Childhood clumsiness and peer victimization: a case–control study of psychiatric patients

All participants in this study were consecutive admissions referred to a specialized ASD and ADHD outpatient psychiatric clinic for diagnosing and treating adult ADHD or ASD at St. Göran Hospital, Stockholm, Sweden. Almost all of the patients were tertiary referrals from licensed psychiatrists or psychologists and almost half of the patients had received psychiatric care in childhood. No exclusion criteria were used but patients with obvious alcohol or drug dependence were referred elsewhere. Various treatment programs, including cognitive-behavioral therapy, support groups, medication, and courses about ADHD and ASD were provided to patients and their family members at the unit. A semi-structured protocol, including questions on civil status, educational level, employment status, current medication, psychiatric care in childhood and adulthood, and peer victimization, was utilized in the assessment of the patients. In 277 patients (133 males and 144 females, age range 18–57 years), information about peer victimization was available. In a subset of 235 individuals, a questionnaire on childhood symptoms was completed by a parent, and a motor function examination was performed in 206 patients (Figure 1).

All of the patients were assessed with a standardized research protocol to collect data on various aspects of adults with neurodevelopmental disorders. This included a diagnostic interview completed by a board-certified psychiatrist and psychologist. The assessment procedure took 12–18 h to complete over a period of 2–3 weeks. A parent, or in a few cases a significant other, was interviewed in person about early signs of ASD and ADHD based on the instrument Five to Fifteen (FTF) [33, 34]. This questionnaire includes neuropsychiatric symptoms noted by parents when the child was between 5 and 15 years of age. It was sent home to the parent prior to the first visit and the results were discussed with the parent and patient together at the consultation with the psychiatrist. The FTF comprises 181 statements that can be endorsed as “does not apply” (= 0), “applies sometimes or to some extent” (= 1), or “definitely applies” (= 2). The items are organized into 22 subdomains that compose eight general domains: motor skills, executive functioning, perception, memory, language, learning, social skills, and emotional/behavioral problems. The “motor skills” domain contains two subdomains: gross motor skills and fine motor skills. Individual items are shown in the Appendix. In this study the parents’ endorsement of gross motor problems in childhood is equated with “clumsiness”, and social skills problems are equated with “poor social skills.”

Peer victimization was assessed using a specific question: “Have you ever been bullied?” If the answer was “Yes,” then the patient was asked whether the peer victimization was regarded as “a little” (i.e., notable to the patient but not necessarily known all over school) or “a lot” (i.e., clearly notable to people around) and during which time span it occurred (i.e., in nursery school, primary school [7–9 years old, 10–12 years old, 13–15 years old], or upper secondary school [16–19 years old]).

The psychiatrist conducted a physical examination that included motor function prior to determining the diagnosis of the patient. Gross motor coordination was observed during hopping with alternating vertical jumps and swinging the arms forward on the contralateral side. This was scored as “normal” (= 0; i.e., no difficulty accomplishing the task), “slightly abnormal” (= 1; i.e., difficulty coordinating the arms and legs), or “clearly abnormal” (= 2; i.e., great difficulty and only able to do at most three dis-coordinated hops). A score of 1 or 2 on this test was defined as “abnormal.”

A number of assessments were used for assisting in the diagnostic procedures for ASD and ADHD. The Autism Spectrum Screening Questionnaire (ASSQ), a 27-items screening tool for detecting high functioning children with ASD, was completed by a parent [35]. The Asperger Syndrome Diagnostic Interview (ASDI) [36], a highly structured diagnostic interview, comprising 20 different items including assessment of social impairment, narrow interests, repetitive routines, speech and language peculiarities, non-verbal communication problems, and motor clumsiness was used for almost all patients (n = 258). For a substantial group (n = 229) the self-rated Autism-Spectrum Quotient (AQ) [37] was also administered. All patients with a possible ADHD were assessed with a structured interview for ADHD, the Wender-Reimherr Adult Attention Deficit Disorder Scale (WRAADDS), which covers attention difficulties, hyperactivity/restlessness, temper, affective liability, emotional over-reactivity, disorganization, and impulsivity in adulthood [38]. The Wender Utah Rating Scale (WURS) was used in order to capture childhood symptoms of ADHD [39].

In addition, general functioning was assessed using the Global Assessment of Functioning (GAF) ranging from 0 to 100 [21], and depression was measured with the Montgomery-Åsberg Depression Rating Scale [40]. Wechsler Adult Intelligence Scale, neuropsychological version, (WAIS III and WAIS III-NI) was used in order to identify intellectual disability. In those few cases where WAIS was not administrated, a high educational level excluded mental retardation. Personality traits were assessed using the screening version of Structured clinical interview for DSM-IV (SCID II Screen) and with the Swedish universities Scales of Personality [41].

