Abstract
School refusal is a problem affecting approximately 1–2%% of all school-aged children and about 5%% of all clinic-referred children and adolescents [1–3]. Persons with school refusal differ from truants in that, while physically able to attend school, the former stay at home with the knowledge of their parents and do not show serious antisocial behaviour. The young person refusing to attend school often has some emotional distress relating to school attendance and the parents usually make efforts to encourage the young person to go to school [4–6].
Over the past 60 years, a shift has occurred in the understanding of school refusal [2]. In the 1930s, young people with school refusal were described as anxious [7]. The term ‘school phobia’ was coined in the following decade [8]. By the early 1960s child attachment problems were seen as the primary cause of school refusal and ‘separation anxiety’ was regarded as the main diagnostic category accounting for the problem [9]. Both ‘school phobia’ and ‘separation anxiety’ have persisted as a way of characterizing young people with school attendance difficulties. Recent studies, however, have noted a high prevalence of mood disorders in adolescents with school refusal [10–18]. Today, school refusal is viewed as a symptom associated mainly with anxiety disorders in children and with anxiety and depressive disorders in adolescents [2].
Studies on school refusal have often been limited by small sample, an emphasis on outpatient groups, lack of comprehensive diagnostic information, and negligible data about precipitants [9], [10],[14–19]. These data are crucial to a better appreciation of some of the factors associated with the problem and in planning effective management.
We aimed to describe a large sample of clinic-referred adolescents with school attendance problems. The information thus gained may be useful for a greater understanding of this symptom, its prevention and treatment. It was hypothesized that: (i) young people with school refusal would mainly have anxiety and depressive disorders; (ii) those with school refusal admitted as inpatients would have a more severe condition, for example would show greater functional impairment on Global Assessment of Functioning (GAF) and greater comorbidity, compared with those not admitted; (iii) family or peer conflict and academic difficulties would be major stressors associated with the onset of the problem.
Method
Subjects
The subjects were 192 young persons (10–17 years), assessed or treated for school refusal between 1994 and 1998 at the Rivendell Unit. Rivendell is a university-affiliated, child and adolescent psychiatric facility in Sydney, Australia, that provides inpatient and outpatient services. Treatment offered by the unit for young people with school refusal problems includes graded exposure, individual cognitive behaviour therapy, family therapy and, in some cases, psychotropic medication. Inpatients receive more intensive treatment and more focused educational input.
Subject selection
Patients with school refusal problems were identified by reviewing all the medical records of the unit between January 1996 and December 1998. Individuals were included in the study if (i) the parent gave a positive response (a rating of 1 or 2) to the Child Behaviour Checklist (CBCL) item ‘fears school’ [20] at initial assessment, and (ii) school refusal was corroborated by the parents (who identified school refusal as the presenting problem in the unit's initial questionnaire) or by the referring agent on the referral form. On initial assessment, subjects diagnosed with psychosis, conduct disorder or obsessive compulsive disorder were excluded. However, examination of further data obtained over the time of the study resulted in some patients being diagnosed with conduct disorder.
Medical record review
Following ethics committee approval, the medical records of all patients assessed at the unit between January 1994 and December 1998 were inspected [by GM]. The following data were obtained using a checklist developed by the authors: patient's gender, status (inpatient, outpatient or assessment only), age of onset of school refusal, age at assessment, duration of school refusal, associated events (rated present or not present), number of schools attended, use of psychotropic medication, family psychiatric history and family composition, GAF at assessment [21], CBCL [20] and Centre for Epidemiological Studies Depression Scale for Children (CES-DC) scores at assessment [22], parental academic achievement and length of admission (if admitted).
Diagnoses
The initial diagnosis (using DSM IV criteria) recorded in the medical record at assessment was accepted as the diagnosis, if the discharge summary confirmed the initial diagnosis, or if there was insufficient information to challenge this diagnosis. Consensus diagnoses were made when there was a change in diagnoses or additional diagnoses from the initial assessment till termination of treatment. The consensus diagnoses were made by two of the investigators according to DSM IV criteria [21] [GM, GW], after considering all sources of information.
Statistical analysis
Because this is mainly a descriptive study, few statistical tests were conducted. Chi-square and anova were used when comparing groups. Only differences significant at p < 0.05 are reported. Percentages are rounded to the nearest unit.
Results
One hundred and ninety-two patients who presented to Rivendell between 1994 and 1998 met criteria for inclusion in the study. This accounted for 7%% of all patients assessed at the unit (n = 2561) during this period. Of those presenting with school refusal, 105 (55%%) were male. The mean age at assessment was 14.2 years (SD = 1.3, range = 10–17 years). Almost half (n = 93, 48%%) were subsequently admitted for inpatient treatment to the unit, about a third (n = 58, 31%%) were treated as outpatients and a fifth (n = 41, 21%%) received no further treatment at the unit following assessment (this last group either had a single consultation or were referred elsewhere for further treatment). The average length of treatment for those treated was 13 weeks.
Refusal to attend school generally began in the first or second year of high school (n = 149, 78%%). The mean age of onset of these difficulties was 12.3 years (SD = 2.6).
In 80%% of the cases (n = 153) school refusal problems had been present for 2 years or less prior to their assessment at Rivendell. Patients had attended on average four schools. Patients in this study had a number of circumstances associated with the onset of their school refusal. These circumstances (recorded in the medical record) included conflict at home (n = 83, 43%%), conflict with peers (n = 67, 34%%), academic difficulties (n = 59, 31%%), family separation (n = 40, 21%%), changing school or moving home (n = 48, 25%%), and physical illness (n = 39, 20%%). Of those who had conflict with peers, bullying was identified in 27 (14%%) cases.
