Monday, 25 March 2013

Autism: Stresses and positives for families

The negatives and positives of being a parent of a child with autism

We are used to hearing that having a children with autism in the family can be stressful for parents. Charitable (non-profit) organisations also push this line, and can make it sound like an inevitable consequence of being a parent of a child with autism. This latter implication is perhaps important in raising funds for projects to support families of children with autism. However, there is a significant danger in sending an unclear message about parenting a child with autism. This blog focuses on a number of different questions the answers to which may paint a fuller picture of parenting a child with autism: 1. Whether parents of children with autism are more likely to report stress or other psychological problems compared to other parents, 2. Does increased stress affect all families, 3. Whether positive perceptions and positive parental health are different in families of children with autism, and 4. Why parents of children with autism might report increased stress (if indeed they do so!).

Answers to all of these questions can be derived from a recent research study of ours (

Most existing disability family research has relied on surveys of parents in contact with support services (health, social care, education) or charity/non-profit organisations. Even when recruitment is carried out more widely and on a large scale, if researchers directly try to recruit families of disabled children it is very difficult to know who responds. Some parents may be more likely to respond if they are under considerable stress (they may want to make this clear to others), but other parents under considerable stress may be less likely to respond (they have more pressing priorities). Some parents may not want to identify themselves as having a disabled child, others will be proud to do so. These examples make it clear that data from most research studies will be biased and very unlikely to be drawn from a sample representative of any country’s population.

One solution to this problem is to find studies that include data from a sample representative of a country’s population and within which it is possible to identify disabled children and their parents. Although large scale population-based surveys are typically not designed to answer research questions directly about the families of disabled children, secondary analysis of the data can sometimes be conducted to do just that. We (Totsika et al., 2011) analysed data from the 1999/2004 Office for National Statistics (ONS) population surveys of mental health in children and adolescents in the UK ( The surveys were designed to be representative of the whole population of children in the UK, had a very high response rate, and included data on well over 18,000 children in total. Within this dataset, it is possible to identify children likely to have an intellectual disability, and also to identify children with autism. Autism was identified using a “psychiatric” interview method, and not simply via parents reporting that their child had a diagnosis of autism. So, the data are likely to be representative of the whole population of children with autism living in private households in the UK – not just those currently diagnosed or in contact with services. Data on mothers were available in the ONS surveys in addition to information about the children.

More stressed?

Using a screening measure included in the ONS datasets that can be scored to indicate high levels of emotional problems, 23.7% of mothers of children without intellectual disability or autism scored above the cut-off suggesting that further detailed psychological assessment would be worthwhile. These proportions were higher for mothers of children with autism and intellectual disability (43.6%), mothers of children with autism and no intellectual disability (43.5%), and mothers of children with intellectual disability without autism (32.4%).

We can glean to things from this pattern of findings. The first is that mothers of children with autism, whether or not the child also had intellectual disability, were about 1.8 times more likely to have high levels of emotional problems than mothers of children without these disabilities.

The second conclusion is also important. The majority of mothers of children with autism did NOT report elevated levels of emotional problems. In addition, these results were drawn from a measure designed to screen for emotional problems. The proportion of mothers who may meet criteria for a diagnosable emotional disorder following a detailed clinical assessment is likely to be lower.

Positive mental health

Using the same four groups of mothers, we also looked at the proportion of mothers who had high levels of positive mental health. 28.5% of mothers of children without intellectual disability or autism reported high levels of positive mental health, and this was at a similar level to the mothers of children with intellectual disability only (27%). The proportions of mothers with high positive mental health were slightly lower for the autism and no intellectual disability group (17.4%), and the autism with intellectual disability group (22.5%).

Although these proportions are lower for high positive mental health, they were not statistically significantly different in this study. This confirms findings from other studies that parents of children with autism report considerable positive outcomes for themselves, and perhaps at a level essentially no lower than other parents.

Although parents (especially mothers) may experience more stress when they have a child with autism, at the same time they seem to experience positive well-being to a similar level as other parents.

Why are mothers of children with autism reporting more stress?

