Clinical predictors of severe behavioural problems in people with intellectual disabilities referred to a specialist mental health service
- 108 Downloads
Associations between demographic and clinical variables and severe behavioural problems in people with intellectual disabilities were examined in a cross-sectional survey of 408 adults consecutively referred to a specialist mental health service. Severe behavioural problems were present in 136 (33.3%) of the sample. The demographic and clinical predictors of severe behavioural problems in this sample were identified by logistic regression. Age and gender were not associated with severe behavioural problems. The presence of severe ID independently predicted the presence of severe behavioural problems. Schizophrenia spectrum disorders and personality disorders independently predicted the presence of severe behavioural problems, whereas the presence of an anxiety disorder independently predicted their absence. There is an increasing evidence base of relationships between mental disorders and behavioural problems in people with ID although the pattern of these relationships remains unclear.
KeywordsIntellectual disabilities Behavioural problems Mental disorders
Behavioural problems are a major reason for referrals of people with intellectual disabilities (ID) to specialist mental health services . Multiple factors are thought to be associated with behavioural problems  including demographic and clinical diagnostic variables.
It has often been stated that behavioural problems in ID peak at late adolescence and are less prevalent in later life . For example, it has been reported that people who exhibit aggressive behaviours are typically in younger adulthood [3, 34]. A relationship between self-injurious behaviours (SIB) and younger age has also been reported [6, 28, 31] but not always confirmed .
Several studies have reported higher rates of aggression in males [4, 11] although one found no gender effect . In contrast, one study linked female gender to behavioural problems . Some researchers have reported SIB to be increased in females  but others have found SIB to be unaffected by gender . The results are therefore widely varying and often contradictory. To try to make sense of the overall findings a meta-analysis was carried out to identify risk markers for behavioural problems . It was concluded that males were more likely to show aggression, but not more likely to show SIB, in the only two studies that met their inclusion criteria.
Level of ID
Many studies have reported that behavioural problems may be more common in those with more severe ID [10, 13, 19]. There may be differences between specific behaviours in this regard. Most studies have found a relationship between more severe ID and SIB [6, 16, 28]. One study found that those with severe or profound ID were more likely to show SIB but no more likely to show aggression than those with mild or moderate ID . Another study also did not find any increase in aggression with lower level of functioning .
Despite ongoing methodological problems there is a growing body of evidence that behavioural problems are associated with an increased prevalence of mental disorders in people with ID. For example, one study found that behavioural problems increased the likelihood of almost all mental disorders as much as three-fold . Another group found that those with a mental disorder generally scored higher on scales of maladaptive behaviours . In the non-ID population there is evidence that some mental disorders such as personality disorders and psychotic illnesses are associated with an increased risk of aggressive and violent behaviours . There is much less similar evidence available in the ID population. It has been reported in one study that schizophrenia, affective, personality and anxiety disorders are all associated with higher levels of behavioural problems . However, as with demographic variables there has been much variation in the patterns of relationships found with clinical diagnoses. A range of associations of mental disorders with behavioural problems has been reported [17, 24, 33] although there has been much inconsistency in definitions, terminology and diagnostic criteria used .
It remains uncertain whether there are associations between epilepsy and behavioural problems in people with ID. Some have found no significant associations [14, 36] whilst others have found that epilepsy was negatively correlated with behavioural problems .
Summary of relationships
The pattern of associations of behavioural problems with demographic factors and clinical diagnoses thus remains unclear. Previous studies have not been consistent on which variables most strongly predict the presence of behavioural problems. It would be of interest to services to know which demographic factors and clinical diagnoses may be associated with behavioural problems in the service users they are referred. This information could greatly assist with the assessment and management of individuals but also with service planning. Yet it remains the case that there still has been too few studies in this area. Furthermore, existing studies have often not used large sample sizes or standardised diagnostic criteria or rating instruments.
Aims of the study
The present study aimed to explore relationships between several demographic factors and clinical diagnoses and severe behavioural problems, in a very large sample of people with ID referred to a specialist mental health service, using diagnoses made clinically with standard criteria and with the presence of severe behavioural problems rated using a widely recognised and validated instrument.
A retrospective study design was used. Data were collected between 1983 and 2001 at the initial clinical assessment of each participant as part of the clinical service. Approval from the local research ethics committee was granted. Informed consent from or assent for the participants was obtained and data were anonymised.
The sample (n = 408) consisted of adults (older than 18 years) drawn from consecutive clinical referrals to a specialist mental health service for people with ID in South East London in the UK. Inclusion criteria: ICD-10  diagnosis of mental retardation (equivalent term to intellectual disabilities) and actual or suspected mental health problems. Exclusion criteria: diagnosis of pervasive developmental disorder, including autism.
