Assessment of the causes of self-injurious behaviour


When starting to assess the causes of self-injurious behaviour in people who have Cornelia de Lange Syndrome, it is important to remember that for any child or adult self-injurious behaviour has not always occurred, and right now it does not occur all of the time (it just feels like that sometimes!). This means that the self-injurious behaviour is likely to be related to either internal or external factors that have changed over a long period of time or change on a more short-term basis. Successful assessment of the causes of self-injurious behaviour rests on finding out more about the factors that appear to be present when self-injurious behaviour is occurring, and absent when it is not.
In order to uncover the causes of self-injurious behaviour it is important to adopt a systematic approach to looking for the factors that appear to be associated with the behaviour. To develop this systematic approach we have drawn on two areas of knowledge. First, we have considered relevant factors that might be specific, or at least more common, in children and adults with Cornelia de Lange Syndrome. Second, we have drawn on the general principles of assessing self-injurious behaviour in anyone who has an intellectual disability. In combination, we believe an approach based on these two bodies of knowledge will give us the best chance of understanding self-injurious behaviour in any individual with Cornelia de Lange Syndrome. In this chapter we will describe a systematic approach to understanding factors that may be causing self-injurious behaviour. Some of the assessments can be carried out by anybody who has a good knowledge of the person showing self-injurious behaviour. However, some assessments require more expertise and the reader is encouraged to seek collaboration and co-operation with the right professionals. We do believe it is always better to seek the support of others when conducting these assessments in order to ensure that the approach is both systematic and thorough.
This chapter is broadly divided into three main areas. First, we will consider assessment of internal factors that can be related to self-injurious behaviour. Second, we will consider assessment of external factors that might explain selfinjurious behaviour and third, we will consider assessment of factors that might account indicate a “loss of control” over self-injurious behaviour. It is important when thinking about assessing self-injurious behaviour to remember that the cause of self-injurious behaviour may not be related to a single factor and that a number of internal and external factors might be influential. It is also important to remember that causes can change over time in any individual and consequently it is important to repeat the assessment process at various points.

We are in the process of adding the boxes and graphs referred to

Assessing the role of pain and discomfort in self-injury

In Chapter 2 we noted that children and adults with Cornelia de Lange Syndrome may experience a number of short-term or long-term medical problems that can give rise to physical discomfort and pain at various times throughout their lives. We also noted that when people experience pain and discomfort they naturally try to relieve the discomfort, usually by physically stimulating the area of the body that is associated with discomfort. In particular, we noted that children and adults with Cornelia de Lange Syndrome can experience discomfort around their eyes (due to lack of tear production and eyelashes touching the eye), the ears (with middle ear infections), the sinuses, the teeth (either because of crowding of the teeth, the increased risk of tooth decay due to thin enamel or the effect of reflux), gastrointestinal pain in the upper chest and throat (due to the stomach acid irritating the lining of the digestive tract) and, possibly, joint pain. Any of these medical problems can give rise to localised pain or more general discomfort and this can lead to a self-injurious behaviour being focused on or around the site of the pain or discomfort.

In order to assess whether any of these medical reasons might be associated with self-injurious behaviour it is important to assess the following factors:

The site of the injury.

Is the self-injurious behaviour directed towards any specific site of the body that is associated with a medical condition that we know is more common in children and adults with Cornelia de Lange Syndrome? As we have noted above, this means is the self-injury directed towards the eyes (see Box 6.1), ears, cheeks and bridge of the nose (where the sinuses are), the teeth or mouth, the chest or throat (where the discomfort associated with a reflux would be found).

The association between self-injury and environmental events.

Self-injurious behaviour that is associated with pain and discomfort is unlikely to change as events in the environment change. This is because the cause of the pain or discomfort is highly unlikely to be related to environmental events. So, for example, the self-injurious behaviour will not be triggered by a change in the amount of social contact that an individual is experiencing, the person being asked to conduct a task or any other events that are occurring throughout the day (see Box 6.2). However, it may be the case that self-injurious behaviour that occurs more commonly after mealtimes than before them, is associated with gastrointestinal problems. Keeping a record of the times in a day that self-injurious behaviour occurs might reveal a pattern that shows that self-injurious behaviour is related to meal times in this way (see Section 6.3.4 on Scatterplots).

