Definition of learning disability and identification of populations

The effect that the definition of learning disability can have on the results of epidemiological studies is illustrated by the example of the change in the American Association on Mental Retardation’s definition of mental retardation. In 1973 this was changed from one standard deviation below the mean on an IQ test to two standard deviations below, with the result that many individuals included in learning disability studies before 1973 would now be excluded.

Identifying populations of adults with learning disability from case registers or those in receipt of specifically targeted social funding or learning disability services has a good ascertainment rate for moderate-to-profound learning disability but is not so good for mild learning disability. Typically, adults with mild learning disability are not known to learning disability services or, if they are, it is because they have additional problems such as mental illness, and this leads to a biased sample. Furthermore, there is considerable variation in the methods used to set up and maintain such case registers and therefore samples taken from different case registers, even within the UK, are not always directly comparable.

Accurate epidemiological information can be obtained only by using reliable and valid measurements on population-based samples. The ideal method would be to screen everyone in a certain area for learning disability and then to further screen those identified for mental health problems using clinical assessment and appropriate diagnostic criteria, as discussed below. However, this would be costly and time consuming. A few studies have used population-based samples, but they are limited by small sample sizes.

Definition and identification of mental health problems

In prevalence studies, many researchers have used terms such as mental illness, mental disorder, psychiatric illness, psychiatric disorder, emotional problems and behavioural disorder without detailed definition. Some have excluded personality disorder and behavioural disorder from their reported prevalence rates, whereas others have included them. This difference in the types of disorder counted can have a considerable effect on reported prevalence rates and makes the comparison of studies complicated.

Sampling bias

Epidemiological studies are seriously affected by the target population chosen for study and this has been particularly problematic in learning disability. Many prevalence studies have used institutionalised or psychiatric out-patient populations, both of which are likely to have a high prevalence of mental health problems. However, a significant number of learning-disabled people with mental health problems are unknown to services, and using samples that are not population-based excludes this group.

Total prevalence and level of ability

Studies examining the prevalence of mental health problems across differing levels of ability have produced conflicting results. Some have found higher rates of psychiatric disorder in people with mild learning disability than in those with severe or profound disability, whereas others have found lower rates. The higher rates in mild learning disability are at least partly explained by biased sampling. This has two causes. First, those with mild learning disability that are known to learning disability services tend to have additional problems such as psychiatric, physical or social disorders. Second, it is difficult to identify and diagnose a psychiatric disorder according to standard diagnostic criteria in people with severe and profound learning disability.

When considering possible aetiological factors such as epilepsy, physical disability and sensory impairment, it makes sense that there would be an increased rate of mental health problems in people with more severe learning disability and this has been described by Reference Cooper and BaileyCooper & Bailey (2001). However, before this issue can be answered definitively further work needs to be done on population-based samples large enough to demonstrate differences and using assessment methods and diagnostic criteria appropriate for all levels of ability. Also, in view of the different manifestations of different disorders across different levels of ability, examining specific disorders rather than total prevalence rates may be more relevant in studying this question.

Total prevalence and age

As with the level of ability, the literature has inconsistent findings with regard to total prevalence and age. Several studies have found no differences in the age distributions of groups with and without mental health problems, some have found a higher prevalence of mental health problems in older adults and some have found a reduced prevalence in older age groups (Reference DayDay, 1985). Reference CooperCooper (1997) found the prevalence of psychiatric morbidity to be 68.7% in a group aged over 64 years and 47.9% in a group aged 20–64 years. Most of this additional morbidity was accounted for by increased rates of depression, anxiety disorders and dementia. Reference Deb, Thomas and BrightDeb et al (2001a ) reported a statistically significant association between the rate of psychiatric illness and increasing age in his study of 101 learning-disabled adults aged 16–64 years. Factors likely to increase the prevalence of mental health problems in older people include the increasing sensory deficits and physical health problems associated with age, the cumulative effect of life events and the association of certain conditions such as dementia with age. However, the effect of differential mortality, i.e. those with more severe learning disability die at a younger age than those with milder levels because of associated physical disabilities, probably operates in the opposite direction. Again, larger studies with better methodology are required to determine whether any specific age group is more at risk than others.

Total prevalence and gender

The relationship between gender and mental health problems in adults with learning disability is also still unclear. Most studies have found that gender is unrelated to the overall rate of psychiatric disorder. However, since in the general population female gender is associated with a higher rate of affective and anxiety disorder and male gender with a higher rate of psychosis, it might be more helpful to look at the effect of gender on specific disorders rather than total prevalence rates. Studies so far in adults with learning disability have been of numbers insufficient to demonstrate such an effect, if it is there.