The following demographic variables were recorded: age (collapsed into binary categories of under 35 years and 35 years and over), gender, degree of intellectual disabilities (collapsed into binary categories of mild/moderate ID and severe ID) and ethnicity (collapsed into binary categories of “white” and “non-white”). The level of ID was determined clinically by employing ICD-10 criteria for mental retardation. Historical details of early developmental problems and any previous IQ testing were obtained from past medical records.
A psychiatrist made diagnoses of mental disorders (if present) at the time of the clinical assessment as part of the clinical service delivery. Two psychiatrists agreed on the diagnosis using ICD-10 criteria. Diagnoses were grouped into the following major ICD-10 categories: dementias (F00-03), schizophrenia spectrum disorders (F20-27), depressive disorders ((F32-39), anxiety disorders (F40-48), adjustment reactions (F43) and personality disorders (F60-69).
Past or current history of epilepsy at time of initial assessment was recorded.
Behavioural problems were rated using the Disability Assessment Schedule (DAS) on the DAS-B scale . These include physical aggression to others, destruction of property, over-activity, attention-seeking behaviours, pestering staff, self-injury, wandering or running away, screaming or making disturbing noises, temper tantrums, disturbing others at nights, difficult or objectionable personal habits, scattering or throwing things, antisocial behaviours, sexual behavioural problems and other behavioural problems. The DAS-B scale has been widely used and has good internal reliability . Each item was rated on a 0–2 scale (0 = “severe management problem”, 1 = “lesser management problem” and 2 = “does not occur”) with the maximum DAS-B total score being equal to 28. The presence of “severe behavioural problems” was defined by a DAS-B score of 0–5 in accordance with this validated and widely used rating scale .
The Statistical Package for Social Sciences (12th version; SPSS 12) was used. Bivariate analysis: chi-squared tests were performed to look for any significant associations between demographic and clinical variables of the participants with and without severe behavioural problems. Multivariate analysis: a binary logistic regression was then performed with a stepwise forward method, using those variables that had significantly predicted severe behavioural problems in the preceding bivariate analysis.
Participants were slightly more likely to be aged less than 35 years (53.1%) rather than more than 35 years old (46.1%) and male (59.1%) rather than female (40.9%). Participants were predominantly of white ethnicity (89.1%) compared to non-white ethnicity (10.9%) and with mild or moderate ID (87 %) rather than severe ID (13 %). In the studied areas of South London (boroughs of Lambeth, Lewisham and Southwark), the average population proportions by ethnic group in 2001 were 63.8% white and 36.2% non-white . Therefore this referred sample had an overall under-representation of those in non-white groups from the current total population but with increasing proportion from non-white ethnic groups over time .
Number and percentages of patients with and without severe behaviour problems (DAS-B) by age, gender and level of intellectual disabilities (ID)
DAS severe behaviour problems
DAS Severe Behaviour Problems
n (% of sample)
n (% of sample)
n (% of sample)
Level of ID
Number and percentages of patients with and without severe behaviour problems (DAS-B) by diagnosis of mental disorders and presence of epilepsy
DAS severe behaviour problems
DAS severe behaviour problems
n (% of sample)
n (% of sample)
N (% of sample)
Diagnosis (ICD-10 codes)
Schizophrenia spectrum disorder (F20-27)
Depressive disorder (F32-39)
Anxiety disorder (F40–48)
Adjustment reaction (F43)
Personality disorder (F60–69)
Other mental disorders
No diagnosable mental disorder
Regression model predicting presence of severe behavioural problems
In this study associations of severe behavioural problems with certain diagnoses were observed. Specifically, personality disorders and schizophrenia were associated with increased prevalence of severe behavioural problems whilst anxiety disorders were associated with their reduced prevalence. The factor in multivariate analysis most strongly predicting the presence of severe behavioural problems (more than six-fold) was the diagnosis of personality disorder. Another group has also reported that aggressive and rebellious behaviours rated by the Reiss Screen for Maladaptive Behaviours were more common in their study sample of people with ID when the participants had personality disorders clinically diagnosed using ICD-10 criteria . Although the diagnosis of personality disorder in ID is controversial , it could be that personality disorder is still under-diagnosed in people with ID and behavioural problems. Considering whether a person with severe behavioural problems has a personality disorder may be in some cases very useful in their management. This area needs further research.