There does seem to be a reliable increased risk of stress and other emotional problems for mothers of children with autism. Why might this be? Although it is not the full story, much of the increased risk for stress seems to be related to two variables. These variables were found to predict maternal emotional disorder in the Totsika et al. (2011) study and several other research studies have found the same pattern of findings. First, mothers are most stressed when their child also has significant levels of behaviour problems. Second, mothers are more stressed when they have limited family resources. In the Totsika et al. study, limited family resources were measured in terms of when families had fewer socio-economic resources (e.g., parents out of work, low levels of income, living in deprived neighbourhoods). In other studies, social support networks have also been found to be important: mothers with the highest levels of stress report smaller, less supportive networks.

What does this all mean for supporting families?

The message that I would like to get out there from this research is that parents (certainly mothers) of children with autism are at increased risk for psychological problems, but also the majority of parents are actually coping very positively. Even when parents are reporting stress, they also have positive things in their lives and view their child with autism positively – and they love them of course just like any other child! So, everything is not doom and gloom for parents of children with autism.

Saying that, it is also important to point out that a proportion families face significant challenges. A good piece of news is that the main factors associated with increased stress for mothers of children with autism are things we can do something about! There are evidence-based parenting interventions that can help families manage the behaviour problems displayed by many children with autism. In addition, increasing families’ access to financial and social supports should also help to reduce their experience of stress.

I recognise that I have described results from only one study in this blog and also that the data focused only on mothers. In addition, the data are a snapshot and probably many more parents at some point in their child’s life (whether or not the child has autism) will experience significant stress. However, my reading of the research literature is that the findings I have described are typical. So, the evidence tells us something important about families at risk, why they might be at risk, and also actually there are some pretty obvious ways we could better support families of children with autism. Let’s do it!

Totsika, V., Hastings, R. P., Emerson, E., Lancaster, G. A, & Berridge, D. M. (2011). A population-based investigation of behavioural and emotional problems and maternal mental health: Associations with autism and intellectual disability. Journal of Child Psychology and Psychiatry, 52, 91-99.

Tuesday, 19 March 2013

Autism and Evidence 5: 15 criticisms of ABA and some responses

 15 criticisms of Applied Behavioural Analysis approaches

For many years now, I have heard several criticisms of the use of ABA approaches to working with children and young people with autism. The typical response from supporters of ABA approaches is to say that these criticisms are ill-informed, incorrect, or relate to the vested interests of the proponents of other approaches who cannot identify strong evidence for their own interventions. In this blog, my aim is instead to take each criticism (gathered from a number of sources – special thanks to those who contributed to the list!) and provide a perspective on them. The reason for taking this approach is that criticisms may well have come from people’s direct bad experiences, or from secondhand reports of poor practice. Thus, these points do have to be addressed.

I need to make several points before I begin:

·      I am no spokesperson for ABA. I do hope that I understand the ABA field to some extent but also much more broadly autism, intellectual disability, families, and education. Thus, I am trying to use all of these perspectives to reflect on common criticisms of ABA. However, it is in no sense an “official” response.
·      I am also not writing this for experts in ABA, although I do hope that ABA practitioners will find my observations helpful. In any case, I will not attempt to be technically accurate at all points. Instead, I am trying to explain things for a wider audience.
·      Several of the criticisms that I will deal with below are more generally about poor practice and are not specific to ABA. Thus, there is certainly a need to improve the quality of practice across the board when it comes to working with children and young people with autism.
·      Many relevant points have already been addressed in earlier blogs of mine. I will do my best to refer back to those blogs at appropriate points – so that people can read a fuller treatment of the issues. However, I would encourage everyone to read the whole series of blogs on autism and evidence.

1.         ABA has a “normalizing” agenda

The criticisms here focus on a number of related points. The first is that ABA approaches are focused on taking away something of the child’s autism – trying to make the child “normal” in some way. Critics argue that this also leads to proponents of ABA approaches trying to convince parents and others that they can “cure” a child of autism, or more generally that “ABA can lead to recovery” from autism. A second area of criticism is that ABA focuses on reducing/removing behaviours that cause no harm for the child and in fact are functional for them (especially stimulatory behaviours – “stims”).