A little more than half of the patients (n = 103, 54%%) lived in an intact (dual parent) family, while one-third lived with a single parent (n = 74, 39%%). Physical illness was identified in 70 (37%%) of the patients, 35 (18%%) of the patients’ mothers and 27 (14%%) of the patients’ fathers. Maternal psychiatric illness was reported in about half the patients (n = 102, 53%%); approximately one-third (n = 66, 34%%) had a paternal history of psychiatric illness. Similar proportions of the patients’ fathers (n = 77, 40%%) and mothers (n = 79, 41%%) completed the equivalent of 6 years of secondary schooling.
Mean GAF and CES-DC on assessment were 54 (out of 100) and 25 (out of 60), respectively, suggesting that these adolescents were significantly impaired and had a moderately depressed mood at the time. The CBCL scores showed reduced total competency (mean = 10.9, SD = 3.3) for activity (mean = 4.3, SD = 2) social (mean = 4, SD = 2) and school (mean = 2.6, SD = 0.9) domains. All ratings on the other CBCL scales were within the clinical range (withdrawn [mean = 7.6, SD 3.5], somatic [mean = 5.2, SD 3.6], anxious/depressed [mean = 13.3, SD 6.4], social problems [mean = 5.3, SD = 3.4], thought problems [mean = 2.8, SD = 2.7], attention problems [mean = 8, SD = 4.6], aggression [mean = 13, SD = 8.3], and delinquency scales [mean = 5.6, SD = 5]).
There were no significant differences between those admitted as inpatients and the rest in age of onset of school attendance problems, age at presentation, gender, family composition, family conflict, family separation, history of abuse, number of schools attended and length of treatment. There were also no differences that emerged between inpatients and the rest on CES-DC, GAF ratings and CBCL scores in any of the scales.
Clinical diagnoses of anxiety, mood and disruptive behaviour disorders were very frequent in the sample (Table 1). Inpatients had a higher rate of dysthymia (chi square [1, 192] = 9.14 p < 0.01), while those not admitted as inpatients had a higher rate of panic disorder (p < 0.01) and disruptive behaviour disorder not otherwise specified (NOS) (p < 0.05). Over one half (n = 106, 55%%) had more than one concurrent diagnosis. Dysthymia often coexisted with major depression (chi-square [1, 192] = 4.37, p < 0.05), separation anxiety (p < 0.01) and oppositional defiant disorder (p < 0.05). Inpatients had significantly more comorbid diagnoses than the rest (chi-squared [4, 192] = 20.25, p < 0.001). More than a third of patients (n = 70, 37%%) were prescribed psychotropic medication at some stage of their treatment. Those admitted as inpatients were more likely to be prescribed antidepressant medication than the others (chi-square [1, 191] = 25.82, p < 0.001). A maternal history of psychiatric illness was more prevalent in the inpatient sample (chi-square [1, 192] = 17.56, p < 0.01).
Diagnostic profile of patients presenting with school attendance difficulties
Discussion
To our knowledge, this study reports the largest sample of adolescents with school refusal in the literature. Young people with school refusal in this group mainly had anxiety and depressive disorders. Family or peer conflict and academic difficulties were the major stressors associated with the onset of the problem. Inpatients, however, seemed to be neither more impaired nor to have more severe symptoms than those who were not admitted (based on GAF, CBCL and CES-DC scores). Overall the GAF, CBCL and CES-DC scores for this sample indicates significant impairment and psychopathology. This calls into question the reasons why some were admitted and some were not. Many factors may have contributed to that decision (e.g. some parents were less able to cope, some adolescents may have been unwilling suddenly, etc.). What is clear is that, on average, those admitted were not more disturbed than those not admitted. Young people with mood disorder (dysthymia) or with comorbid diagnoses were more likely to be admitted for inpatient treatment. The study has several limitations. First, the design was retrospective. Second, there was no control group. As a result, it cannot be concluded definitely whether these events are more common among school refusers than the rest. Third, the data were obtained from a single site – a specialist adolescent unit; findings may not generalize to other settings, such as the community. Fourth, as a result of the way information was arranged in case files, ratings could not be made blind to patient status (inpatient vs other).
Many of the findings are consistent with results of earlier studies. For example, anxiety and depressive disorders are generally found in these patients, comorbidity is frequent, and duration of symptoms prior to presentation (mostly 2 years or under) is in keeping with previous research [9–17].
There were some discrepant findings. For example, Borchardt et al. [16] found that inpatients, compared with outpatients, experienced greater family dysfunction and abuse and were more likely to come from single parent families; we found that inpatients and outpatients experienced family dysfunction at the same rate and inpatients were not more likely to be from single parent families. Our study also revealed a high prevalence of comorbid disruptive behaviour disorders (apart from conduct disorder), which has not been highlighted previously. Disruptive behaviour disorders are likely to present an added challenge to treatment.
The finding that ‘school refusal’ is a symptom of a variety of disorders reinforces the view that a diagnosis of separation anxiety should not be assumed in these cases and that treatment should be tailored to the individual patient. Since the early high school years represent the common time of onset, identification of vulnerable individuals and implementation of transitional programs to assist with the adjustment from primary to high school are warranted. Some of the factors associated with school refusal may be modifiable.
The consequences of school refusal include short- and long-term deficits in personal, social, academic and vocational development. The efficacy of treatment for young people with school refusal is the focus of a companion paper on patient outcomes emanating from this study (McShane et al., unpubl. data).