Change in other behaviours around the time of self-injury.

The third aspect of assessing whether self-injurious behaviour is associated with pain and discomfort is to look for other indicators of discomfort that appear to be present when the self-injurious behaviour is more common. For someone who has poor communication it is often difficult to know whether the person is experiencing pain and discomfort. Some of the more obvious signs that someone is experiencing discomfort may be facial expression and vocalisations, loss of appetite, poor or disrupted sleep and a generally higher level of movement. Any or all of these indicators when seen with self-injurious behaviour might indicate that pain or discomfort is an underlying cause for the self-injury. (Behaviours that we believe indicate gastrointestinal problems are described in Chapter 2).

It is, of course, extremely difficult to ever be certain that medical reasons are associated with self-injurious behaviour, particularly when the person is unable to communicate. However, the three indicators listed above would suggest that medical reasons should be explored prior to any other reasons. If more than one indicator is present then we would suggest that a possible medical reason is likely.

Assessing reward by stimulation or the presentation or removal of social contact

When thinking about reasons for self-injurious behaviour it is important to remember the information that was presented in Chapter 5 in which we discussed how self-injurious behaviour might be rewarded or reinforced by sensory stimulation (an internal factor) or social processes (an external factor). When beginning an assessment of possible psychological reasons it is usual to try to distinguish between self-injurious behaviour that might occur for sensory stimulation and that which might occur because of social processes. This is because the types of intervention that would be used are very different for self-stimulatory and socially reinforced self-injurious behaviour. Before describing the main assessment strategies that can be used to evaluate these reasons for self-injurious behaviour it is important to note a number of issues that always need to be considered. First, there are always risks and benefits associated with some psychological assessments.
Risk of injury whilst assessing is a possibility for some of the experimental methods of assessment which often require the creation of conditions under which self-injurious behaviour usually occurs to ensure that we are right about a particular cause. Under these circumstances the risks to the individual from showing self-injurious behaviour and the benefits of being more certain about the causes of self-injurious behaviour need to be balanced and thought through with others. It is also important to note that when we assess psychological reasons for self-injurious behaviour it is unlikely that we will be able to see a one-to-one relationship between self-injury and internal or external factors. It is more likely that we will just see a higher level of self-injurious behaviour when a cause is present compared to when it is absent, as opposed to being able to turn self-injurious behaviour on and off completely. It is also important that we think about different factors in different environments1. This is most commonly the case between school and home environments. The things that happen in school may be very different to those that happen at home and the way that teachers and parents might respond to self-injurious behaviour will also differ. When considering the reasons for self-injurious behaviour it is important to consider differences between the main environments in which the child or adult might spend their time.
Just as there are differences between environments that might be related to the reasons for self-injury, so the reasons can change over time. The reason for self-injurious behaviour to start in a young child might be very different from the reason that the behaviour continues over a longer period of time. The most common example of this is for self-injurious behaviour to begin for a medical reason or because of the stimulation that is produced by the self-injurious behaviour, and for the behaviour to then become socially reinforced which may maintain the behaviour over long periods of time (see Chapter 5)2, 3. For this reason it is important to remember that assessments for psychological reasons should be repeated periodically.
One of the most important aspects of assessment is systematic record keeping and the value of good records cannot be underestimated. Whatever assessment is conducted the results must be carefully documented so that a comparison can be made at different points in time or between environments. It does not matter that an assessment does not show a positive result i.e. that a factor cannot be said to be associated with self-injurious behaviour. The results of this negative finding must still be documented so that others can see the results at a later date and make a comparison should that be necessary. The key to conducting a good assessment is good recordkeeping. We outline some methods of keeping
records in Chapter 7.
There is a very large research literature covering the assessment of psychological reasons for self-injurious behaviour. Basically this literature identifies five different reasons that self-injurious behaviour may occur and these different reasons are directly related to the causes of self-injurious behaviour that are described in Chapter 5. These five reasons are:

  1. Sensory reinforcement or stimulation (an internal factor).
  2. Positive social reinforcement, most commonly by attention. (an external factor).
  3. Positive tangible reinforcement by the presentation of things or activities such as food, drink or toys etc.. (an external factor).
  4. Negative social reinforcement by escape from tasks or activities that the person does not want to do. (an external factor).
  5. Negative social reinforcement by escape from any social contact regardless of whether or not tasks or activities are involved. (an external factor).

