Why is the evidence for Early Intensive Behavioral Intervention for Autism so weak?

One to one autism therapy is pricey – is it worth it?

Only a handful of countries widely apply behavioral interventions to treat toddlers diagnosed with autism.  Behavioral interventions include Applied Behavioral Analysis (ABA), Verbal Behavior (VB), Pivotal Response Treatment (PRT) and the Denver model.

Even after several decades, the published evidence that these interventions actually work is quite weak.  This explains why most countries do not readily provide public funds for ABA.

In the US, efforts are being made to diagnose autism at younger and younger ages, because the child can then benefit from these “proven” interventions, that other countries do not believe work.  Who is right?  

You can read Manuel Casanova’s perspective at the end of this post.  He is not such a fan of expensive US developed therapies and concludes:-

"spending time with your children and group socialization, in my experience, have provided the most favorable outcomes"

Does ABA work?  If so, why can’t you prove it?

From my personal experience, behavioral intervention was very beneficial as a teaching method, but it does not make autism go away.

In today’s study the aim was to determine if behavioral intervention is cost effective.  The conclusion based on all the studies considered is that there is no conclusive evidence that behavioral intervention is cost effective.  So logically the countries that do not widely fund it, like the UK, can be reassured that they are on the “right side” of the argument.

My view is that is that autism is so heterogeneous you can prove almost nothing, with any degree of certainly.  It is always going to be a case of ifs, buts and maybes.  This also very much applies to clinical trials of drugs to treat autism.

Why did ABA ever catch on in the first place?  People want hope and the more expensive something is, the more people want it.  Forty hours a week of ABA is very expensive and nice to have, if someone else is paying.  

We saw in an earlier post that Lovaas (the founding father of ABA) later admitted to selectively retiring non-responders from his clinical trials, to improve the apparent success of his methods.  This pretty much means you have to ignore all his data and his papers should be retracted. 

Many parents want curative treatments for autism.

Lovaas claimed that ABA is curative and that the treated kids end up like typical kids.  Sadly, this is an exaggeration.

Is two years of ABA cost effective for severe autism?  I guess it depends whose money is paying for it.  Is two years of ABA going to be life changing for a person with severe autism?  Unfortunately, even after 20 years of ABA, that person will likely still have severe autism, if you have not treated their underlying biological problems.

Some parents rave about ABA and make comments like “after two years of ABA my son now makes eye contact”.  Great, but would you pay $120,000 of your own money for that?  I think not.  Should your local government regard that as money well spent?  I think they should be more demanding; the results of just $1,000 spent on the right personalized medicine will be much more impressive.

Today most people currently being diagnosed with autism have mild cases.  If they can talk and do not have intellectual disability (ID) / mental retardation (MR), they will likely see little benefit from 40 hours a week of discrete trial training.  It would be a huge waste of money and probably just annoy the child.  

Many children with mild autism need a different kind of therapy, they need to learn social and emotional skills they may not naturally possess - how to make friends, how to avoid making enemies and so how not to get bullied at school.  This will only be effective started very young, before being a victim becomes a badge of honour.

Interventions based on early intensive applied behaviour analysis for autistic children: a systematic review and cost-effectiveness analysis

Autism is a lifelong condition that affects how people understand the world and interact with others. Early intensive applied behaviour analysis-based interventions are an approach designed to help young (preschool) autistic children. This approach is often delivered on a one-to-one basis, for 20–50 hours per week, over a period of several years.

This project obtained and analysed the original data from studies of early intensive applied behaviour analysis-based interventions, to determine whether or not these interventions are beneficial. It also investigated whether or not the interventions represent good value for money.

The results suggest that early intensive applied behaviour analysis-based interventions may improve children’s intelligence, communication, social and life skills more than standard approaches. However, some results could be inaccurate or incorrect, and there was no evidence about other important outcomes, such as the severity of autism and where children went to school. Most studies lasted for around 2 years, which means that it is not known if early intensive applied behaviour analysis-based interventions have meaningful long-term benefits.

It was not possible to fully assess whether or not these interventions provided value for money, as the benefits of early intensive applied behaviour analysis-based interventions were unclear, although the available evidence suggested that they did not. Early intensive applied behaviour analysis-based interventions may, however, provide value for money if their effects were to last into adulthood, or if receiving early intensive applied behaviour analysis had a large impact on the type of school children attended.

Future studies of early interventions may be helpful, but should consider looking at which components of early applied behaviour analysis-based interventions are the most important, rather than at whether or not they work better than other interventions. Future studies should also follow best current research practice and evaluate outcomes that matter to autistic people and their families. 

Economic evaluation

Using National Institute for Health and Care Excellence decision rules to benchmark the results of the cost-effectiveness analysis and adopting a £30,000 (USD 40,000) per quality-adjusted life-year threshold, these results indicate that early intensive applied behaviour analysis-based interventions would need to generate either further benefits or cost savings to be considered cost-effective.

Implications for service provision

Although individual participant data meta-analyses have shown small to moderate improvements in child cognitive ability and adaptive behaviour for early intensive applied behaviour analysis-based interventions relative to treatment as usual or eclectic approaches, all of the identified studies were at risk of bias, limiting the strength of conclusions that can be drawn from these results. Furthermore, results from individual studies varied considerably, with some showing no relative benefit of early intensive applied behaviour analysis-based interventions. 

Conclusion

For cases of severe autism, if you can afford intensive (and expensive) 1:1 intervention of any credible kind (Floortime, ABA, Denver etc - whatever works best in your case) it makes sense to use it.  It should improve skill acquisition and will make the parents feel better.

None of these interventions are curative, the child will still have autism.  When you no longer pay for the 1:1 intervention, the effects most definitely will start to fade away.  Don’t mortgage your house to pay for ABA.

Nothing stops you making your own 1:1 intervention program using family, friends and volunteers.  This does not cost much and is sustainable over many years; it is likely to be much for effective that 2 years of "professional" therapy.

I do find it odd that in the US there is free early intervention for toddlers and then provision just stops, as if it suddenly is no longer needed.

If you use ABA to teach a child to tie shoe laces, he/she will retain the skill as long as you keep buying shoes with laces.  If you do not practice/apply the skill for 6 months, do not be surprised if it has to be re-taught.

Our final ABA consultant was very experienced, she worked for 10+ years in the US before moving home to Athens, Greece.  She told me that in her experience all children with autism benefit from ABA, but the level of progress they make varies widely.  If a child does not respond to ABA, it very likely is not being done correctly.  ABA should be seen as fun, not like a punishment. If your child hates ABA sessions, they have no chance of working.

I come back to my earlier recommended strategy. Find your most effective novel medical treatment, which will inevitably be a polytherapy and combine this with a method of learning that works best for your particular child.

Then just keep going and let time do its work.

In countries like the UK, with free health and education provision, the government does not generally pay for early intervention because their medical advisors do believe it to be cost effective, which really means they think it does not work and so do not want to pay for it.  The cynic might just say they do not want to fund it. 

The idea was supposed to be that by investing upfront in ABA during the early years, you save money later on, by having a more functional child and then adult who requires less expensive provision.  Unfortunately, there is absolutely no proof this is true.  

If you go from early intervention, to an ABA special school and then ABA college, things clearly did not work out.

In the US early intervention is assumed to be very effective and the current idea is that doctors should hurry to diagnose autism before 24 months so as to get into the intervention program as soon as possible.  Where is the evidence to support the US view?  Are US outcomes any better?

We saw in recent research from UC Davis that looked at outcomes over time in autism that the best outcomes are not associated with any particular therapy.  The best outcomes happen because of the biological characteristics of that child, rather than any amount of behavioral intervention.

I expected the UC Davis study to show a relative benefit for those who received ABA therapy, but it did not.  We do have to take note.  I am actually pro-ABA and have spent a vast amount of money on this kind of therapy and 1:1 instruction.   

Ignoring treating the biological dysfunctions in autism while spending hundreds of thousands of dollars on 1:1 therapy and special education does not make a lot of sense.

Here is a relevant excerpt from a recent post by the neurologist, autism researcher and autism Grandfather, Manuel Casanova, from his Cortical Chauvinism blog: -

AUTISM: HOW CULTURE INFLUENCES OUTCOMES

Despite marked differences in geography, non-Westernized countries see autism as a social responsibility rather than a medical condition.  These countries offer a collectivist perspective that downplays individuality and prioritizes maintaining relationships within a given group of people.  In this regard, I have often marveled as to how vastly different countries, like Colombia and the more desolate regions of Eastern Russia (Siberia), share similar perspectives regarding autism. Indeed, due to a lack of resources, interventions in these countries are usually parent-mediated and heavily influenced by cultural norms.  Lack of personnel trained in behavioral analysis has been supplanted by art and music instruction.  Classes are provided in group settings where outperforming other members is not seen as conductive to the overall benefit of the group. Members are encouraged to adopt the norms of the group while teachers emphasize cooperation and nurturing. Students arrive early to school to participate in team building exercises.