The presence of schizophrenia spectrum disorders increased nearly three-fold the likelihood that a person would have severe behavioural problems. It has been reported that psychosis (and personality disorder traits) as recorded by scales (not clinical diagnoses) was related to aggression in people with severe ID . It has also been found that schizophrenia in people with ID is associated with maladaptive behaviours  The findings of the present study thus give substantial support to previous suggestions that people with ID and coexisting schizophrenia spectrum disorders show increased levels of behavioural problems.
An anxiety disorder diagnosis decreased the likelihood that a person would have severe behavioural problems. This may seem to contradict the findings of another study in which an increased prevalence of “anxiety” in people with behavioural problems was reported . However, this did not use standard diagnostic categories as with the present study. It may be necessary also to consider here the difference between anxiety disorders and symptoms. In people with ID there is little or no evidence whether specific anxiety disorders such as obsessive-compulsive disorder, agoraphobia and social phobia may be more or less associated with behavioural problems. However anxiety symptoms may often be part of other mental disorders that can be associated with an increased prevalence of behavioural problems.
Depression has often been the mental disorder most commonly suggested to be associated with behavioural problems in people with ID . For example, one study found that those with aggressive behaviours were more than four times likely to have depression . Associations have been found between severity of behavioural problems and prevalence of depression  using a screening instrument, the PAS-ADD Checklist . However, the diagnostic categories used also did not use standard criteria. The present study did not show any clear association with depression and severe behavioural problems. In fact, depression actually reduced the likelihood of severe behavioural problems in this sample in bivariate analysis although this effect did not persist in multivariate analysis. So the present study seems to contradict the findings from most other previously reported studies, which have tended to link affective disorders with increased behavioural problems. There is a theoretical possibility that depression can sometimes reduce the risk of severe behavioural problems through such symptoms as reduced motivation, social withdrawal and reduced psychomotor activity . It may also be that the stringent measure of severe behavioural problems (DAS score of 0–5) used in this study gave a different pattern of associations than would have been seen if relatively less severe or frequent behavioural problems were considered. It is of course also possible that the link between depression and behavioural problems may not be as strong as the results of some other studies may have suggested.
This study found no evidence of any associations between severe behavioural problems and dementia in this sample. However, it has been previously reported that dementia in ID is associated with a high prevalence of aggression . Although the total sample in the present study was large there were relatively few with a diagnosis of dementia. It would be worthwhile for future research to continue to look further at associations of dementia with behavioural problems as well as other specific mental disorders in people with ID. There was also no evidence found of any association of severe behavioural problems with epilepsy in this sample.
Many previous studies that have been drawn upon to describe the relationships between demographic variables and behavioural problems have had similar selected samples to the present study, and often they have been of much lower size. However it remains that the present study used a selected sample so its findings regarding the patterns of associations of demographic variables with severe behavioural problems must also be viewed with great caution. There has been a consensus that behavioural problems are more common in those younger, male and with severe ID . The present study did not find that age or gender was associated with severe behavioural problems in this referred sample. Severe ID was associated with severe behavioural problems, in keeping with the majority of previous studies. It is not possible though to conclude that severe ID is associated with behavioural problems, because the participants with severe ID may not have been representative of their community population. Whilst the finding of an association of severe ID with severe behavioural problems may still be a valid finding, at least some of this association may be accounted for by the presence of undiagnosed autism and/or mental illness in those with severe ID. Autism is increased in prevalence in more severe ID  and mental illnesses also become more difficult to diagnose with certainty. In a previous meta-analysis it has been reported that too little data exists currently to report on relationships between ethnicity and behavioural problems . In the present study the participants were of mostly white UK origin and it is therefore not possible to draw any conclusion about any possible association of ethnicity with severe behavioural problems in this sample.
With the use of so many different approaches and definitions previously it is unsurprising that there has been little consistency found in relationships between demographic and clinical variables with behavioural problems. There are many methodological problems in this area of research . One is that there is increased difficulty of making definitive diagnoses in people with ID. Another crucial issue is deciding on the thresholds used to define behavioural problems. There do not exist clear definitions of specific behavioural problems. For example, although widely used, an umbrella term such ‹self-injurious behaviours’ encompasses behaviours with sometimes little in common . Changing the definitions used of specific behaviours can produce quite different patterns of association and this has frequently happened in previous studies. This study benefited regarding this issue from the use of the recognised DAS-B category of ‹severe behavioural problems’ rather than the use of a broader measure such as the presence of any particular behavioural problem. Another strength of the present study was that it used clinical diagnoses using ICD-10 criteria rather than assessment by a screening instrument such as the PAS-ADD checklist. A further advantage of the present study was that it used a standardised rating scale for behavioural problems, the DAS-B .