Some discussion of this issue can be found in two of my previous blogs, so please take a look at those first:

In summary, ABA approaches in autism do not “decide” which behaviours to focus on in terms of developing new skills or reducing existing problematic behaviours. Multiple perspectives are used to identify targets for intervention including the child, their family, teachers etc. Ethically speaking, behaviours ought to be the target for reduction only when suitable alternatives are available for children (and these may have to be taught), and when the behaviour in question is clearly interfering with an aspect of the child or family’s experience of quality of life. Sometimes, “stims” may be targeted for reduction for these reasons, but there is no prescription to do so as a part of an ABA programme.

When it comes to a broader normalizing agenda, this is an issue that is not specific to ABA. In fact, educationally focused interventions in general ought to be about making a positive difference and not succumbing to a medical model agenda of cure or recovery.

That said, unfortunately, it is true that some proponents of ABA sell their services on the basis that recovery “is possible”. Thus, many people may have heard these qualified promises, or perhaps stronger promises made. However, proponents of many other interventions in autism also make similar (sometimes much stronger) claims. The fact is that many interventions exist within a marketplace where they are trying to attract attention of parents as “consumers”. Until autism organizations and government bodies recommend (and fund) only interventions with a clearly demonstrated evidence base, this marketplace will continue to function.

An important point to make is that just because some individuals or organizations argue that ABA can lead to some sort of recovery from autism does not mean that this is what ABA is all about.

2.         Children fail to generalize skills

Under this general heading of criticisms are perhaps two related points. First, children become dependent on one-to-one teaching, or learning in particular contexts only. Second, one-to-one teaching does not encourage independence.

A defining feature of interventions that can be called Applied Behavioural Analysis is that they directly address generalization of learning (to new teachers, settings, and skills) and also the maintenance of newly acquired skills in the child’s typical environment. It can be hard to establish generalization for children with autism whatever intervention approach you chose. However, ABA includes a variety of methods and perspectives to directly teach generalization of skills. If a child on an ABA intervention has not yet generalized their new skills, the programme or intervention is not yet completed.

I can imagine that people will have seen children reported to be on ABA programmes who are currently failing to generalize their skills. However, this is most likely a feature either of where the child has got to (generalization is still being worked on), or perhaps a reflection of the lack of competence/experience in their intervention team. In neither case is the failure to generalize an inherent feature of ABA.

In fact, quite the opposite. ABA approaches explicitly recognize the need to teach and establish generalization, and (unlike many other intervention approaches) practitioners have worked for many years to establish methods of teaching to maximize generalization. Some of these methods will still rely on one-to-one teaching. It is not one-to-one teaching that causes problems in terms of generalization, but the failure to properly plan for and teach generalization causes generalization problems!

3.         ABA is adult led

I think this criticism focuses on the experience of ABA programmes for children with autism where a teacher (therapist, tutor) presents as many learning opportunities to the child as possible. Although an adult will also incorporate natural learning opportunities (see below), the adult can perhaps be seen as the main active participant in the intervention. My understanding of this criticism is also that the implication is that this active involvement of an adult teacher is somehow “bad”. So, perhaps one reflection on this criticism is to ask why that might be a bad thing. The alternative might be to offer very little by way of teaching support to children with autism who may just not interact with other people much at all.

A more positive perspective is to clarify that ABA interventions are very clearly child focused and individualized. Thus, initial targets for intervention are defined by young people themselves, their families/carers/advocates, and the expectations of society. Delivery of an ABA programme also focuses on the hour-by-hour collection of data about the child’s learning. Thus, bespoke evaluation is used. In addition, reinforcers (in common language often called “rewards” – also see below) are chosen for the fact that they can be used to motivate an individual child (through a formal process of reinforcer assessment). Also, ABA programmes begin with a very detailed individual assessment of the child’s strengths and weaknesses that informs what should be taught first and what might follow once foundational skills are established.

So, depending on what critics mean by “adult led” perhaps ABA interventions can be described in this way. However, they are also child centred and individualized at multiple levels.