Total prevalence and epilepsy

In the general population it is widely accepted that epilepsy confers an increased risk for mental health disturbances. Depression, anxiety disorders and psychosis are all common in people with epilepsy. However, it is uncertain whether this is also the case in people with learning disability. Reference LundLund (1985a ) found a rate of 52% for psychiatric diagnosis in people who had had seizures within the previous year compared with 26% in those without seizures, and Reference Corbett, James and SnaithCorbett (1979) found a rate of 60.8% in those with epilepsy and 40.4% in this without. Reference Deb and JoyceDeb & Joyce (1998), however, found no increased rate of problem behaviour or psychiatric illness in learning-disabled people with epilepsy. More recently, a study by Reference Espie, Watkins and CurticeEspie et al(2003) found a rate of psychiatric disorder according to the PAS–ADD Checklist of 33% in a sample of 186 people with learning disability and epilepsy. They compared this with the 33% PAS–ADD Checklist prevalence found by Reference Roy, Martin and WellsRoy et al(1997) in a community sample of 127 adults with learning disability identified from a learning disabilities register and through liaison with primary healthcare teams, and concluded that epilepsy in itself was not a risk factor for psychiatric disorder. However, normative data for the PAS–ADD Checklist used with adults with learning disability have been published since then (Reference Taylor, Hatton and DixonTaylor et al, 2004) showing a prevalence rate of 20.1%, suggesting that there might indeed be an increased rate of psychiatric symptoms in people with epilepsy.

Total prevalence and physical illness

A relationship between physical and mental health has been demonstrated in the general population. In a community-based sample of people aged between 16 and 64 years with mild-to-moderate learning disability, Reference Deb, Thomas and BrightDeb et al (2001a ) found a statistically significant association of physical disability with psychiatric illness, but Reference Moss, Goldberg and PatelMoss et al (1993a ) failed to demonstrate any such relationship in a community sample of people with severe learning disability aged over 50 years. This may have been due to the difficulties in identifying psychiatric disorders in people with severe learning disability and the confounding effect of increased physical morbidity with more severe levels of learning disability. Reference CooperCooper (1997) found a relationship between physical ill-health and dementia in older adults with learning disability, but not between mental and physical ill-health.

The existing information on the total prevalence of mental health problems in adults with learning disability is summarised in Box 2.

Life events and abuse

Although conclusive evidence on the relationship between mental ill-health and factors such as level of learning disability and physical illness is still lacking, there has been some progress in our understanding of life events as possible aetiological factors. A recent study by Reference Hastings, Hatton and TaylorHastings et al(2004) measured the rate of life events in the preceding year for a large sample of adults with learning disability in the north of England. They found that 46.3% had experienced one or more life events, a rate similar to that in the general population. They also found that the presence of one or more life events in the preceding year was significantly associated with scoring above threshold on the affective/neurotic sub-scale of the PAS–ADD Checklist. There are some limitations to this study, but it does suggest that life events are associated with affective and neurotic symptoms in adults with learning disability.

When comparing a group of 50 people with learning disability who had been bereaved with a matched control group, Reference Hollins and EsterhuyzenHollins & Esterhuyzen (1997) found a higher rate of depression, anxiety and adjustment disorders in the bereaved group. However, that group had also experienced more life events, which may have affected this result.

A systematic review of the literature on the clinical effects of sexual abuse in people with learning disability (Reference Sequeira and HollinsSequeira & Hollins, 2003) found several studies suggesting that a range of psychopathology, including traumatic stress reactions, depression, anxiety and behavioural problems (e.g. aggression, self-injury and sexual behaviour) may follow sexual abuse. However, because of methodological limitations the results are not conclusive.

From these studies it seems likely that adults with learning disability experience at least as many life events and are at least as vulnerable to these as are adults without learning disability and that affective and neurotic disorders, and probably other types of mental ill-health, are associated. Further work evaluating the role of life events in the mental health of this population needs to be performed.

Psychosis

Most of the research literature on psychotic disorders focuses on people with mild learning disability. There is very little on psychosis in people with moderate, severe or profound disability. This is mainly because of the complexity of diagnosing these disorders in people with more severe levels of learning disability. First-rank symptoms are conceptually too difficult for most learning-disabled people to express and negative symptoms and thought disorder may be inappropriately counted as symptoms of psychosis when they are in fact due to institutionalisation or are developmentally appropriate. Nevertheless, it is recognised that there is an increased prevalence of psychotic disorder in people with learning disability compared with the general population and this is one of the most consistent findings in the literature.