When psychologists undertake the assessment of self-injurious behaviour they are trying to find out which of these reasons apply to the self-injury so that they can match the treatment to the reason. The assessment task therefore is to systematically evaluate each of these potential reasons.
There are a number of ways in which these reasons can be assessed that range from informal information gathering, by asking those who know the child or adult well and conducting informal observations, through to conducting experimental methods in which environmental events are systematically manipulated and the effect on self-injurious behaviour is recorded. In addition to these assessments of self-injurious behaviour there are some additional assessments which are important to conduct in order to help develop the best intervention and we will consider these towards the end of the chapter.

Informal information gathering
The benefits of this type of assessment are that it is quick to conduct and that it can cover a wide range of environments and events. However, a problem is that the information that is collected may be unreliable, and there is good evidence that if you ask different people about the events that are associated with self-injurious behaviour for any individual then you tend to get different answers. For this reason psychologists would usually use this method to collect some basic information and then would go on to use one of the other methods that have been described below.
Basically, when conducting this kind of assessment you are trying to identify the events that happen before self-injurious behaviour occurs (these are called antecedents) and the events that happen after self-injurious behaviour has occurred (these are called consequences). Thus, the overall analysis is called antecedentbehaviour- consequence or ABC for short and we outlined this idea in Box 5.7.

The main strategy in this assessment is to ask about different antecedents that appear to be occurring before self-injurious behaviour and that would indicate one of the five reasons that are described above. So, if self-injurious behaviour tends to occur more when there is a boring and unstimulating environment than when there are things to do, then it is likely that the self-injurious behaviour is maintained by sensory stimulation (and consequently the model outlined in a Box 5.4 is likely to be applicable). If self-injurious behaviour tends to occur when a carer is not attending to the person who shows self-injury and is attending to someone else, then it is likely that the self-injurious behaviour is maintained by a positive social reinforcement (and consequently the model outlined in a Box 5.5 is likely to be applicable). If self-injurious behaviour occurs more when the person is asked to do something they do not want to then it is likely that the self-injurious behaviour is maintained by negative social reinforcement of escape from task (and consequently the model outlined in Box 5.6 is likely to be applicable). The questions that might be asked and the informal observations that might be conducted in order to find out which of the five reasons outlined above are likely to be applicable are described in Box 6.3.

ABC and STAR charts
There is really very little difference between an ABC chart and a STAR chart and both seek to obtain a record of what happens before and after the self-injurious behaviour. STAR stands for Setting, Trigger, Action, Results and so gives a bit more information than ABC charts in that you learn about the settings in which the behaviour is likely to occur4. ABC and STAR charts are commonly used by psychologists to conduct the assessment of psychological reasons for self-injurious behaviour when the behaviour is occurring at a relatively low frequency, say a couple of times a day. If the behaviour is much more frequent then ABC and STAR charts tend to be less informative and the observational or experimental methods that are discussed below might be more appropriate.

There are two types of ABC charts that might be used for assessing self-injury. Open ended ABC charts simply require people to record in boxes the events that they saw just before the self-injury occurred (A, antecedent), a brief description of the behaviour itself (B, behaviour) and a description of the events that happened just after the behaviour (C, consequence). An example of a completed ABC chart is shown in the Box 6.4. together with an interpretation of how the completed form is related to the five causes.

A second way in which ABC or STAR charts may be used is by designing a closed-ended form that is completed for each incident or burst of self-injurious behaviour5. Using this method it is possible to help people who are keeping the records to keep an eye out for the types of event that would help to understand the possible reasons for self-injurious behaviour. An example of a closed ended ABC chart is shown in Box 6.5.

Research that we have conducted in the past in which we have compared the information that can be taken from open ended and closed ended ABC charts strongly suggests that the closed ended ABC charts give better information, probably because they help the person completing the charts to look for the right things.