4.         ABA is rigid and reduces the opportunity to respond naturally to spontaneous initiations and interactions with the child

This may lead on from the previous criticism about being adult led. Any high quality ABA programme will, however, incorporate learning opportunities within the typical environment. Some approaches major on this approach – natural environment teaching/training is a widely applied method. All high quality ABA programmes will also take advantage of learning opportunities outside of formal teaching sessions. This is one reason why family members are often encouraged to be involved in a child’s programme – so that they can help with the maintenance of the child’s learning within their day-to-day environment.

Perhaps the error is to focus only on one part of ABA intervention approaches – the more formal and typically table-top delivered Discrete Trial Training method. For me, DTT is about direct teaching of new skills that are practiced so that the child becomes fluent in the skill within the teaching environment. Next steps are obviously to extend these new skills to other settings and to ensure that the skills can be maintained in the child’s everyday environment (generalization and maintenance again). Unless a child becomes fluent in a skill, it is hard to achieve long lasting intervention effects. This is an error in some other intervention approaches where it is not clear if a child truly masters a skill, and the likelihood of successful generalization and maintenance is likely reduced.

So, perhaps DTT might look “rigid” to an outside observer not looking at the whole context of a child’s programme. I would also be concerned if a child’s programme only consisted of DTT because it seems unlikely that proper attention is being paid to generalization and maintenance of skills. Thus, “rigidity” may be serving an important purpose within ABA programmes, but ongoing reliance on DTT especially later on in a child’s programme MAY be an indication of poor practice. Thus, ABA isn’t rigid but it may be applied in lower quality programmes in a more rigid manner.

5.         Reliance on “external rewards”

There is a collection of criticisms that seem to me to focus on the idea that ABA interventions somehow force a child to learn. Children are offered “rewards” to perform/behave in certain ways and so this is not real learning – the children are not learning real skills. A related criticism is that ABA interventions make the child too reliant on the structure of the intervention and so they do not learn to occupy themselves.

These points touch on the motivation to learn. I cannot see how a child can be forced to learn. It is certainly the case that skilled ABA practitioners are very good at motivating a child to engage with learning. The reason that reinforcers are applied in the context of ABA interventions is because basic research on learning demonstrates that we learn through the gradual shaping through reinforcement of successful ways of behaving. New skills, once learned to a fluent level, “self-maintain” because they lead to success in the child’s environment. The whole process of generalization and maintenance (see above) is all about “real” learning and long term change in behaviour.

Why do children with autism often have to be taught using “rewards” delivered reliably by a teacher in the context of the intervention programme? Because they have often failed to learn key skills by other means. Therefore, they need to be taught these skills. The way to teach new skills is by the behavioural teaching methods used in ABA interventions.

An important outcome of any intervention, and ABA is no exception, is that a child develops learning-to-learn skills. These skills can be taught like any other. They might be examples of what ABA folk sometimes call pivotal skills or behavioural cusps – skills that once learned open access to other learning reliant on these building block skills. In addition, children can be taught self-management skills. They can monitor/record their own learning and make adjustments to their approach to learning based on this information.

6.         Lack of focus on sensory issues

Some experts in the field of autism adhere to a position that sensory sensitivities are a core feature of autism. In the new DSM definition of autism, such sensory issues are to be more centrally represented. The implication seems to be that any intervention for children with autism should include a piece that states clearly how sensory issues are addressed.

Sensory issues ARE special in ABA intervention but at a very individual level – there is no one set of intervention supports that would be thought of as the “sensory bits”. First, sensitivities to sensory information need to be understood to be able to plan for how best to teach an individual child. Second, certain sensory stimuli may have unusually aversive properties for a child with autism. This would also be discovered in the context of a reinforcer assessment (how best to motivate the child to learn, and by implication how not to motivate them!). Third, on some occasions a child or young person may seek help to tolerate sensory experiences that they struggle with in the service of a higher level goal. For example, working in the film and television business may require tolerance to bright lights. ABA methods can often be used to enable children and young people with autism to develop tolerance to self-manage their sensory sensitivities. Complete avoidance is often not an option for those who wish for inclusion in society and independent living.

