Intended for healthcare professionals


Association of common health symptoms with bullying in primary school children

BMJ 1996; 313 doi: (Published 06 July 1996) Cite this as: BMJ 1996;313:17
  1. Katrina Williams, lecturera,
  2. Mike Chambers, lecturera,
  3. Stuart Logan, senior lecturerb,
  4. Derek Robinson, consultant community paediatricianc
  1. aDepartment of Epidemiology and Medical Statistics, St Bartholomew's and Royal London School of Medicine and Dentistry, Queen Mary and Westfield College, London E1 4NS
  2. bDepartment of Epidemiology, Institute of Child Health, London WC1N 1EH
  3. cLondon WC1N 1EH, c Newham Health Care, Beacontree House, London E15 4EE
  1. Correspondence to: Dr Logan
  • Accepted 12 April 1996


Objectives: To estimate the prevalence of bullying in primary school children and to examine its association with common symptoms in childhood.

Design: Semistructured health interview conducted by school nurses as part of a school medical.

Setting: Newham, east London.

Subjects: All children in year 4 of school during the academic year 1992-93.

Main outcome measures: Reported bullying and common health symptoms.

Results: 2962 children (93.1% of those on the school roll) were interviewed (ages 7.6 to 10.0 years). Information about bullying was not recorded for 114 children. 22.4% (95% confidence interval 20.9 to 24.0) of children for whom information was available reported that they had been bullied. There was an association between children reporting being bullied sometimes or more often and reporting not sleeping well (odds ratio 3.6, 2.5 to 5.2), bed wetting (1.7, 1.3 to 2.4), feeling sad (3.6, 1.9 to 6.8), and experiencing more than occasional headaches (2.4, 1.8 to 3.4) and tummy aches (2.4, 1.8 to 3.3). A significant trend for increasing risk of symptoms with increased frequency of bullying was shown for all reported health symptoms (P<0.001).

Conclusions: Health professionals seeing primary schoolchildren who present with headaches, tummy ache, feeling sad or very sad, bed wetting, and sleeping difficulties should consider bullying as a possible contributory factor.

Key messages

  • Many children report having frequent headaches or tummy aches, that they sleep poorly, wet the bed, or feel sad

  • Children who report these symptoms also report being bullied substantially more often than do their peers

  • Although it is not clear whether the association is causal, health professionals seeing such children should ask about bullying


A large proportion of children report being bullied at school.1 2 3 4 In a study in Sheffield of 2623 children aged 8-11, based on self-completion questionnaires, 27% reported being bullied, 10% once a week or more.3 The reported forms of bullying, in decreasing frequency, were being called names, being physically hurt, being threatened, having rumours spread about you, not being talked to, and having belongings taken.

It is generally accepted that bullying causes substantial distress and that both short and long term adverse effects may occur. Interventions to prevent bullying have been developed,1 5 but they have not been fully evaluated.6

A semistructured health interview conducted by school nurses with children in year 4 in Newham, east London, provided an opportunity to examine the relation between reports of being bullied and common childhood symptoms.

Subjects and methods

Since 1988 all children in year 4 (age 8-9) attending mainstream schools in Newham have been invited to attend a health interview with the school nurse. This is intended to be both a therapeutic exercise and a way of collecting data. The child is seen either alone or in the presence of a parent or carer. Data reported here refer to the 1992-93 academic year.

The semistructured interviews were conducted by 18 school nurses, who completed the questionnaire. The questions were asked in sequence (appendix), but the approach to the interview was not rigid. If children responded in a way that suggested a health problem this was further explored at the time. If the nurse decided that a reported problem was in fact “normal” then it was not recorded as a problem. Most questions had a range of responses (appendix), and each child was told of the response options.

Although no standard definition for bullying was used, the nurse asked if the child knew what bullying was. The nurse then offered an explanation of bullying, asked if the child understood this explanation, and then asked if he or she was bullied.