When reviewing the completed ABC and STAR charts, you are looking for antecedents and consequences that might indicate which of the five causes appear to be important. For example, if the records show a pattern of low levels of attention and carers being with other people and then the consequences show that some form of attention was paid following the behaviour then it is likely that the behaviour is maintained by positive social reinforcement (see Box 5.5).

However, if the records show that prior to the behaviour occurring tasks are being presented or the person is being asked to do something and following the behaviour the task is no longer present, then it is likely that the behaviour is maintained by a negative social reinforcement by escape from a task (see Box 5.6). These are just two of the reasons that might be revealed by ABC and STAR charts and boxes 6.4 and 6.5 shows the sorts of antecedents and consequences that might be associated with the different reasons for self-injurious behaviour.

Questionnaire methods
There are two questionnaires that can be used to assess the factors that might influence self-injurious behaviour. They are the Motivation Assessment Scale (MAS)6 and the Questionnaire About Behavioral Function (QABF)7. These have been designed to ask specific questions about possible factors. The person who is completing the form gives a rating for different questions and these ratings can then be added to give a score for different causes. Examples of the questions that are used in the questionnaires are given in Box 6.6 together with an interpretation of what the responses might mean. Unfortunately these questionnaires are not easy to get hold of but Clinical Psychologists will usually be able to obtain them.

Scatterplots are slightly different from ABC and STAR charts and are usually used to get an indication of the times of day that the self-injurious behaviour tends to occur8. The advantage of this method is that it is not very time consuming for those who are being required to complete the charts and the information can be used as a good baseline by which to compare the effects of an intervention that is put into place following assessment. Basically, a scatterplot consists of a grid in which a rating of the frequency or intensity of self-injurious behaviour is made for each, say, half-hour or hour of the day. A completed example is shown in Box 6.7.

When reviewing a completed scatterplot it is helpful if the ratings that are made throughout the plot can be related to a diary of events and activities that the individual was involved in. This will help to try and identify the relationship between events and activities and levels of selfinjurious behaviour so that conclusions might be drawn about possible external factors that are related to the behaviour. So, if high level selfinjurious behaviour recorded in the scatterplot appears to be related to times at which there are one-to-one teaching sessions it is possible that the self-injurious behaviour is occurring because it is socially negatively reinforced by escape from the tasks (see Box 5.6). If high levels of self-injurious behaviour are recorded in the scatterplot at times when there is little activity and the person is left to their own devices, then it is possible that the self-injurious behaviour is occurring because it provides stimulation for the person. An example of the interpretation of the information from scattered plots are given in Box 6.8.

While scatterplots can give some very useful general information about the times of day that self-injurious behaviour is occurring, successful interpretation of the information depends on how well the level of self-injurious behaviour can be tied to a diary of events and activities, and how accurately the diary of events and activities is described. Sometimes, although self-injurious behaviour can clearly be shown to be associated for example with group activities, it is difficult to know whether this tells us enough to work out the reason for the self-injurious behaviour. It could be that the individual is not receiving a great deal of stimulation in this setting, and consequently self-injures because of self-stimulation, or it could be that there is shared attention and that the self-injury is reinforced by social attention from the carer who is working with the group. Results of scatterplots therefore should the reviewed alongside other assessment information.

Probability plots
Probability plots are similar to scatterplots but may prove to be more accurate when trying to evaluate whether a particular antecedent is associated with self-injurious behaviour and thus which of the five causes is important. We should emphasise that this is a relatively new method that we are currently developing in order to aid the understanding of the environmental reasons for self-injurious behaviour and to date it has not been widely used9.

A probability plot is conducted by a drawing up a chart that has times recorded down the lefthand side and environmental events across the top. There is also a box to record self-injurious behaviour. An example of a probability plot is shown in Box 6.9. Once the probability plot has been developed the observations are conducted throughout the day across a number of days. The method of observation to complete the plot is called Momentary Time Sampling (see Box 7.3) and involves the person who is spending time with the child or adult who is showing self-injurious behaviour making a record at given time intervals. In the past we have used ten-minute intervals although more frequent observations can be used.