I have often marveled at the achievements of troupes of autistic children performing autochthonous musicals and their accompanying choreography.  Adopting the norms of the group have served them far better than any Westernized behavioral intervention.  Participants in these groups seem genuinely happy; in part, given the sense of achievement at contributing to a piece of artistic expression.  In addition, the structured activities in such groups offer norms that minimize uncertainty.  Participants feel a sense of security in a group that fast becomes their extended family.

Autism is a medical condition but, without a cause that we can target, treatment options have remained symptomatic.  This is one of the reasons for looking at other countries and learning what has worked for them.  Indeed,  I believe that we can gain from adopting the cultural perspective of other countries to benefit our own children. Whether it is an improvisation on an autism chair, electroacupuncture, or using a zen bowl, spending time with your children and group socialization, in my experience, have provided the most favorable outcomes.

Manuel is one of a very small group of thoughtful researcher-clinicians, who have been working in the field of autism for decades, like Dr Kelley from Johns Hopkins and that psychologist Dr Siegel who wrote the Politics of Autism and revealed how Lovaas really did his "research". 

Manuel's researcher son-in-law is interested in precision medicine and drug re-purposing, I guess driven by his own young son's rare genetic "autism", NGLY1 deficiency. This very severe condition leads to the body not being able to breakdown and remove damaged and misfolded proteins.  You would think that reducing Endoplasmic Reticulum (ER) stress, that produces misfolded proteins, might be useful. This was covered here, along with a long list of possible therapeutics:-

https://epiphanyasd.blogspot.com/2019/12/er-stress-and-protein-misfolding-in.html

Some readers are following the details of the Covid-19 situation.

The Indian Experiment rather than the Swedish Experiment

A recent study suggests that more than half of the 6 million slum dwellers in Mumbai have had Covid-19; another 6 million do not live in slums. Government research showed that in the capital Delhi 23% have Covid-19 antibodies.

Mumbai slums have an extremely high population density, extreme poverty and so not much social distancing. So they show what Covid-19 does with no serious intervention, better than Sweden does.  Mumbai has reported 6,200 deaths in total.

You can extrapolate from the data (57% of slum dwellers and 16% of non slum dwellers with Covid antibodies) for the total 12 million population of Mumbai.  4.4 million had the virus and 0.14% died.  In the worst case scenario, when everyone finally gets infected in the next few years, there would be another 7.4 million with the virus and another 10,800 deaths.  The death/mortality rate for the city would be 0.14%.  (In reality it will probably be less than 0.14%, because some people will not get the virus)

The 0.14% Covid-19 mortality rate compares to the 2.5% mortality rate of the 1918/9 global flu pandemic; worse still that flu pandemic affected fit young people the most, making the demographic impact huge. 

The crude death rate from all causes in the US is around 0.8% each year (just 0.7% in India).  That puts the 0.14% from Covid-19 into some perspective. If Americans are as healthy as Indians and India did not under-report the number of Covid deaths in Mumbai (both are big ifs), you could apply the 0.14% mortality from Covid-19  to 330 million Americans and get 460,000 people. I think the realistic number would be higher, given deaths to date in the US.  

I think the world has been very lucky to have been affected by a pandemic that has such a low mortality rate.  It could easily have been 20 times worse, perhaps next time?  In the Middle Ages, the Black Death killed hundreds of millions of people - a truly apocalyptic pandemic.

There is no certainty that a vaccine is going solve the Covid-19 problem, indeed the UK government is buying 12 different vaccines, in the hope that one is effective.  Vaccines are often least effective in older people, who are main risk group for Covid-19.

If no vaccine turns out to be 90% effective, the Mumbai slum dwellers and the Swedes will have been the smart ones.


Controlled Infection vs Vaccination

If I was a dentist I would be seriously worried about Covid-19. I would favor a small infection today, caught from my party-going offspring, rather than in two year's time catch it while peering into a stranger's mouth during an hour long procedure, and get a huge initial exposure, leading to a more severe infection.  The fact that Mumbai policemen, London bus drivers and of course doctors and nurses without good PPE have had so many fatalities does suggest the amount of virus you are initially exposed to is a critical factor to the outcome.  This would be logical anyway.

I am really glad at least my older son and myself have had Covid-19.  If I was a dentist, I would be hugely relieved. A few months ago we assumed Covid-19 was both highly infectious and often deadly, now we know the reality.  If you are youngish, slim and healthy the risk is very low.  Many in rich societies are old, overweight and in poor health.

I did take my younger son Monty, aged 17 with autism, for a visit to the dentist two months ago and I really felt sorry for her.  She was wearing a mask, but that is no guarantee of her safety.