In the comparative analyses several options have been condensed. For the question “how do you usually feel?” answers of happy and very happy were condensed into a “happy” category, answers of sad and very sad into a “sad” category. Reports of being bullied “sometimes” or more often were classified as being bullied. Headache and tummy ache were reported a problem only if experienced once a week or more often. The prevalence and 95% confidence interval (exact binomial) of reported bullying and other health symptoms were calculated with SPSS/PC. Odds ratios with 95% confidence interval (Cornfield)7 and χ2 analyses of linear trend were calculated with the packages Confidence Interval Analysis8 and Epi Info.9.Denominators in the tables vary slightly because of missing data.


Questionnaires were completed for 2962 (93.1%) of the 3180 eligible children. Of 218 not interviewed, 42 attended a school without a school nurse and 176 failed to attend.

Age at interview was known for 2953 children. The mean age was 9.0, range 7.6-10.0 (98.7% were between the expected ages at interview of 8 years, 3 months and 9 years, 9 months). A parent was present for 1644 (55.5%) of interviews.


Relevant questions from questionnaire in order of asking, with possible responses

View this table:

Information on bullying was not available for 114 children. Forty children (1.4%) said they were bullied nearly every day and 638 (22.4%) bullied at some time (table 1). A higher proportion of boys reported bullying (347/1428 (24.3%), 95% confidence interval 22.1% to 35.0%) than girls (290/1419 (20.9%), 23.1% to 28.3%). Most bullying occurred at school, although 22 children reported bullying in the home. table 2 gives the frequency of different types of bullying, and table 3 shows the number of children reporting common symptoms.

Table 1

Reported frequency of bullying

View this table:
Table 2

Frequency of different types of bullying

View this table:
Table 3

Frequency of health symptoms reported

View this table:


Reporting being bullied was associated with an increased risk of reporting not sleeping well (odds ratio 3.6, 95% confidence interval 2.5 to 5.2), bed wetting (1.7, 1.3 to 2.4), feeling sad (3.6, 1.9 to 6.8), and experiencing more than occasional headaches (2.4, 1.8 to 3.4) and tummy aches (2.4, 1.8 to 3.3) (table 4).

Table 4

Association between reported bullying and other health symptoms

View this table:

Table 5 shows children reporting not sleeping well, headache, and tummy ache by frequency of bullying. Linear analysis of trend for different frequencies of reported bullying was significant for all symptoms: not sleeping well (χ2 = 61.8; P<0.0001), bed wetting (χ2 = 13.5; P = 0.0002), feeling sad (χ2 = 26.2; P<0.0001), headaches (χ2 = 38.0; P<0.0001), and tummy aches (χ2 = 38.6; P<0.0001).

Table 5

Percentage (number) of children with reported tummy ache, headache, and not sleeping well according to frequency of bullying

View this table:

Analyses stratified for parental presence at interview, sex, ethnicity, and reporting of multiple symptoms made little difference to the estimated odds ratios.


Our results suggest a strong association between reported bullying and common symptoms. Most of the problems with the design of this study, including random misclassification due to lack of standardisation of practice, are likely to weaken these associations. If, however, the likelihood of recording reported bullying was increased by the presence of other symptoms, the relative risk would be biased upwards. Although this potential bias cannot be excluded, any symptom was discussed as it was raised during the interview. This meant that the interviewer felt that problems had been sufficiently explored and explained before proceeding.

It has been suggested that the prevalence of bullying is best assessed through self completion questionnaires.10 The prevalence reported here is, as expected for an interview, slightly lower than reported for this age group in other studies using self completed questionnaires in Britain and Ireland,2 3 4 although the difference between the sexes is similar.

The linear association found between risk of health symptoms and increasing frequency of bullying suggests a “dose effect,” strengthening the argument that this is a causal relation. We cannot, however, exclude the possibility that the risk of reporting being bullied and of reporting other symptoms are both manifestations of some other underlying problem.

Health professionals seeing children with these common symptoms should ask whether the child is being bullied. If bullying is reported, whether or not it is causally related to the presenting complaint, it should be taken seriously.

We thank the founding members and steering group of the Newham 8-plus health survey, Margaret Baker and the school nurses for their enthusiasm, and Rumana Omar and Sandra Eldridge for statistical advice.


  • Funding Newham Health Authority (now East London and the City Health Authority) and Newham Community Health Services (now NHS Trust).

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
View Abstract