To complete the records the carer does not need to watch the child or adult all of the time. Rather, they will wear a watch which will give a signal, say, every 10 minutes (this is usually called the countdown function on digital watches) and at that point the carer will look at the child or adult for about five seconds and make a record of which environmental event is occurring and whether or not self-injurious behaviour is occurring. If the carer is unable to make the observation precisely at the time at which the signal occurs then this does not matter. However, it is important for the carer to make the observation as soon as possible after the signal has occurred. The important thing is that the observation does not occur just because self-injury is happening. Once a number of probability plots have been completed the information can be analysed to try and identify a pattern of association between selfinjurious behaviour and the environmental events. This is a little complicated but is worth the effort. Basically, the analysis consists of working out the probability (or chance) that self-injurious behaviour would occur. This means dividing the number of times that self-injurious behaviour was observed by the total number of observations that were carried out.

The result of this calculation gives us a probability value. For example, if selfinjurious behaviour was observed on 58 occasions and the total number of observations was 325, then the probability of self-injurious behaviour is 58 divided by 325 or 0.18 (these are odds of about 5 to 1 against). We can then calculate whether the self-injurious behaviour is more likely to occur given that an antecedent event has occurred. This means only looking at the occasions in which an antecedent has occurred and then calculating the probability of self-injurious behaviour. 

For example, if the social event of tasks was recorded on 65 occasions and when we look at those 65 occasions we note that self-injurious behaviour occurred on 30 occasions then we would divide 30 by 65 and this gives us a probability of 0.46 (these are odds of about 2 to 1 against). When we compare the two probabilities we can see a difference (5 to 1 against is more of a long shot than 2 to 1 against). We describe this by saying that the simple or unconditional probability of self-injurious behaviour is 0.18 (5 to 1 against) and the conditional probability of self-injurious behaviour given that tasks are occurring is 0.46 (2 to 1 against). We can conclude from this analysis that self-injurious behaviour is approximately two and a half times more likely when tasks are occurring (0.56 divided by 0.18 or 5 divided by 2). When we look at our five causes it then seems likely that the reason for self-injurious behaviour is negative reinforcement by escape from task demands.
Examples of the calculations for this analysis are given in Box 6.10.

Unstructured natural observations
Unstructured natural observations are difficult to conduct and the information can be difficult to analyse and carers will need help and advice from a clinical psychologist or somebody familiar with behavioural techniques. Basically, the method consists of using either paper and pencil or a palmheld computer to record at each second or, for example within a ten second interval, the selfinjurious behaviour and the environmental events that are observed10, 11. It is then necessary to analyse the data by looking at the probability that self-injurious behaviour is associated with events that happen before or after the behaviour with the probability that self-injury will occur regardless of what is happening. Thus, the analysis is similar to the comparisons of unconditional and conditional probabilities described above. Further information about this method can be found in some of the texts on applied behaviour analysis listed in the Bibliography and we have included this information here for the sake of completeness.

Analogue or experimental conditions

Over the last 20 years psychologists have developed a method of assessing the psychological reasons for self-injurious behaviour that involves systematically manipulating environmental conditions and then observing the effect on self-injurious behaviour12. This methodology is usually referred to as applying analogue conditions (analogue meaning a model, in this case a model of what happens in the natural environment). The ideas behind analogue conditions are not themselves complicated but actually conducting the conditions can be difficult and it is important to use this method of assessments in collaboration with a clinical psychologist or another professional who is familiar with behavioural methods.

We have already described the different psychological reasons that might be associated with self-injurious behaviour. In the observational methods we have described so far we rely on the natural occurrence of environmental conditions and self-injurious behaviour in order to see whether the self-injury occurs more frequently when specific environmental conditions are seen. Analogue conditions are designed as an experiment in which specific environmental conditions are presented and the frequency or duration of self-injurious behaviour is recorded. So, instead of waiting to see if low attention or high task demands trigger self-injury, these conditions are artificially created to see if they make self-injury more or less likely. There are two main types of analogue conditions that have been used in the past that can be employed to assess both the social and stimulatory reinforcement that form the basis to our five psychological reasons for self-injury.

