month period. Moreover, epidemiological studies pointed out that the prevalence of headache has increased over the last decades in the school-age population.5-8 Quite recently, studies on the potential risk factors for youth's headache have drawn attention to the role of psychological and social factors, including negative experiences at school.9-12 For example, stressors in the school environment, such as schoolwork pressure,[13] negative feelings about school,[14, 15] perception of being treated badly or unfairly by teachers,11-13 fear of failure,[16] and harassment by peers[13, 16] turned out to be associated with higher levels of headache in children and adolescents.

A serious and frequent source of concern in children's and adolescents' school life is bullying, that is, a repetitive physical or psychological abuse by a stronger schoolmate or group on a weaker peer.[17, 18] Epidemiological studies across countries indicate that 10-20% of students are frequently bullied by schoolmates.[18, 19] Importantly, pediatric and psychological research is increasingly demonstrating the adverse consequences of being bullied at school for children's and adolescents' psychosocial adjustment,[20, 21] health,[22, 23] and medicine use.[24] The well-established cognitive appraisal model of stress and coping developed by Lazarus[25] and the “sustained activation hypothesis”[26] can help to understand stress reactions to bullying. These models suggest that repeated bullying experiences in children's life might cause a state of emotional distress that can lead to adverse health outcomes such as recurrent headache.

To date, 2 meta-analyses[22, 23] have shown that bullied students can be affected by poor physical health and that these youths are about 2 times more likely than non-bullied agemates to report a variety of symptoms, such as headache, backache, abdominal pain, skin problems, vomiting, etc. However, both these meta-analytic reviews only reported an overall risk estimate for victims' health problems and did not specifically focus on headache. The current meta-analysis aims at (1) estimating the risk for headache in children and adolescents who are bullied by peers (ie, victims) compared with non-bullied peers; (2) performing separate meta-analyses of longitudinal and cross-sectional studies; (3) testing for potential moderators of variation in the magnitude of effect sizes, that is, testing whether certain study features explain differences in the strength of the effect sizes.


Literature Search

Several methods were used to identify relevant studies. First, electronic searches in PsychInfo, Pubmed, EMBASE, the Cochrane Library database, the Campbell Collaboration database, and Scopus were conducted in September 2013 with the following keywords: “bullying,” or “peer victimization” and “headache,” “somatic,” “psychosomatic,” and “physical health.” Second, the “cited by” function in Scopus was used to retrieve empirical articles that have cited the 2 meta-analyses[22, 23] on the association between bullying and health problems. Third, previous issues of the journal “Headache” were searched for relevant studies. Finally, review articles regarding consequences of bullying and the reference sections of the collected articles were reviewed for possible relevant citations. If a study was not available in full-text, the corresponding author was contacted. This meta-analysis was planned, conducted, and reported in adherence to the Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines.[27]

Inclusion Criteria

A study had to meet the following a priori criteria to be included. The most basic requirement was the inclusion of measures of bullying victimization at school in childhood or adolescence and of headache. Consistent with the international literature, bullying was defined as a deliberate, repeated exposure to aggressive acts performed by a peer or a group of peers with higher power or strength than the bullied schoolmate (ie, the “victim”).[17, 18] These measures could include (1) self-report questionnaires; (2) peer or adult reports; or (3) an interview that resulted in a quantitative rating of peer victimization and headache frequency. The majority of studies defined frequent headache as occurring “about every week/at least weekly,” with few others using a slightly broader definition (“at least monthly/several times in the last 12 months”). Second, studies were required to have reported effect sizes and related confidence intervals or enough information to calculate these data – for example, by reporting comparisons between bullied children and a control group (defined as children from the same population of victims who were classified as not bullied). Both cross-sectional and longitudinal studies were included. We excluded the following types of studies: studies that did not include a control group; studies that measured headache with items included in a larger scale, as this problem could not be clearly distinguished from other symptoms; studies with duplicated data; studies that did not report analyses on the variables of interest; and studies with adults or psychiatric patients. Two authors (GG, TP) independently assessed whether articles met the inclusion criteria. In the case of disagreement, a consensus was reached through discussion.

Coding of Studies

Studies were coded on design (cross-sectional vs longitudinal), length of follow up for longitudinal studies, type of bullying and of symptoms measure (self-report questionnaire vs peer/adult reports vs interview), confounding variables (eg, age, gender), type of sampling procedure, sample composition and characteristics, and geographical location of study. Two authors (GG, TP) independently coded the studies. Quantitative data were extracted from text and tables; for the sake of comparability with the results of the former meta-analyses,[22, 23] the data adjusted for confounders were preferred.