Several studies have reported the prevalence of schizophrenia in learning disability as 3% (and this is probably an underestimate) compared with 1% for the general population. The explanation for this increased prevalence of psychosis in people with learning disability and the direction of the relationship is much less apparent. The concept of pfropfschizophrenie was revisited by Reference Doody, Johnstone and SandersonDoody et al(1998) in a between-group study comparing people with mild learning disability and a history of schizophrenia and two control groups: one of people with schizophrenia but no learning disability and the other of people with mild learning disability but no history of psychosis. They found that individuals with schizophrenia and learning disability had a tendency to belong to families with a history of both disorders, showed a high rate of chromosomal variants in karyotypic testing and were more likely to have negative symptoms, epilepsy, neurological soft signs and episodic memory deficits than individuals with schizophrenia and no learning disability. The authors concluded that a severe form of schizophrenia may occur where schizophrenia and learning disability arise together from a common genetic aetiology.

Other genetic associations with psychosis that have been described include Prader–Willi syndrome and velocardiofacial syndrome. There are three genetic subtypes of Prader–Willi syndrome (deletion, maternal disomy and imprinting defect) and, interestingly, the association with affective psychosis is now known to relate principally to the genetic subtype of maternal disomy. One study found a 100% prevalence of psychotic symptoms by age 28 years in a sample of people with this genetic subtype of the syndrome (Reference Boer, Holland and WhittingtonBoer et al, 2002).

There have been no other recent epidemiological studies that have examined psychotic disorders in adults with learning disability.

Affective disorders

Affective disorder in people with learning disability has attracted a high level of research interest, but the unresolved difficulties encountered in its assessment and diagnosis have hampered progress in determining prevalence rates and associated factors. Most authors agree that standard diagnostic criteria are appropriate for diagnosing depression in people with mild learning disability, but there is some disagreement as to whether modifications are required for people with moderate-to-profound learning disability. Furthermore, the debate as to whether behavioural disorder is an atypical symptom of depression is yet to be resolved. However, from the studies that have been carried out it seems that the point prevalence of depression in people with learning disability is around 3–4%. This is considerably higher than the 1.7% found in the general population (Reference Meltzer, Gill and PetticrewMeltzer et al, 1995).

Risk factors for depression in the general population include stress, life events, low socio-economic status, lack of social support, old age and female gender. These factors are shared by people with learning disability, who may be further disadvantaged because of their limited coping skills, experiences of discrimination, rejection, ridicule, stigma and abuse. However, whether they are also risk factors for depression in learning disability has not been adequately studied, although some results have been published. For example, Reference MeinsMeins (1993) found that both men and women with little support tended to have more symptoms of depression than well-supported individuals. Reference LunskyLunsky (2003) found that learning-disabled individuals with higher self-reported depression scores were lonelier and had higher stress levels than individuals with lower scores. None of the population-based studies have demonstrated a relationship between affective disorder and gender in learning disability, but this may be because of a lack of power due to the small sample sizes. However, a study by Reference Hastings, Hatton and TaylorHastings et al(2004) of a population-based sample of over 1000 adults with learning disability demonstrated a significantly increased rate of affective and neurotic symptoms, as measured by the PAS–ADD Checklist, in women than in men.

These results do suggest that some of the risk factors for depression in people with learning disability might be the same as those for the general population. There is also the possibility that additional and as yet unknown factors are involved in the aetiology of depression in people with learning disability. An interesting study by Reference Richards, Maughan and HardyRichards et al(2001) demonstrated a fourfold increased risk of affective disorders for people with mild learning disability even after controlling for social and material disadvantages and medical conditions. A higher prevalence of depression in adults with Down’s syndrome than in adults with other causes of learning disability has been suggested by some researchers (although not found by others) and a relative serotonin deficiency has been proposed as the reason for this increased vulnerability (Reference Collacott and BarnesCollacott, 1999). Clearly, further research in this area is required to confirm this and establish whether there are any treatment implications.

Anxiety disorder and obsessive–compulsive disorder

The identification of anxiety and stress disorders in people with learning disability is particularly difficult and they are frequently missed by carers and clinicians, primarily owing to communication problems. In severe and profound learning disability only behavioural symptoms can be assessed, and as a result many anxiety disorders are misdiagnosed as problem behaviours. Classification is also difficult because anxiety can be a perfectly normal and appropriate reaction but it also occurs in a number of very different psychiatric conditions. As a result of these difficulties, much of the epidemiology of anxiety disorders in adults with learning disability is still unknown and studies have tended to describe the occurrence of anxiety symptoms rather than specific disorders. A few studies have examined the types of fear of adults with learning disability and they seem similar to those reported by children of equivalent developmental age (except that adults with learning disability report more fears of doctors). However, the usual gender difference (female gender is associated with greater reporting of fears) has not been found. When a group of 30 adults with learning disability were compared with a similar group without learning disability, the learning disability group reported more fears (Reference Pickersgill, Valentine and MayPickersgill et al, 1994). The authors suggested that overprotection and poor linguistic skills, leading to greater difficulties in discussing or dismissing fears and resulting in over-generalisation, were responsible for this.

Other suggested aetiological factors for anxiety disorders in people with learning disability include the effects of deinstitutionalisation and certain behavioural phenotypes. Fragile-X syndrome is associated with social anxiety, and phenylketonuria and Williams syndrome with generalised anxiety.

Obsessional thoughts are hard for people with learning disability to describe and, although compulsions are more readily observed, it can be difficult to distinguish them from stereotypies, tics and autism. Despite these diagnostic problems, Reference CooperCooper (1997) found a rate of 2.5% for obsessive–compulsive disorder in learning disability, which is higher than in the general population. The most common compulsions in the general population are hand-washing, checking and cleaning, but in adults with learning disability, ordering is the most common, with hand-washing, cleaning and checking rarely seen.

Behavioural disorders

Some researchers believe that problem behaviours are symptoms of underlying psychiatric disorder, whereas others think that they are purely learned. All agree that problem behaviours are common in learning disability (although exactly how common very much depends on the definition of problem behaviour used and the sample studied). Reference Qureshi and AlborzQureshi & Alborz (1992) focused on highly disruptive, seriously challenging behaviour and found a prevalence rate of 5.7%. In a similar study, Reference Emerson, Kiernan and AlborzEmerson et al(2001) found that 10–15% of people known to local education, social and healthcare services showed challenging behaviour. At the other end of the spectrum, Reference Deb, Thomas and BrightDeb et al (2001b ) found a rate of 60.4% for any behavioural disorder of any frequency or severity.

In people with learning disability, the most common forms of problem behaviour identified by the various studies are demanding behaviour, verbal aggression, physical aggression, destructiveness and self-injury. The majority of those with problem behaviour have more than one type, with some having as many as five different problem behaviours. Most studies have found a higher prevalence of problem behaviours in males, although Reference Deb, Thomas and BrightDeb et al (2001b ) found a significantly higher prevalence of severe problem behaviours in females. The general trend is towards a greater prevalence of problem behaviour with increasing severity of learning disability, although people with profound learning disability tend to exhibit less outwardly directed behaviour because of their restricted mobility.

The overall prevalence of problem behaviour in learning disability increases with age during childhood, reaches a peak in the age range 15–34 years and then declines. Reference Thompson and ReidThompson & Reid (2002) found that a high number of behavioural symptoms persisted over a 26-year period in a cohort of 53 adults but the severity decreased. Reference Moss, Emerson and KiernanMoss et al(2000) found that increasing severity of challenging behaviour was associated with increased prevalence of psychiatric symptoms as measured by the PAS–ADD Checklist, with depression showing the most marked association. These authors also found self-injury to be associated with anxiety symptoms, a relationship suspected by other researchers. A study by Collacot et al(1998) on a population-based sample of 2277 adults found a prevalence of 1.7% for severe self-injury. No gender difference was apparent but self-injurious behaviour was associated with younger age, lower IQ, autistic symptoms, hearing impairment and immobility.

Another aetiological factor in problem behaviour is genetic disorder. Prader–Willi, Lesch–Nyhan, Aicardi, Rubenstein–Taybi and Smith–Magenis syndromes are all associated with self-injury. Angelman syndrome is associated with impulsivity and overactivity, Sotos syndrome with aggression and Cri du Chat syndrome with irritability. Physical health problems such as toothache and any febrile illness are also important precipitants of problem behaviour. Although some studies have found an increased rate of problem behaviour in people with epilepsy, others have found the opposite, and it remains unknown whether epilepsy plays a role in the aetiology of problem behaviour. The epidemiology of problem behaviours in adults with learning disability is summarised in Box 3.