Brian Iwata and his colleagues developed the first method in the early 1980s13. In this method the person who shows self-injurious behaviour is exposed to four different conditions a number of times and the conditions are presented in a random order. The conditions are:

  • Alone: The individual being assessed is left alone in a room with no toys or other forms of stimulation. No social consequences are delivered if self-injury occurs. If self-injury occurs more in this condition than others then it is probably maintained by self-stimulation.
  • Task demands: An experimenter (or teacher/parent) provides instructional tasks to the individual, using a three-step prompting system (verbal, gestural and physical prompts). The task is removed for several seconds if self-injury occurs. High levels of self-injury in this condition would siggest that the behaviour is maintained by negative reinforcement in the form of escape.
  • Social attention: The individual has access to toys. The experimenter provides several seconds of attention (i.e. saying “don’t do that”) following self-injury, but ignores the individual the rest of the time. High levels of self-injury in this condition would tell us that the behaviour is maintained by positive reinforcement in the form of attention.
  • Control (or play): The experimenter and individual play with toys as normal but no demands are delivered. No social consequences are delivered if self-injury occurs. This condition serves as a control condition to rule out the effect of other factors such as the experimenter being present.

Each condition lasts about ten minutes and throughout the condition a record is kept of the frequency or duration of self-injurious behaviour. Each condition is usually conducted about seven to ten times in order to insure that the results are stable. Once completed the results are graphed in order to examine which condition appears to cause a higher rate of the self-injurious behaviour. Box 6.11 shows an example of how the results of analogue conditions can be graphed in order to evaluate the possible psychological reason for self-injurious behaviour.

The second method was developed by Ted Carr and Mark Durand and was first described in 198514. In this method the person who shows self-injurious behaviour is exposed to three different conditions in a systematic order. The conditions are:

  • High attention, easy task demand: In this condition, children carried out an easy task whilst receiving high levels of attention in the form of instructions, praise or neutral comments.
  • Low attention, easy task demand: In this condition, the task difficulty was the same as the baseline above but children only received attention in 33% of trials instead of 100%.
  • High attention, hard task demand: In this condition the child received attention on 100% of trials as in the baseline but the difficulty of the task was increased.

Again each condition lasts about ten minutes and a record of self-injurious behaviour is kept. Each condition is repeated approximately six times and the results graphed to examine which condition is associated with self-injurious behaviour. This method of analogue conditions only looks at the possible social reinforcement of self-injurious behaviour and not at whether the self-injurious behaviour occurs for self-stimulation. The interpretation of the results is carried out by comparing the rates of self-injury across the conditions. So, if the self-injurious behaviour occurs at much higher levels when there is low attention, easy task demands than when there is high attention and easy task demands it is likely that the self-injurious behaviour is occurring because of the low attention and consequently it is likely that the reason for self-injurious behaviour is to gain attention. However, if the self-injurious behaviour is occurring when there is high attention, difficult task demands in comparison to high attention, easy task demands then it is likely that the selfinjurious behaviour is occurring to escape task demands (social negative reinforcement by escape from task demands.)
Whilst these two methods of analogue conditions are the most commonly used there are numerous variations that have been employed in order to try and understand the psychological reasons for self-injurious behaviour. Researchers have used so-called brief analogues15, with each condition lasting only five minutes in order to conduct the assessment more efficiently, or have used just two conditions to assess whether selfinjury might be related to low attention (and thus socially reinforced by the presentation of attention). We used this method to look at the selfinjurious behaviour of some children with Cornelia de Lange Syndrome who attended the Chester conference in 2001. Box 6.12 shows the results of our assessments.

Additionally, researchers have explored some of the more idiosyncratic or unusual causes of self-injurious behaviour using analogue conditions. For example, for one child although generally difficult task demands did not necessarily cause self-injurious behaviour, when the demand was that the child took part in a medical intervention this did cause high levels of self-injurious behaviour16. This is presumably because in the past medical interventions were associated with pain and discomfort for the child, whereas normal task demands were not, and the self-injury was reinforced by the medical examination being terminated. This brief example shows us two things. First, that we can use analogue conditions in order to understand specific causes for any given individual and second, that sometimes the normal analogue conditions that are used to may not reveal causes that are important for every child. These idiosyncratic causes might only be revealed in the informal interviews and observations that are described above.