Statistical Analyses

Analyses were done using Comprehensive Meta-Analysis.[28] We extracted odds ratio (OR) and their 95% confidence interval (CI) from each study. With very few exceptions, studies did not report results for boys and girls separately; therefore, we were not able to compare effect sizes by gender group. Because most of the studies reported the proportion of girls in the sample, we used this information to test for possible moderation by gender composition of the sample.

Data from individual studies were pooled using a random-effects model. Each study was weighted by the inverse of its variance, which, under the random-effects model, includes the within-study variance plus the between-studies variance tau-squared (Τ2). The Z statistic was calculated, and a 2-tailed P value of less than .05 was considered to indicate statistical significance. Statistical heterogeneity was assessed using the Q statistic to evaluate whether the pooled studies represent a homogeneous distribution of effect sizes. Also reported is the I2 statistic, indicating the proportion of observed variance that reflects real differences in effect size.[29]

To address the possible “publication bias” – that is, the fact that studies with non-significant results are less likely to be published – we computed the “fail-safe N” (Nfs) according to the method that Orwin[30] proposed, which is more conservative than the traditional Rosenthal's Nfs.[31, 32] Orwin's Nfs determines the number of additional studies in a meta-analysis yielding null effect sizes that would be needed to yield a “trivial” OR of 1.05. Researchers suggest that meta-analysts calculate a tolerance level around a fail-safe N that is equal to 5 times the number of effects included in the meta-analysis plus 10 (the “5k + 10” benchmark).[32, 33] Moreover, the association between the standardized effect sizes and the variances of these effects was analyzed by rank correlation with use of the Kendall tau method. If small studies with negative results were less likely to be published, the correlation between variance and effect size would be high. Conversely, a lack of a significant correlation can be interpreted as the absence of publication bias.[34]


Study Identification and Characteristics

After the removal of duplicates, a list of 137 potentially eligible studies was generated (Fig. 1). Based on titles and abstracts, 57 articles were excluded at the first screening because they were qualitative studies, reviews or commentaries, or studies that did not measure school bullying. Seven studies35-41 were not available in full text. Full-text copies of the remaining 73 potentially relevant studies were obtained. Thirty-seven studies were excluded because they did not meet the inclusion criteria (eg, they did not have a control group). Fifteen studies did not report enough data to compute effect sizes or confidence intervals. As a result, the remaining 20 studies were included for this meta-analysis. Three studies were longitudinal studies, and 17 employed a cross-sectional design.

Figure 1.

Flow diagram of study inclusion. ES = effect size.

The Table summarizes the characteristics of the studies included in this meta-analysis, including sample-size and response rate, age and gender composition of the sample, type of measures, study design, and type of sampling. A total of 173,775 children and adolescents participated in the 20 studies. Across the 17 studies that provided information about the sample's gender composition, 51.3% (range: 32.8-62.4%) of the participants were girls. Fourteen studies reported data on the prevalence of headache, which was on average 32.7% (range: 9.1-71.7%) in the bullied group and 19.1% (range: 5.3-46.1%) in the control group.

Table –. Characteristics of the Studies Included in the Meta-Analysis
First Author (Year of Publication)Sample Size (Response Rate)Age Range (% of Girls)Bullying MeasureHeadache MeasureAdjustment for ConfoundersStudy DesignType of Sampling
  1. SES, socioeconomic status.
Arslan et al (2012)[50]1315 (97%)11-19 (53.4)Self-report questionnaireSelf-report questionnaireSES, school grade, genderCross-sectionalCluster random sampling
Biebl et al (2011)[52]65 (n/a)12-20 at time 3 (52.9)Time 1: play session; time 3: self-report questionnaireSelf-report questionnaireGenderLongitudinalConvenience sample
Due et al (2005)[18]123,227 (>90%)11-15 (51)Self-report questionnaireSelf-report questionnaireAge, family affluence, countryCross-sectionalCluster random sampling
Fekkes et al (2004)[48]2766 (100%)9-12 (50)Self-report questionnaireSelf-report questionnaireGenderCross-sectionalUnknown
Fekkes et al (2006)[49]1118 (70%)9-11 (50.3)Self-report questionnaireSelf-report questionnaireGender, age, having friendsLongitudinalUnknown
Gini (2008)[56]565 (94%)8-11 (52.9)Self-report questionnaireSelf-report questionnaireGender, ageCross-sectionalSimple random sampling
Haavet et al (2004)[42]8316 (88%)15 (54.4)Self-report questionnaireSelf-report questionnaireNoneCross-sectionalPopulation study
Hesketh (2010)[54]2191 (80%)9-12 (44)Self-report questionnaireSelf-report questionnaireGender, age, residence, parental educationCross-sectionalSimple random sampling
Karatas and Ozturk (2011)[51]92 (82%)10-12 (51.1)Self-report questionnaireParent-report questionnaireNoneCross-sectionalSimple random sampling
Kshirsagar (2007)[46]500 (100%)8-12 (62.4)Semi-structured interviewSemi-structured interviewNot specifiedLongitudinalSimple random sampling
Lien et al (2009) (sample 1)[67]3790 (88%)15-16 (49.3)Self-report questionnaireSelf-report questionnaireFamily structure, SES, ethnicity, exposure to violence, having close friendsCross-sectionalPopulation study
Lien et al (2009) (sample 2)[67]3790 (80%)18-19 (55.9)Self-report questionnaireSelf-report questionnaireFamily structure, SES, ethnicity, exposure to violence, having close friendsCross-sectionalPopulation study
Løhre et al (2011)[43]419 (100%)7-16 (n/a)Multi-informant (self-, teacher-, parent-reports)Self-report questionnaireGender, gradeCross-sectionalConvenience sample
Luntamo et al (2012)[44]2215 (91%)13-18 (50)Self-report questionnaireSelf-report questionnaireNoneCross-sectionalPopulation study
Natvig et al (2001)[45]856 (83.7%)13-15 (50.6)Self-report questionnaireSelf-report questionnaireGender, age, schoolCross-sectionalUnknown
Ramya and Kulkarni (2011)[47]500 (n/a)8-14 (32.8)InterviewInterviewNoneCross-sectionalSimple random sampling
Schnohr and Niclasen (2006)[55]891 (n/a)11-15 (n/a)Self-report questionnaireSelf-report questionnaireGender, ageCross-sectionalCluster random sampling
Srabstein et al (2006)[53]15,305 (83%)11-15 (53.5)Self-report questionnaireSelf-report questionnaireGender, age, race, overweight/obesity, maternal educationCross-sectionalCluster random sampling
Stickley et al (2013)[58]2892 (96%)13-17 (57.6%)Self-report questionnaireSelf-report questionnaireAge, parental education, family structureCross-sectionalSimple random sampling
Williams et al (1996)[57]2962 (93.1%)7-10 (n/a)Semi-structured interviewSemi-structured interviewNot specifiedCross-sectionalPopulations study

Five studies were from Norway,42-45 2 of which were from the same publication; 2 respectively from India,[46, 47] the Netherlands,[48, 49] Turkey,[50, 51] and the United States;[52, 53] and 1 respectively from China,[54] Finland,[44] Greenland,[55] Italy,[56] the United Kingdom,[57] and Russia.[58] One article reported data from multiple countries.[18] Information about race/ethnicity and socioeconomic status (SES) of the participants was not systematically reported in all studies. Overall, the heterogeneity of racial and SES classification within and across the studies was such that it precluded any analysis by race/ethnicity or SES.

Meta-Analysis of Longitudinal Studies

Three studies used a longitudinal design. The follow-up duration ranged between 9 months and 11 years. Across the 3 samples, bullied children were found to have a significantly higher risk for headache than non-bullied agemates were (OR = 2.10, 95% CI = 1.19-3.71, Z = 2.57, P = .01). Figure 2 shows the forest plot for this meta-analysis. Studies were not completely homogeneous (Q = 4.11, P = .13, I2 = 51.37%).

Figure 2.

Forest plot for random-effects meta-analysis of the association between being bullied and headache: longitudinal studies. Note. Effect sizes are expressed as odds ratios. Studies are represented by symbols whose area is proportional to the study's weight in the analysis. CI = confidence interval.

Meta-Analysis of Cross-Sectional Studies

Across the 17 samples that were included in the cross-sectional studies, bullied children were found to have a significantly higher risk for headache than were non-bullied peers (OR = 2.00, 95% CI = 1.70-2.35, Z = 8.43, P < .001). Figure 3 shows the forest plot for this meta-analysis. Effect sizes within this group of studies were not homogeneous (Q = 65.64, P < .001, I2 = 75.63%).

Figure 3.

Forest plot for random-effects meta-analysis of the association between being bullied and headache: cross-sectional studies.

Note. Effect sizes are expressed as odds ratios. Studies are represented by symbols whose area is proportional to the study's weight in the analysis. CI = confidence interval.

Moderator Analyses

Moderator analyses with gender composition of the sample, number of confounders, and geographical location were performed to explore possible explanations for heterogeneity in the effect sizes across cross-sectional studies. The proportion of girls in the sample was available for 15 out of the 17 cross-sectional studies, and it was used as a continuous predictor in a weighted mixed-effects meta-regression. Results indicated that the magnitude of the effect size significantly decreased with the increase of the number of female participants in the study sample (B = −.06, 95% CI: −.07 to −.04, P < .001). Conversely, the number of confounders considered in the study (range: 0-6) did not moderate the magnitude of the effect (B = .005, 95% CI: −.04 to .05, P = .82). Also the study's geographical location (coded as Europe vs other countries) was not a significant moderator (k = 11, OR = 2.03, 95% CI: 1.59-2.60, and k = 5, OR = 2.00, 95% CI: 1.32-3.02, respectively; Q = .48, P = .79).

Sensitivity Analysis

Finally, consistent with the MOOSE guidelines[27] and to the former meta-analyses,[22, 23] a sensitivity analysis was performed based on 2 aspects of study quality (beyond those required as inclusion criteria): (1) the use of a randomized sampling design or a whole population of students; and (2) a good response rate (>80%). Thirteen cross-sectional studies satisfied both criteria. We then performed a separate meta-analysis of this subgroup of studies, and the resulting OR and confidence interval was OR = 1.90, 95% CI = 1.61-2.25.

Another sensitivity analysis was performed with the 13 studies that used only self-report questionnaires to gather data from participants. Estimated OR was 1.87, with a 95% CI ranging from 1.57 to 2.23.

Publication Bias

No evidence of publication bias was present. Kendall's tau was 0.13 with 2-tailed P = .44. An additional 253 studies with null effect sizes would be needed to attenuate the effect size to a negligible value (“5k + 10” benchmark = 110).


The results of this meta-analysis confirmed that bullied youths are about twice more likely than non-bullied agemates to suffer from headache. Same results were found both in longitudinal and cross-sectional studies. The sensitivity analysis, which removed those studies with potential to introduce bias, revealed similar results. These findings confirm and complement those of 2 previous meta-analyses[22, 23] on the health consequences of peer victimization. The present study differs from those meta-analyses in important ways. First, only this review reports meta-analytic effects specifically for the association between bullying victimization and headache. Second, it includes 11 new studies that were not included in the first meta-analysis (% of overlap of the 2 meta-analytic databases: 40.9%), and 2 new studies that were not included in the latter one (% of overlap of the 2 meta-analytic databases: 42.8%). Third, another moderator that was not considered before was tested (ie, number of confounders). Finally, only in this work a sensitivity analysis with studies that used self-report measures was performed.

Similarly to what have been reported on a former meta-analysis,[23] a meta-regression analysis showed that the strength of the relationship between being bullied and suffering from headache is higher when samples contain proportionally more boys. A possible explanation of this finding might deal with the fact that a school/classroom environment with a higher proportion of male students is a social context in which bullying behavior is more likely to happen, and where supportive and helping behaviors in favor of the bullied pupils are less frequent.[59] This could increase the negative impact of being bullied on children's health status. However, given the explorative nature of the moderator analysis, a significant finding is not to be considered definitive, but it does suggest a direction for additional research. The influence of the school environment's gender composition on peer victimization and its consequences on children's well-being is a topic that warrants further research.

The current meta-analytic findings confirm for headache what have been reported for health problems in general. A possible explanation of this significant association may be related to the well-known psychological vulnerability of bullied students, given that research has shown a strong association between emotional difficulties and somatic complaints.[60] The feelings associated with being bullied represent a form of social pain, a term used to describe the feelings of pain that follow the experiences of peer rejection, ostracism, or loss.[61] Neuroscience research is increasingly showing that social pain is experienced in a similar way to physical pain, at least as far as the brain is concerned. Indeed, recent studies have shown that social pain and physical pain rely on similar neurobiological and neural substrates and are experienced physiologically in a similar manner.[61, 62] Similarly, peer victimization is linked to dysregulation of the hypothalamic-pituitary-adrenal axis, the body's stress response system.63-65 From a physiological perspective, peer victimization represents a relatively extreme and/or persistent stressor, which ultimately leads to lower cortisol levels compared with that of non-bullied peers.[64, 65]

Of course, other psychological characteristics of bullied youth may influence the relationship between bullying and health problems. For example, one may hypothesize that students who lack adequate coping skills, as well as have low self-esteem or lack assertiveness, in front of victimization experiences are at increased risk for negative outcomes compared to peers who possess more developed psychological and social competencies. This is certainly an interesting hypothesis that should be tested in future longitudinal studies.