This study does have a number of limitations. The most evident one is that the sample may be a highly selective one, as it comprised of referrals to a specialist mental health service for people with learning disabilities in London, UK. Another source of potential bias in this study was that it was retrospective. Another is that clear-cut diagnosis of mental disorders at initial assessment is often difficult for people with ID. It is also necessary to consider that some of the results may have varying degrees of circularity. For example, personality disorders include patterns of difficult behaviour as part of the diagnostic criteria. Using standard criteria (ICD-10) to diagnose in people with ID as in this study can be considered a methodological strength but also simultaneously a methodological weakness as it is possible that some mental disorders may be missed by only considering unmodified criteria. The issue of autism is also crucial when considering behavioural problems. The presence of autism causes increased difficulties with diagnosis of mental illness and personality disorder . Associations between the major diagnostic categories and behavioural problems were therefore relatively less difficult to clarify in this study due to the exclusion of people with a diagnosis of autism. Yet the exclusion of people with autism is also simultaneously a weakness of the study in that it has long been recognised that autism is associated with behavioural problems (2).
As this study was based on a referred sample it is not able to definitively address the conflicting findings from previous studies. A population based study would help further in this respect. It would be useful for future studies to look in more depth at the behavioural presentations associated with specific mental disorders diagnosed or at the clinical predictors of treatment outcomes. However despite this limitation the present study still strengthens the existing evidence base regarding relationships of demographic and clinical variables with behavioural problems in referred samples with good quality data and appropriate analyses. The findings provide further evidence why the assessments of people with significant behavioural problems must be comprehensive. A behavioural assessment such as functional analysis is essential but may not be sufficient on its own to understand all of the possible determinants or concomitants of the behavioural problems. Assessment must also include mental health assessment to determine whether or not there are disorders such as schizophrenia and personality disorders present or likely. If there is a history of behavioural problems and these are not currently present then care must be taken to consider whether or not a mental disorder such as an anxiety disorder might be decreasing an overall tendency for a person to display behavioural problems. Considering whether mental disorders are also present in a person with ID and behavioural problems should be an essential part of a holistic approach to their care.
- 9.Corbett JA (1979) Psychiatric morbidity and mental retardation. In: James FE, Snaith RP (eds) Psychiatric illness and mental handicap. Gaskell Press, London, pp 11–25Google Scholar
- 10.Crews WD, Bonaventura S, Rowe F (1994) Dual diagnosis: prevalence of psychiatric disorders in a large state residential facility for individuals with mental retardation. Am J Ment Retard 98:688–731Google Scholar
- 15.Hemmings CP (2007) The relationships between challenging behaviours and psychiatric disorders in people with severe intellectual disabilities. In: Bouras N, Holt G (eds) Psychiatric and behavioural disorders in intellectual and developmental disabilities, 2nd edn. Cambridge University Press, Cambridge, pp 62–75Google Scholar
- 16.Hillery J, Mulcahy M (1997) Self-injurious behaviour in persons with a mental handicap: an epidemiological study in an Irish population. Ir J Psychol Med 14:12–15Google Scholar
- 18.Holmes N, Shah A, Wing L (1982) The disability assessment schedule: a brief screening device for use with the mentally retarded. Psychol Med 2:879–890Google Scholar
- 26.Nihira K, Price-Wiilliams DR, White JF (1988) Social competence and maladaptive behavior of people with dual diagnosis. J Ment Handicap Persons 1:185–199Google Scholar
- 27.Office for National Statistics (2001) Census 2001. Stationery Office, LondonGoogle Scholar
- 32.Rojahn J (1994) Epidemiology and topographic taxonomy of self-injurious behaviour. In: Thompson T, Gray DB (eds) Destructive behaviour in developmental disabilities: diagnosis and treatment. Sage, Thousand OaksGoogle Scholar
- 34.Sigafoos J, Elkins J, Kerr M, Attwood T (1994) A survey of aggressive behaviour amongst a population of persons with intellectual disability in Queensland. J Intellect Disabil Res 348:369–381Google Scholar
- 36.Turkistani IYA (2004) Epilepsy in learning disabilities: relevance and association with mental illness and behavioural disturbances. J Learn Disabil 8:89–99Google Scholar
- 37.World Health Organisation (WHO) (1992) The ICD-10 classification of mental and behavioural disorders: clinical descriptions and diagnostic guidelines. World Health Organisation, GenevaGoogle Scholar