7.         The speech children produce when on an ABA programme isn’t “real”

This criticism is that ABA practitioners do not work on the precursors to language skills first when they teach children to speak. However, my understanding would be that any good quality ABA intervention would be developmentally informed. A good understanding of typical development is needed to inform which skills to teach in steps towards a more complex skill and in what order these steps should be taught. The most important issue is also that language is functional – that the child can use it to obtain what they want to obtain. Thus, the real test of whether good skills have been taught is whether they work in the child’s environment.

Speech production itself is also probably a good example of an area of knowledge where other professionals are expert and could be collaborating with those who know best how to teach children complex skills (ABA practitioners).

8.         ABA is a one size fits all approach

ABA is a child centred and individualized approach as described at several points in generating answers to other common criticisms, and in my earlier blogs. What is taught and how this is done, will vary depending on the child’s strengths and weaknesses, theirs and their parents’ wishes, and how quickly they learn the skills.

In thinking about this criticism, which just doesn’t make sense to me given my opening comments, I suspect that it may come from seeing practitioners at work who aren’t yet experts. In all areas of practice, experts are those so fluent in an intervention approach that they know the theory and practice inside out and can problem solve quickly and effectively as they go along. Thus, they are very good at responding to individual variations and creating new ways of supporting individuals where the intervention isn’t quite working. This is no different for ABA practitioners. Early on in people’s development, there is likely to be a reliance on using manualised approaches and perhaps ABA programmes may look quite similar across different children. Things will look very different when you watch a real expert at work.

Again, my point is that “one size fits all” is not a feature of ABA intervention per se. Rather, it is perhaps an indication of inexpert delivery of any intervention approach. At heart, ABA is much more of an individualized approach to intervention.

9.         Use of punishment

Several critics point to the use of punishment in ABA programmes historically and also very rare cases internationally of services calling themselves ABA and using electric shock, amongst other things.

Punishment is not a process that leads to the learning of new behaviour, but it can be used to stop children engaging in challenging behaviours that may be dangerous to themselves or others. The use of punishment is outlawed or very heavily restricted in codes of ethics and professional practice in the field of ABA generally (as it is by all professional bodies) and so also in ABA interventions applied to children with autism.

ABA practitioners are also trained to a much higher level than other practitioners to recognize and understand punishment so that they can avoid its unethical application. This is actually quite a technical point. What most people mean when they talk about punishment is either degrading practices (which must be outlawed on an ethical basis), or things that many people find aversive. Punishment in ABA is different in the second respect because what is aversive is known to be very individual (just like the things that positively motivate each child). In addition, “punishers” are defined by their effects on behaviour. Most typically, if something happens after a child’s behaviour that leads to that behaviour being less likely to occur in similar situations in future, that “something” is acting as a punisher.

In summary, modern ABA programmes and practitioners do not use punishment. Practices agreed to be degrading are outlawed. At the same time, ABA practitioners have a clearly developed understanding of what punishment actually is and how individual it is. Thus, they should be less likely than other practitioners to punish a child’s behaviour. Without a nuanced understanding of punishment, other practitioners may be inadvertently punishing a child’s behaviour. This is a potentially serious issue in practice.

10.         Lack of professionalism amongst ABA practitioners

For me, this is an area again where a category error is being made by critics. Thus, behaviour that may be considered less than professional that they might observe in some ABA practitioners is somehow attributed to the approach per se rather than to the individuals concerned, or the system in which they work. Common criticisms include that the qualifications of ABA staff are not clear, little attention is paid (in the UK) to the National Curriculum, and ABA staff will not collaborate with other professionals in school settings.

Putting it very simply, ABA interventions do not require practitioners to be unprofessional.

That said, there are very important questions to be asked about how ABA interventions should be best governed at least within the UK education system. Because local education authorities have typically failed to invest in ABA-based services, the ABA interventions that they have funded have historically been as a result of parental demand and even parents seeking legal redress to get funding for their child’s ABA programme. Thus, education authorities have been reactive rather than proactive. This means that the staff with the ABA skills are often outside of the system. And, Yes, this can mean that people are employed to work on ABA programmes who lack a broader professional training (despite the fact that they may be very good at directly teaching children).

My argument would be that the UK education system must properly engage with ABA practice and work out how to make the benefits available to as many children with autism (and other special educational needs) as possible. The ABA community in the UK has carried out a piece of collaborative work to address professional standards including clarity about the range of competencies that high quality practitioners should be able to demonstrate. The UK ABA Autism Competencies Project focused on ABA competencies but also competencies relating to autism, working professionally, and working within the four countries of the UK in the education system (including the context of the National Curriculum).

My message to local authorities – stop using perceived lack of professionalism in ABA as an excuse for ignoring its benefits. Instead, engage with the ABA practice community.

Follow this link to the competencies framework for a free download of the full information:

11.          ABA programmes can be never-ending, and the long term benefits of ABA are not clear

Never ending ABA programmes is a criticism that probably deserves a separate discussion. In fact this is quite an interesting issue. The first is that this criticism often comes from local education authorities in the UK who typically fund the interventions. It is not clear why they haven’t got appropriate governance in place to prevent extended early intervention programmes from running. From families’ perspectives I can see why they would want to extend funded intervention time for as long as they can. There may also be confusion about what ABA intervention actually is (see Many people confuse Early Intensive Behavioural Intervention with ABA more generally. My own position is that running an EIBI model beyond early childhood (i.e., when intervention would be “early”) is not appropriate, but ongoing use of ABA teaching methods throughout life for a person with autism is a very good idea.

Evidence-based support shouldn’t end just because young people with autism move beyond early childhood. However, I can also see that an unchanging ABA programme in adolescence that still looks like the intervention the child was receiving at 3-4 years of age is unlikely to be sensible.

In terms of long term benefits, it is the case that long term follow-up of children who received ABA early intervention has not yet been the subject of significant research effort. Thus, in that sense the long term outcomes of EIBI as an early intervention are unknown. However, again, this is not unique to ABA intervention. There are no significant similar data either for other autism interventions. There is, however, international consensus about the value of early intervention in autism and the notion of intervening early for longer term impact on quality of life.

If this “long term benefits” criticism relates to the value (or otherwise) of extended ABA intervention, this point has been addressed above already.

12.         ABA is very American and doesn’t work for the UK education system

Both of the major UK-based evaluation studies that I’ve been involved in as a researcher have been carried out within the UK education system and in collaboration with local education authorities. In SCAmP, we worked with 11 education authorities in Southern England to deliver and evaluate a home-based EIBI model (Remington et al., 2007). In the Westwood School project, we worked with two education authorities in North Wales and delivered and evaluated an early years school based ABA model integrated into a mainstream school (Grindle et al., 2012).

Both studies reported positive results for the children with autism who received the intervention. We certainly did have to work hard to develop a model that used the best of ABA intervention evidence and delivered this appropriately within the UK education system. However, all of the many people involved (children, parents, teachers, education administrators, ABA practitioners, other professionals, researchers) worked together to deliver the positive outcomes. It can be done in the UK.

13.         ABA is too intensive and expensive

As I have mentioned before in this blog series, I do not believe that ABA comes with a prescription for intensity. International consensus, not “ABA”, suggests that early intervention for autism should be intensive. There are some indications that intensity of ABA intervention is associated with outcome. However, intensity only explains a small amount of the variability in outcome in EIBI. There are likely to be several other factors in play.

ABA practitioners strive to deliver the best quality that they can with the resources available. Resourcing decisions are ones that are made by education authorities and other official bodies.

Those points made, there are several examples of evaluation studies in the research literature (e.g., Eldevik et al., 2012) suggesting that ABA interventions produce better outcomes than eclectic special education funded at the same level of cost. It seems that you can spend the same amount of money and still get better outcomes.

14.         ABA is hard on family life and family members

I am always interested to hear this criticism since our research group has been one of the few worldwide to carry out research on this question. It is often assumed that an intensive home-based intervention programme may be stressful for parents and other children in the family. This assumption is unlikely to apply to ABA programmes run primarily in school settings away from the home. However, even in the home context the research data do not support the anecdotal assumptions. Stress did not increase over time in parents whose children received EIBI in the family home (Remington et al., 2007), parents of children with autism on home programmes and siblings do not seem to have elevated levels of psychological problems compared to other parents/siblings (Hastings & Johnson, 2001; Hastings, 2003), and although parents did report that the intervention can be stressful they would do it all again and strongly recommend ABA to other families (Grindle et al., 2009).

15.         ABA is only for the “severe” children

A final criticism that I will address at this point is one that ABA might be useful only for those with severe autism and/or who have significant challenging behaviours. Higher functioning children do not need ABA.

By this point, I should ask you the reader to generate your own answer to this criticism. My view is that again we are probably dealing here with a mis-understanding about what ABA is. ABA teaching technologies are for all children and young people – nothing to do with autism, and certainly nothing to do with the severity of autism. Most children at some stage will require some additional support to develop a new skill, manage problematic behaviour, or extend their skills in an area where they are already expert. The evidence-based way to teach children when they need this assistance is to use methods informed by our understanding of learning theory. Perhaps this criticism is linked to a mis-understanding that EIBI is the only ABA intervention. I can certainly see that perhaps children with the highest levels of need are the ones who may need a comprehensive early intervention programme – teaching them across multiple skills domains. Children and young people with more specific needs would still benefit from a focused ABA intervention dealing with a particular domain of skill development (e.g., reading, numeracy, social skills, managing anxiety).

Thank you for reading, and sorry this ended up being so long – there was a lot to say!!

References to research studies mentioned above

Eldevik, S., Hastings, R. P., Jahr, E., & Hughes, J. C. (2012). Outcomes of behavioral intervention for children with Autism in mainstream pre-school settings. Journal of Autism and Developmental Disorders, 42, 210-220.
Grindle, C. F., Kovshoff, H., Hastings, R. P., & Remington, B. (2009). Parents’ experiences of home-based Applied Behavior Analysis programs for young children with autism. Journal of Autism and Developmental Disorders, 39, 42-56.
Grindle, C. F., Hastings, R. P., Saville, M., Hughes, J. C., Huxley, K., Kovshoff, H., Griffith, G. M., Walker-Jones, E., Devonshire, K., & Remington, B. (2012). Outcomes of a behavioral education model for children with autism in a
mainstream school setting. Behavior Modification, 36, 298-319.
Hastings, R. P. (2003). Behavioral adjustment of siblings of children with autism engaged in applied behavior analysis early intervention programs: The moderating role of social support. Journal of Autism and Developmental Disorders, 33, 141-150.
Hastings, R. P., & Johnson, E. (2001). Stress in UK families conducting intensive home-based behavioral intervention for their young child with autism. Journal of Autism and Developmental Disorders, 31, 327-336.
Remington, B., Hastings, R. P., Kovshoff, H., degli Espinosa, F., Jahr, E., Brown, T., Alsford, P., Lemaic, M., & Ward, N. J. (2007). Early Intensive Behavioral Intervention: Outcomes for children with Autism and their parents after two years. American Journal on Mental Retardation, 112, 418-438.

Wednesday, 6 March 2013

Recovery from autism: false hope, or the wrong question?

There has been considerable media interest in recent weeks in the question of whether children recover from autism – whether they can be “cured” of autism. This interest has come in response to a new research study published in the Journal of Child Psychology and Psychiatry by Deborah Fein (from the University of Connecticut) and colleagues (JCPP, 2013, volume 54, issue 2, pages 195-205).

The Fein et al study focuses on the idea of optimal outcome for some children with autism. The researchers found a group of children who were diagnosed earlier in childhood as having autism and then when followed up they did not meet diagnostic criteria for autism any more. Thus, these data have been discussed in terms of the possibility of recovery from autism. The published study is very carefully done and all effort has been made to try to ensure that the initial diagnoses were valid and that currently these optimal outcome children do not seem to have autism.

The potential for recovery makes a great media story. Experts have weighed in on both “sides” – some suggesting that recovery may be possible and others simply stating that autism is a lifelong difference or disorder and so recovery/cure is not possible. I have been asked several times whether these research results, and the stories about them, offer false hope for families. This blog is a reflection on some of the issues as I see them.

Are the data believable?

As already stated, the Fein et al research has been carried out very carefully. However, the authors themselves are careful to point out the limitations of their design. Most importantly, because diagnosis is an imperfect process they cannot rule out the possibility that the children were misdiagnosed earlier in childhood. Since researchers (at least currently!) cannot go back in time to examine their sample as younger children, we can never know whether misdiagnosis happened. The children were recruited precisely because they were once thought to be autistic and were now thought not to be. Thus, they were diagnosed by a variety of professionals in a variety of different geographical areas. Therefore, the factors in play at the time of their diagnoses are likely to be many and varied.

A second issue is that it is not enough to show the existence of a group of children who once did but do not now “have” autism. We would need to be able to say how this happened. Although not the focus of the published study, media reports have suggested that the research team have some other evidence to suggest that many or all of these “recovered” children have had intensive early interventions. This is certainly a hypothesis that is worth exploring. However, a clear causal theory underlying the purpose of an early intervention would also be needed (see my Autism and Evidence 1 blog –

Let me explain this theoretical point. My understanding is that there is as yet no single identified cause of autism. In fact, a long list of potential variables has been identified as linked to autism including genetic factors, other biological causes, and social and environmental factors. We would need to be able to track backwards through a logical argument to establish with more certainty whether recovery has happened. That is, we would need a clear statement of what the causal factors are, how to turn this information into an early intervention format, and then to show that this intervention (and not something else) resulted in the loss of an autism diagnosis. I doubt if this logical model, with related intervention research evidence, is available to support the idea of recovery.

Some researchers have suggested that we can look to changes in the brain to see if early intervention has “normalised” brain function in children with autism exposed to early intervention. Given that existing early intervention models currently rely on increasing skills or teaching new skills to children, then brain changes would be expected – our brains change in response to experience such as new learning. However, understanding what such research might mean in terms of recovery is very complicated, and would require some demonstrated links between “biology” and behaviour.

In summary, there are limitations to the research evidence so far, and also a serious lack of evidence clarifying how recovery might happen if indeed it does.

The wrong question?

A more fundamental problem for me is that discussion about recovery or cure in autism is not a helpful question in the first place. There are two main approaches to understanding or working with autism, and only one of these leads one to ask questions about recovery. Following a medical model, autism (putting things simply) is a disorder with a biological underpinning. From this perspective, it makes sense to explore medical causes of autism (e.g., genes, environmental poisons, and differences in various biological processes in the body). A medical discourse is also strongly rooted generally in “cure”. Consider, for example, research and media representations of cancer – basic medical research seeks to understand why cancers occur and to use this information to devise new pharmacological or other treatments.

Non-medical perspectives on autism are varied but essentially they have a social-educational focus. These varying perspectives consider autism as a difference, and understanding the nature of the difference is important. Support or intervention is then not about treatment but about developing ways to ensure the best quality of life for individuals with autism and their families and carers.

Early intervention models with existing good evidence (e.g., the Early Start Denver Model, Early Intensive Behavioural Intervention and the like) have a clear educational focus – they are about developing or accelerating important cognitive and social skills in young children. These interventions do not simply focus on trying to change or take away autistic behaviours. Thus, I believe that their underlying rationale is to achieve the best outcomes that can be achieved for every individual child with autism. Early educational interventions were not designed to somehow cure autism but to make a positive difference to each individual child and their family.

So, here’s the point. Linking educational early intervention to the notion of recovery or cure is mixing two completely different understandings of, or approaches to, autism. These approaches are, in my mind, directly opposed.

In the absence of convincing biomedical research and “treatment”, we should be asking the “right” question – what can be done to provide the best quality of life outcomes for children with autism and their families? And not just for children. Whilst millions are spent on biomedical research, we are struggling for evidence about how best to organise education for children with autism, what social care services in adulthood will lead to the best outcomes, and what supports are needed for adults with autism to engage with employment and the independent life they want to lead.