The diagnosis of ADHD or ASD was made after a consensus was reached between psychiatrist and psychologist according to the criteria of the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM–IV criteria). Ninety-three patients (male, 49; female, 44) fulfilled the criteria for ASD, and 128 (male, 61; female, 67) fulfilled the criteria for ADHD. Fifty-six (male, 23; female, 33) had other severe psychiatric problems (e.g. chronic depression or severe anxiety disorders) (Table 1).

Data from this cohort on the outcome of central stimulant treatment in ADHD patients [42] and on psychiatric comorbidity in adults with ASD [43] have previously been reported. Approved written consent to use the information in research was obtained from the patients, and ethical approval was obtained from the Ethics Review Board, Karolinska hospital in Stockholm.

The peer victimized group consisted of all participants who reported having been bullied in nursery school or school, regardless of severity and duration of the peer victimization. Dichotomous variables were analyzed using the Pearson χ ² test. Spearman’s rank correlation test was used to assess the correlations between motor skills and various domains on the FTF, with a significance level p < 0.001.

A bivariate logistic regression analysis, with peer victimization as the dependent variable, was used to screen the subdomains of the FTF (Table 2). The analyses that showed significant relationships with peer victimization were included in a multivariate logistic regression analysis, in which risk factors indicated from previous research (i.e., sex, diagnosis, and psychiatric care in childhood) were also included. For all of these analyses, the mean FTF subdomain scores (ranging from 0 to 2) were used. In the logistic regressions, odds ratios were calculated as unit change (i.e., the changed likelihood that resulted from an increase of subdomain score by 1). In a third step, the number of time spans of peer victimization, gross motor skills problems according to the FTF, sex, ASD, ADHD, and other diagnosis, and sex × diagnoses interactions were analyzed in an ordinal regression model (polytomous universal model). The significance level was set to p < 0.05. Statistica version 10 software (StatSoft) was used for all of the analyses, with the exception that SPSS version 19 software was used for the ordinal regression (PLUM).

Of the total sample, 64% (68% of the males and 60% of the females) reported that peers had victimized them in childhood. Most afflicted were individuals diagnosed with ASD, 83% of whom had been victimized. This rate was markedly higher than the rate among those with ADHD (53%) or those with other psychiatric diagnoses (57%). Of those who were victimized at any time period, 69% were also victimized during at least one additional period, confirming the repetitiveness of the experience.

In our sample, education, civil status, employment status, GAF, Body Mass Index in adulthood, Intelligence Quotient, current depressive symptoms, antidepressant treatment, and history of childhood psychiatric care did not differ between the victims and non-victims groups. However, those with a history of being victimized were lonelier compared with the non-victimized group (Table 1).

In 235 participants, information on childhood symptoms was retrospectively obtained from a parent. Sixty-eight patients were reported to not have had any gross motor skill problems in childhood, whereas in 167 patients clumsiness was reported. Seventy per cent of the individuals with reportedly poor gross motor skills had been victimized by peers compared with 47% of those with normal motor skills in childhood, corresponding to an unadjusted odds ratio of 2.63 (95% confidence interval: 1.47-4.70). Girls who were reported to have poor motor skills had the highest risk for being victimized by peers at 10–12 years of age compared with a peak risk at 13–15 years of age for boys with poor motor skills (data not shown). Peer victimization at the different age spans is presented in Figure 2 for those with and without reported gross motor skills problems in childhood.

Impaired visual perception was similarly associated with an increased risk for peer victimization (unadjusted odds ratio 2.00, 95% confidence interval: 1.10-3.63); however, none of the other domains of the FTF was associated with an increased odds ratio for peer victimization in the bivariate regression. In this particular cohort, externalizing and attention problems reported on the FTF, suggesting ADHD, seemingly decreased the risk of being victimized. In Table 2, the odds ratios for peer victimization associated with the various FTF domains (as continuous variables), sex, previous child psychiatric treatment, diagnosis, and educational level are presented, showing an almost three-fold increased risk of being victimized by peers among those with poor gross motor skills in childhood.

In an ordinal logistic regression model, with the number of time spans of peer victimization as the dependent variable, poor gross motor skills in childhood remained an independent risk factor when sex, diagnosis (i.e., ADHD, ASD, and other), and sex × diagnosis interactions (which turned out non-significant) were introduced in the model. The association between poor gross motor skills in childhood and number of peer victimization periods had an odds ratio of 1.78 (95% confidence interval: 1.13-2.79; n = 227, p = 0.013).

Strong correlations were found between gross motor skills in childhood and body perception (ρ (n = 232) = 0.57, p < 0.0001), fine motor skills (ρ (n = 235) = 0.55, p < 0.0001), and social skills in childhood (ρ (n = 233) = 0.51, p < 0.0001). Visual perception correlated strongly with fine motor skills (ρ (n = 230) = 0.55, p < 0.0001) and body perception (ρ (n = 228) = 0.53, p < 0.0001). Weaker correlations, but highly significant, were found between vertical jumps in adulthood and gross motor skills in childhood (ρ (n =168) = 0.26, p = 0.0008), body perception (ρ (n =166) = 0.29, p = 0.0002), and social skills in childhood (ρ (n =168) = 0.25, p = 0.0009).

Of those with impaired gross motor function (i.e., poor vertical jump performance), 75% had been victimized by peers in childhood compared with 59% of those with normal performance (χ ² (df =1) = 6.01, p = 0.01, n = 206, unadjusted odds ratio 2.22 [95% confidence interval: 1.15-4.28]).

A larger proportion of patients with ASD performed poorly on vertical jumps compared with those with ADHD or other diagnoses (χ ² (df =2) = 8.11, p < 0.05, n = 205), showing impairments in 44.0% of the ASD subjects, 29.0% of the ADHD subjects, and 18% of the subjects with other diagnoses. However, the relationship between peer victimization and poor motor skills was not dependent on the ASD diagnosis (Figure 3).

Motor deficits in children with ADHD and ASD are well documented [30, 45, 46], and comorbidity between ASD, ADHD, and developmental coordination disorder has long been recognized [13]. Children with a combination of motor coordination deficits and ADHD have less favorable outcomes and more autistic traits [24], including perceptual problems, supporting the concepts of Deficits in Attention, Motor control, and Perception (“DAMP”) [25, 30] and Early Symptomatic Syndromes Eliciting Neurodevelopmental Clinical Examinations (“ESSENCE”) [47]. This also supports the idea that atypical brain development can explain the interrelatedness between these disorders [48].

Although motor abnormalities are common in ADHD and ASD and pivotal for the diagnosis of developmental coordination disorder, the underlying neurophysiologic impairments and type and persistence of motor problems appear to vary between the diagnostic groups and subgroups. Research in adults with ASD suggests a combination of cerebellar and basal ganglia deficits [49], and imaging studies of ASD and ADHD have consistently found cerebellar abnormalities [50‐52]. The cerebellum controls sensorimotor coordination and is critical to timing computations in both motor and non-motor tasks [53]. Working memory, the shifting of attention, implicit learning, emotional regulation, executive function and facial recognition all depend on cerebellar functioning [50, 54]. Moreover, symptoms of cerebellar lesions resemble the typical impairments of ASD [55].

In this study, the parents reported poor social skills in the majority of the cases (victims and non-victims), consistent with the social deficits related to ASD and to some extent to ADHD. However, poor gross motor skills in childhood but not poor social skills predicted peer victimization in the present study. Because poor social skills are frequent in both ADHD and ASD [56, 57], this by itself may explain the increased risk for peer victimization compared with normal controls [11, 12, 58]. Nevertheless, we propose that poor gross motor functioning constitutes an independent risk factor for peer victimization across various populations, regardless of provocative behaviors or anxiousness. Poor motor skills may manifest as impaired social skills, thereby contributing to peer victimization. Success in social interaction is largely a matter of timing and harmonious integration of verbal and non-verbal cues. In other words, cerebellar function may be critical for social success. Impairments in social functioning could be viewed as disintegrated motor skills, reflected by subtle aberrations in facial expressions, gaze, pitch, prosody, wording, posture, gestures, turn-taking, and sense of appropriate physical distance. Even slight signs of impairment, registered at a subliminal level, may affect others’ judgments and determine whether the person will be viewed as socially “normal” and attractive as a companion or considered “strange” and result in social exclusion or peer victimization. Preadolescents and adolescents are exceptionally sensitive to deviations [59], are easily embarrassed [60], and strive to conform to their peers [59, 61], possibly explaining why peer victimization peaks in the preteen years [62]. We suggest that motor skills are associated with social likeability and that poor skills may contribute to the loneliness in adulthood reported by the victims in this study. Thus, the feeling of loneliness often reported by bullied children [1] seems to continue into adulthood.

Another possible explanation for our findings could be the close association between poor motor skills and visual perceptual dysfunction. Visual perceptual dysfunction has been shown to be associated with the number of co-occurring disorders in children with developmental coordination disorder [63] and contribute to poor outcome in preterm born babies [64] and in children with hyperopia [65]. Poor motor skills without visual perception dysfunction may represent a dimension of normality whereas the combination of the two reflects a neurological dysfunction with increased risk for social exclusion, and/or cognitive, behavioral and emotional disability. In a study on adults with either social phobia or obsessive-compulsive disorder (OCD), childhood peer victimization was much more often reported in the OCD group [66]. Although both disorders have similar ages of onset, chronicity, and relationships with avoidant personality, OCD differs from social phobia by being associated with motor problems and soft neurological signs [67‐69], including deficits in visuospatial skills [70]. Interestingly, poor visuospatial skills predicted persistence in pediatric-onset OCD [71].

None of the established factors previously shown to be associated with peer victimization (e.g., education level, civil status, employment, general functioning, overweight, intelligence, reported social skills, internalization, depression, and antidepressant treatment) were implicated in this study. Notably, however, childhood Body Mass Index data were not obtained. By studying a sample at high risk for peer victimization, risk factors beyond the expected may be revealed, such as poor gross motor skills. Hence, our lack of findings about the established risk factors cannot be generalized to normally developing children.

Psychiatric patients are often targets of peer victimization in childhood. Bullying amongst normally developing children peaks in the pre teens, but children with both poor gross motor skills and psychiatric problems have often been bullied for a prolonged period of time. Prolonged peer rejection tends to result in poor self-esteem and may decrease the ability to appeal to others. In adult psychiatric patients a history of poor gross motor skills and peer victimization in childhood may indicate a severe and pervasive neuropsychiatric disorder with social skills deficits. In such cases the clinician should be aware that a range of support and treatments often are indicated.

Simple assessments of motor skills in pre school children should be helpful for identifying children at risk. Furthermore, the possibility should be explored whether early intervention programs using specialized physical education adapted for ‘clumsy’ children (e.g. [72]) reduce the risk of loneliness and peer rejection.

The patient group studied here may be perceived as marginal; on the other hand ADHD is not by any means a rare disorder in adults. ADHD has a chronic course in approximately 4% of the adult population [73] and can be diagnosed in almost one of four adult psychiatric outpatients [74, 75]. Concerning ASD, the prevalence rate in the general population exceeds 1% [76], and ASD frequently co-occurs with other psychiatric disorders. In a recent study, 70% of adults with ASD had experienced at least one episode of major depression [77]. Possibly, the co-occurring disorders draw the clinician’s attention, whereas ADHD and ASD often remain un-diagnosed.

Nevertheless, this study has several limitations, partially due to lack of relevant instruments and the naturalistic setting. First, our retrospective use of FTF could interfere with the results, as the FTF was designed to be filled out by parents, when evaluating their children. Consequently, since parents may forget earlier signs of problematic behavior, false-negative responses may constitute a problem. However, false positive results seem less likely. In our experience, the parents recalled signs of atypical development remarkably well and our results are consistent with a previous report on children with ASD and ADHD that used the FTF [18]. Secondly, we did not use any specific definition of peer victimization, but the concept of bullying is well known [78], and the patients never questioned the meaning of the wording. They knew whether or not they were victimized by peers and remembered how long it lasted and who where the perpetrators. Being victimized by peers is such a traumatic experience that it may reside as a humiliating memory throughout one’s life. Although some children may deny being victimized by peers or belittle it as “teasing”, we question whether this bias prevails into adulthood. In fact, self-reports of peer victimization by adults may be more reliable than children’s self-reports because the shame that hindered them in childhood may eventually subside. In this study we did not inquire about type of bullying that the participants had experienced, which is another limitation. Most of the participants were in their preteens during the eighties and nineties, thus cyber bullying had not emerged on the scene yet. The three-point scale to determine severity used in this study is not a validated method for measuring severity. However, the question is straightforward and the responses were dichotomized in the analyses; thus the peer victim group consisted of all participants who reported being bullied, regardless of severity. Attrition was partially attributed to the fact that many patients were originally investigated at other clinics and only referred to us for medication, thus providing a smaller proportion of patients to examine. Of the total sample, 21% were not included in the study because of missing data. However, no significant demographic differences were found between these individuals and those included in the study.