Strengths and Limitations

This is the first meta-analytic study that estimated the relationship between being bullied and headache. Strengths of this meta-analysis include the large overall sample size and the wide geographic distribution of the samples, which support the generalizability of the overall findings. Moreover, the large majority of the studies included in the meta-analysis were characterized by good methodological quality, as defined, for example, by the use of a random sampling design. Furthermore, we did not find evidence of publication bias that may have led to overestimating the association between bullying experiences and headache. Finally, we were able to perform separate meta-analyses of longitudinal and cross-sectional studies, which yielded the same results, even though the lack of large longitudinal studies is still a limit of the literature in this field.

The results of this meta-analysis should be interpreted in the context of the study limitations. The fact that the available studies neither explicitly compared male and female samples, nor reported separate effects for different ethnic groups limited the possibility for more detailed analyses. In particular, youths' cultural background could influence how bullying victimization is experienced, as well as their ability to cope with it and the negative consequences that may arise from this socially adverse experience. Moreover, much variability exists in the methods and instruments used to assess the prevalence of headache and peer victimization experiences. The majority of studies used a variety of self-report questionnaires, both for peer victimization and for children's health complaints. In some cases, these measures were reduced to a single-item questionnaire. Self-report measures are very common in bullying research and are usually considered to be valid and reliable.[66] However, possible problems with these instruments are that they require a good level of respondents' self-consciousness and that some bullied children may tend to deny their condition. To avoid these problems, future studies should collect information about youths' bullying experiences through multiple independent informants, such as children themselves, their peers within the class, and their teachers or parents. Moreover, future investigations should also cover the issue of cyberbullying victimization, that is, victimization experienced through information technologies (ie, cell phones, Internet). Also, the assessment of children's health complaints must be improved. For example, none of the available studies included independent objective information, such as children's school absenteeism extracted from school attendance records or their visits to the school nurse office; further improvement on the accuracy of headache reports in these age groups would profit from the use of prospective measurement in diaries, instead of only retrospective recalls. Moreover, studies in this field do not report information on the type of headache (migraine vs tension-type headache [TTH]) suffered by bullied youth. It is important that future research works address this limitation by comparing the specific effects of bullying as a stressor on both migraine and TTH. Finally, our meta-analysis shares the same limitations of all meta-analyses of observational studies. Because individuals cannot be randomly allocated to groups, the influence of confounding variables cannot be fully evaluated. Although many studies controlled for important confounding variables, such as parental education and SES, other unknown confounders could be partially responsible for the effect observed.


Bullied youths are about 2 times more likely than non-bullied agemates to report frequent headache. This meta-analysis complements the growing body of research that documents the poor personal adjustment of bullied children and adolescents, in terms of both internalizing and externalizing problems, which other recent meta-analyses[12, 13] on the psychosocial consequences of peer victimization have summarized. It is important that pediatricians, school nurses, and other professionals be ready to identify children who are at risk of being bullied at school because the potential negative health, psychological, and educational consequences of bullying experiences are far reaching.

Statement of Authorship

Category 1

  • (a)Conception and DesignGianluca Gini; Tiziana Pozzoli; Michela Lenzi; Alessio Vieno
  • (b)Acquisition of DataGianluca Gini; Tiziana Pozzoli
  • (c)Analysis and Interpretation of DataGianluca Gini; Tiziana Pozzoli

Category 2

  • (a)Drafting the ManuscriptGianluca Gini; Tiziana Pozzoli; Michela Lenzi; Alessio Vieno
  • (b)Revising It for Intellectual ContentGianluca Gini; Tiziana Pozzoli; Michela Lenzi; Alessio Vieno

Category 3

  • (a)Final Approval of the Completed ManuscriptGianluca Gini; Tiziana Pozzoli; Michela Lenzi; Alessio Vieno