The assessment of setting events

In a Chapter 5 we introduced the idea of setting events and illustrated how events such as periods of illness or fatigue can influence self-injurious behaviour by interacting with an antecedent. The most common examples that appear in the research literature are interactions between somebody being tired, unwell or experiencing low mood and the presentation of a demanding task. This is an example of how internal and external factors can interact. Under these circumstances the self-injurious behaviour may occur because the task is made so much more unpleasant by the fact that the person is tired or in discomfort. The assessment of these potentially relevant setting events has been conducted in two ways. First, a record of the setting event (such as fatigue, low mood, being in pain or unwell) is kept using a diary or the scatterplot method. An additional record of the environmental events and self-injurious behaviour is also kept using the ABC or probability plot methods. The resultant records are then examined to see if the probability of self-injurious behaviour is higher when both the setting event and the antecedent are present as opposed to when either the antecedent or setting event is present. The second method is to combine a diary record of setting events with analogue conditions. Using this method the effect of a given analogue condition is compared when the setting event is present with when it is absent. So, if the self-injurious behaviour only tends to occur when a person is experiencing in low mood and the analogue condition of task demands is presented, as opposed to when either of these alone is present, then this would indicate that there is an interaction between a setting event and an antecedent. An example of this analysis is shown in Box 6.13.

Assessment of communicative and adaptive behaviours and the broader environment

In addition to the assessment of self-injurious behaviour it is also important to assess a number of adaptive behaviours that might be shown by a child or adult in order to provide the groundwork for increasing behaviours that might replace self-injury. There are two areas of assessment requiring particular attention that are extremely important when self-injurious behaviour is occurring because it is reinforced by social events such as attention, more tangible things such as access to materials or food and drinks, and escape from demands or social interactions. These are communication and other adaptive behaviours that help the person exert control over their environment.
You will remember from Chapter 5 (Section 5.4) and the descriptions in boxes 5.5 and 5.6 that when self-injury is socially reinforced we can think of the behaviour as being very similar to communication. For this reason we must attend to the ability of an individual to communicate with others and by this we really mean the ability of an individual to affect the behaviour of others. This is called the pragmatics of communication. Additionally, we must also attend to the adaptive behaviours that an individual may already show or could show that would allow them to have control over their environment and thus satisfy their own needs. If the child or adult is unable to control aspects of their environment directly then they will require others to help them to achieve this, and this makes it more likely that social contact is very rewarding. So, any assessment of self-injurious behaviour that is occurring because of social rewards should also attend to communication and other adaptive behaviour.

Loss of control

In Chapter 5 we considered the possibility that for some individuals who are showing self-injurious behaviour the psychological or medical reasons that we have outlined so far are less influential and that, for reasons that we do not fully understand, the behaviour seems not to be completely under the control of the individual. We would emphasise that research into self-injurious behaviour that appears not to be under the individual’s control has not really been able to explain why this might be, so consequently some of the information presented here is speculative. There is evidence that in Cornelia de Lange Syndrome the self-injury can become severe and it can become difficult to identify any medical and psychological reasons for the behaviour. It is possible that the behaviour has a “compulsive quality” and the individual has reduced control over the behaviour19.

Assessing self-injurious behaviour to determine whether there is evidence of a loss of control is not easy. We think there are two important factors that might indicate that self-injurious behaviour is not under the control of the individual. The first of these is self-restraint and the second is the presence of some other specific behaviours.

Self-restraint In some individuals who show self-injurious behaviour the person shows a preference for imposed restraint (will clearly prefer to wear items such as splints, gloves or a helmet) or shows self-restraint behaviours that they initiate themselves (such as covering or sitting on their hands). In the first instance a preference for imposed restraint may be seen when individuals actively help in putting on devices such as splints or helmets on and actively seek out these devices if they do not have them on. Additionally, they may become extremely distressed when these devices are removed and appear anxious, tearful or very angry. At the same time, when they do not have the devices they may also try to show forms of self restraint, apparently in an attempt to restrict their own movements.

Self-restraint is different because it does not involve a device that someone else has provided for the individual, and instead the individual has learned a way of controlling their own behaviour. The forms of self-restraint are numerous and Box 6.16 gives a list of the types of selfrestraint that we have observed in individuals with intellectual disability who show self-injurious behaviour and in individuals with Cornelia de Lange Syndrome who show self-injury. Selfrestraint can have some similarities to the preference for imposed restraint. Individuals may become distressed if the self-restraint is terminated or if they are showing a form of self-restraint such as covering hands with their sleeves and the clothes that they are offered does not allow this behaviour to occur.

The association with other behaviours.
Another indication that self-injurious behaviour may not be completely under the control of the individual or that it may in the future not be under control is the presence of compulsive behaviours. We would emphasise at this point that we do not have strong evidence that this is the case and that we are speculating that the presence of these compulsive behaviours indicates present or future lack of control. However, our research to date shows that for those individuals with Cornelia de Lange Syndrome who show self-injurious behaviour and compulsive behaviours, self-restraint occurs more than we would expect by chance. It is possible that this is explained by the self-injurious behaviour itself being compulsive, and consequently out of the individual’s control, and as a result the individual shows self-restraint in order to limit the behaviour.

Using assessment information

Once the assessments of these different potential causes of self-injurious behaviour have been conducted it is important to collect the information together and to start to build a model of why self-injurious behaviour might be occurring.
This is called the process of formulation and it is helpful because it will tend to show numerous potential points for intervention. In building a model of self-injurious behaviour it is likely that there will be different levels of strength of evidence for any given cause. It is also quite unlikely that there will be a clear-cut cause that indicates a single specific intervention that needs to be undertaken. It is more likely that there are a number of ways in which an intervention can be put together starting with changes which address the cause for which there is the strongest evidence.

The important thing about the assessment process, and how it is associated to interventions, is being prepared to collect careful information and being prepared to identify a number of causes, prioritise them and address them with interventions one by one whilst carrying out an evaluation of effectiveness. If you feel that a given cause might be influential you should be prepared to experiment to see what happens if you change things in a way that would tell you whether or not something is influential. The second important thing is to keep an open mind in terms of the potential causes of self-injurious behaviour. It is important to note that when we look for the causes of behaviour there is a strong tendency for us to seek information to confirm what we already think. Rather than do this, it is much more important to look at all the evidence and to work systematically through the assessment process from start to finish and to keep looking hard for evidence that contradicts what we think as well as evidence that confirms what we think.

Finally, if you have worked through the assessments and you are unable to find any way in which self-injurious behaviour appears to be linked to any of the reasons that have been described then all is not lost. There are still interventions that we can try and whilst you may not be guided towards a particular intervention you can still work through different interventions keeping a record of their effect on self-injurious behaviour and operate on a trial-and-error basis. That is, instead of putting into place an intervention because you know the reason for the selfinjurious behaviour you can try an intervention and see if it works. There is nothing wrong with this approach when we cannot find causes. There are two final and important aspects of completing the assessment process. First, you must share the results of the assessment with anyone else who is involved with the child or adult who are showing self-injurious behaviour.

You must also be prepared to listen to contradictions from others that indicate that the model you have built of self-injurious behaviour might need to be modified in some way. Finally, whatever model of self-injury you build at one point in time may not be appropriate later on. You must be prepared to change the model over time and to go back and look at specific assessments, repeat assessments and collect more information.


Systematic assessment of the causes of self-injurious behaviour underpins successful intervention. The first step is to evaluate whether the selfinjury is related to pain and discomfort by looking carefully at the site of the injury, the variability of self-injury across environmental events and the presence of other behaviours that might indicate pain and discomfort. If this reason for selfinjury is ruled out, then sensory and social reward should be considered. This will mean collecting informal information, record keeping and observing to examine the association between the self-injury and environmental events. When this process is complete the information should be shared with the key people who have contact with the child ready for the intervention to begin.


Chris Oliver

by Chris Oliver, Jo Moss, Jane Petty, Kate Arron, Jenny Sloneem, Scott Hall.

Page history
Last modified by Gerritjan Koekkoek on 2021/05/23 22:28
Created by Gerritjan Koekkoek on 2021/05/23 21:59



About the website contents

All of the information on this WebSite is for education purposes only. The place to get specific medical advice, diagnoses, and treatment is your doctor. Use of this site is strictly at your own risk. If you find something that you think needs correction or clarification, please let us know at: 

Send a email: